06:22Nick Saraev · Tutorial
How to Use Fable In Every Aspect Of Your Life
A six-minute walkthrough of a screen-watching AI agent that logs your habits, waits for a pattern to repeat three times, then hands you the fix.
July 9thThe argument in one line.
Clear thinking is not a discipline trait but a stack of hardware and software upgrades, and almost every gain comes from redesigning your environment once instead of resisting temptation daily.
Read if. Skip if.
READ IF YOU ARE…
- An operator or founder who makes a handful of expensive calls each quarter and wants a repeatable way to price them instead of going on gut feeling.
- Someone who has read productivity content for years, knows what to do, and still does not do it, and wants the mechanism behind that gap named.
- A knowledge worker whose output has plateaued and who suspects the ceiling is environmental (air, noise, notifications, sleep) rather than motivational.
- Anyone using AI daily who has noticed the model agrees with them too often and wants a concrete prompting method that fixes it.
- A self-taught learner who wants the rationality canon (Bayes, expected value, Goodhart, Chesterton's fence) explained in plain language rather than notation.
SKIP IF…
- You want tactics for a specific tool or workflow. This is decision theory and physiology, with no software walkthrough anywhere in it.
- You already have a working rationality practice and log your own predictions. Most of the software half will be revision.
- You need clinical guidance on sleep, nutrition, or supplements. The percentage improvements quoted are the presenter's own first-principles estimates layered on top of the literature, not medical advice.
- You dislike long single-take lecture formats. It is one person reading a document he wrote, for five straight hours.
The full version, fast.
Thinking is cognition running on hardware (attention, working memory, executive function) and software (the models you reason with), and both are upgradeable without extra willpower. On the hardware side, sleep, air quality, intelligible speech, light, clutter, notifications and cardio stack multiplicatively to roughly 150 to 200 percent more clarity, and none of them require discipline once you change the default. On the software side, expected value, power laws, the chain law, local versus global maxima, Bayes in natural frequencies, Fermi estimates, opportunity cost and the planning fallacy multiplier give you named bins for problems you were previously solving by gut. The execution layer is trigger action plans, noticing your own confusion, premortems and five-minute resolve cycles. AI belongs on the adversarial side of the loop: ask it why your plan fails, never let it write the decision line.
Free for members
Chat with this breakdown — free.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
Create a free account →Where the time goes.
00:00 – 06:05
01 · Cold open and the course map
The promise, the neuroscience background, and a full table of contents read aloud before anything starts
06:05 – 11:19
02 · The three parts of cognition
Attention as gatekeeper, working memory as workbench, executive function as foreman
11:19 – 14:13
03 · A stingy brain on a fixed budget
Two percent of body mass, twenty percent of energy, and no measurable increase when the task gets hard
14:13 – 23:21
04 · Working memory and chunking
The four-chunk limit, and how meaning compresses 22 letters into three items and then into one
23:21 – 32:08
05 · Enriched cages and clean air
Mice with toys grew 15 percent more granule neurons; knowledge workers in ventilated air scored double
32:08 – 46:16
06 · Willpower is environment design
The reservoir model is wrong, asymmetric bets, and the if-then rule that ended a weed habit
46:16 – 56:48
07 · Choice architecture and defaults
Organ donation at 12 percent versus 99.98 percent, and the do-nothing diagnostic for finding your own defaults
56:48 – 1:03:40
08 · Friction, the golden rule
Cafeteria studies on ten inches and tongs, and Scott Alexander on the trivial inconvenience
1:03:40 – 1:20:16
09 · Hardware: sleep first
Rule out sleep-disordered breathing, then 7 to 9 hours, fixed wake time, caffeine curfew, light and temperature
1:20:16 – 1:28:34
10 · Bloodwork, alcohol, protein, creatine
The standard panel, why alcohol costs REM sleep, and the one supplement with real evidence
1:28:34 – 1:34:15
11 · VO2 max and the REHIT protocol
Cardio as the single largest cognitive lever, and 18 total minutes of hard effort over six weeks
1:34:15 – 1:44:30
12 · Workspace: air, speech, light, clutter
CO2 thresholds, the verbal reasoning tax of overheard conversation, window light, and biased competition
1:44:30 – 1:54:09
13 · Notifications and single-tasking
Batching alerts into two or three windows a day and stripping the screen down to one thing
1:54:09 – 2:10:47
14 · Expected value and the resulting trap
Coin flips, a SaaS versus a service business, a sales hire, and why a losing bet can still be right
2:10:47 – 2:27:24
15 · Power laws and the chain law
One client equals sixteen, Pareto in Microsoft's bug queue, and why chains compound down
2:27:24 – 2:39:52
16 · Local maxima and attentional residue
Stepping down a hill to climb a taller one, and the resume note that unsticks task switching
2:39:52 – 2:47:50
17 · Akrasia: the doing bottleneck
Scott Alexander on rationalists, the limiting reagent, and 45,500 community threads that agree
2:47:50 – 2:57:47
18 · The map is not the territory
CRMs, dashboards, direct reports, and rats that got faster because they were hungrier
2:57:47 – 3:03:48
19 · Ulysses contracts
Odysseus at the mast, phones in other rooms, public commitments, and paying a competitor
3:03:48 – 3:12:03
20 · Opportunity cost
Bastiat's broken window, price versus true cost, and every yes as a no
3:12:03 – 3:23:07
21 · Goodhart's law and Chesterton's fence
Rat tails in colonial Hanoi, countermetrics, and why you should learn a rule before deleting it
3:23:07 – 3:31:28
22 · Bayes in natural frequencies
Priors, odds instead of percentages, and which sales-call sentences are actually evidence
3:31:28 – 3:38:45
23 · Fermi approximations
Decompose, estimate, multiply, sanity-checked against real cold-email numbers
3:38:45 – 3:50:38
24 · Hourly rate and the offload rule
Average versus marginal rate, the hidden supervision cost, and the 2.5x delegation threshold
3:50:38 – 3:57:04
25 · The planning fallacy multiplier
258 transport projects, confidence curves, and picking a personal constant
3:57:04 – 4:05:08
26 · Explore versus exploit
Slot machines, time horizons, and a revenue history mapped onto explore and exploit seasons
4:05:08 – 4:20:54
27 · TAPs: trigger action plans
Obvious, visceral, reliable triggers, small physical actions, ten rehearsals, and chained morning taps
4:20:54 – 4:33:30
28 · Noticing: confusion, flinch, should
Three signals and three five-second moves that compound like an index fund
4:33:30 – 4:47:14
29 · Premortems and resolve cycles
Writing the failure story in past tense, then five-minute timers that end procrastination loops
4:47:14 – 4:55:32
30 · Using AI without losing your edge
Sycophancy as a broken map, four red-team prompts, and never letting a model write the decision line
4:55:32 – 5:02:20
31 · Recap and close
A spoken index of all 27 concepts, framed as an exponential return on better decisions
Lines worth screenshotting.
- The brain burns the same energy staring out a window as it does solving a calculus problem, so the calories are spent either way.
- People with apparently high willpower are not winning fights against temptation. They arranged their lives so the fight never starts.
- Working memory holds about four meaningful chunks, so expertise is mostly the ability to pack more meaning into each chunk.
- Twenty-four knowledge workers scored roughly double on cognitive tests on days when the office air was well ventilated. Nobody slept more or tried harder.
- Moving a bowl of peas ten inches further away in a cafeteria dropped consumption 13 percent, and swapping the spoon for tongs dropped it 16.5 percent.
- In opt-out countries organ donation registration sits above 85 percent. In opt-in countries it is under 28 percent. The difference is one line of policy.
- Office CO2 at 1,400 parts per million cut cognitive function scores in half against a 550 ppm baseline.
- Following an intelligible conversation while you work taxes verbal reasoning by roughly 44 percent, and writing is the task that suffers most.
- A single notification can cost 10 to 15 percent of cognitive capacity for up to 20 minutes, and most people get interrupted more often than that.
- Six hours of sleep a night accumulates the deficit of a full sleepless night within two weeks, and subjects cannot detect their own impairment.
- Three ten-minute bike sessions a week, with only one minute of hard effort in each, raised sedentary adults' VO2 max 15 percent in six weeks.
- Between 1988 and 2003 hardware got about a thousand times faster while algorithms got ten thousand times faster, which is the case for upgrading your software first.
- A good decision that loses is still a good decision. Judging choices by their outcomes is called resulting and it teaches you the wrong lesson.
- Power laws govern which system to pick. The chain law governs how to improve the one you already have. Confusing the two wastes years.
- Chains compound down: a three-step process where every step succeeds 90 percent of the time finishes only 72.9 percent of the time.
- Most people who feel stuck are not at the bottom of a hill. They are at the top of a small one, and the only move is down before up.
- Writing a resume note before you switch tasks cuts attentional residue from 5 to 10 percent down to 1 or 2 percent.
- Knowledge is almost never the bottleneck. Across 45,500 community threads, the people who win are the ones who act before they understand.
- A prospect saying 'this looks great, send me the details' is worthless evidence, because buyers and non-buyers say it at the same rate.
- Nine out of ten large transport projects overran their budget, averaging 28 percent, and those were experienced engineers with decades of comparable data.
- People who claim 99 percent confidence in a deadline hit it 45 percent of the time.
- A 2006 meta-analysis put if-then implementation intentions at 0.65 standard deviations of improvement, which moves you from the 50th to roughly the 75th percentile.
- Premortems beat brainstorming risks because the brain answers 'why did it fail' with a story and 'why might it fail' with an abstraction.
- Asking an AI 'is this good' leaks your hope into the prompt and gets your own opinion back wearing a lab coat.
- Your time is only worth your hourly rate if you actually use the hour you bought, otherwise you paid someone $150 so you could scroll.
Redesign the room before you blame the discipline.
WHAT TO LEARN
Nearly every gain in this course comes from a decision made once while clear-headed, not from resisting anything in the moment.
02The three parts of cognition
- Every thought passes through attention, working memory and executive function, so a focus problem is really a question of which of those three is failing.
- Expertise is compression. A chess master and a tax accountant do not hold more items than you, they hold bigger ones.
03A stingy brain on a fixed budget
- The brain spends roughly the same energy on hard and easy tasks, so the metabolic cost of thinking well is already paid whether you use it or not.
- Because the brain is stingy, it defaults to habit over decision, which is exactly the lever every environmental technique in the course pulls on.
04Working memory and chunking
- Working memory holds about four meaningful chunks, and everything from note-taking to writing while you think exists to expand that window.
- Meaning is what compresses information, so learning a domain deeply raises how much of it you can hold at once rather than raising raw memory.
05Enriched cages and clean air
- Mice in enriched cages grew 15 percent more granule neurons with no drugs and no training, purely from a better enclosure.
- Knowledge workers scored about double on cognitive tests when the office air was clean, which is a reminder that environment beats effort at the margin.
06Willpower is environment design
- Willpower is not a reservoir. The people who look disciplined have simply removed the situations where a mistake is an available move.
- Treat willpower as limited even if it is not, because the belief that costs you nothing to hold builds a better environment either way.
- Never write a plan whose step is 'I will resist X'. Reduce access to X instead and the resisting disappears.
07Choice architecture and defaults
- Almost every option you encounter was arranged by someone whose KPI is not yours, from autoplay to pre-selected annual plans.
- Opt-out organ donation runs above 85 percent and opt-in under 28 percent, which is how much a single default can move a whole population.
- Ask what happens at every transition in your day if you do nothing. Whoever benefits from that answer set the default.
08Friction, the golden rule
- Friction is proportionate: make something three times harder and you will do it roughly three times less, which works in both directions.
- A trivial inconvenience of one or two seconds stops most people most of the time, which is why small environmental edits outperform resolutions.
09Hardware: sleep first
- Rule out sleep-disordered breathing before optimizing anything else, since roughly 30 percent of US men are suspected to have some form of it.
- Six-hour sleepers accumulate real deficit within two weeks while their self-reported sleepiness plateaus, so self-assessment cannot detect the damage.
- A fixed wake time entrains the circadian clock better than a fixed bedtime, and morning light is what actually sets it.
- Caffeine's half-life means a noon cutoff for a 10pm bedtime, and 400mg six hours before bed costs a measured hour of sleep people never notice.
10Bloodwork, alcohol, protein, creatine
- Bloodwork is a one-time few-hundred-dollar check that catches iron, ferritin, B12, vitamin D and thyroid issues that read as laziness.
- Alcohol's cognitive cost is not the liver, it is the suppressed REM in the second half of the night that blocks memory consolidation.
- Creatine at three to five grams a day is the only supplement with real evidence, and it matters most for people who eat little red meat.
11VO2 max and the REHIT protocol
- VO2 max is both the best single cognitive proxy and a stronger longevity predictor than most other risk factors.
- REHIT put 18 total minutes of hard effort across six weeks and raised sedentary adults' VO2 max 15 percent, which is why time is rarely the real barrier.
- Stack habits so one change covers several levers at once: exercising outdoors in the morning with a friend hits light, wake time, cardio and social contact together.
12Workspace: air, speech, light, clutter
- CO2 above about 800 parts per million starts measurably degrading cognition, and a closed office passes 1,000 within an hour or two.
- Overheard intelligible speech taxes verbal reasoning by up to 45 percent, and writing suffers the most of any task.
- Every object in your visual field consumes suppression effort, so clutter is a continuous low-grade tax rather than an aesthetic issue.
13Notifications and single-tasking
- One notification can cost 10 to 15 percent of capacity for up to 20 minutes, which for most people means permanent partial impairment.
- Batching notifications into two or three scheduled windows converts a constant tax into a couple of short ones.
- Single-tasking the screen means full screen, no other apps, no URL bar, no taskbar, with design and trading as the honest exceptions.
14Expected value and the resulting trap
- Expected value is the sum of each outcome's value times its probability, and it is the smallest upgrade that turns a gut call into a defensible one.
- The exciting option often loses on EV: a one-in-a-thousand shot at a billion is worth less than a boring business at half a million a year for five years.
- The real value of expected value is writing the probability down, because a written 20 percent is revisable and a feeling is not.
- Resulting is judging a decision by its outcome, and it is what makes people abandon correct bets that lose 80 percent of the time.
- Cold outreach is the everyday version of this: a 20 percent visibility rate means four failures in five are not evidence of a bad method.
15Power laws and the chain law
- Power laws make averages useless. If one client can be worth 25 others, the strategy is finding that client, not adding more of the small ones.
- Find the 20 percent by sorting your actual outputs in descending order, then either double down on the head or stop funding the tail.
- The chain law is the inverse: reliability is the product of the steps, so the weakest link caps everything and extra links lower the ceiling.
- Power laws tell you which system to pick. Chain laws tell you how to fix the one you have. Applying the wrong one wastes the effort entirely.
16Local maxima and attentional residue
- A plateau usually means a local maximum, and the three tells are shrinking returns, unresponsive inputs, and less skilled people outperforming you elsewhere.
- Cross the valley by quantifying the loss you will accept in advance and by allocating 10 to 20 percent of hours to the new hill rather than jumping.
- Attentional residue means part of your attention stays on the old task after a switch, costing roughly 5 to 10 percent until it clears.
- A one-minute resume note written before switching cuts that residue to 1 or 2 percent, which is the cheapest fix in the entire software half.
17Akrasia: the doing bottleneck
- Akrasia, not ignorance, is the bottleneck: across 45,500 community threads the winners were consistently the ones acting, not the most knowledgeable.
- The fix for akrasia is not more information. It is shrinking the next action until doing it is trivial and attaching it to a trigger.
18The map is not the territory
- Every dashboard, CRM, metric and direct report is a lossy compression of reality, and compression is exactly what makes them useful and wrong.
- When the map and the territory disagree, the territory wins. Go look at the source instead of checking another summary.
- AI output is a map of a map, so verify against the raw call recording or source data before making an expensive decision on a summary.
19Ulysses contracts
- A Ulysses contract moves the decision to when you are clear-headed and removes the tempted version of you from the loop entirely.
- The three practical forms are physical distance, public accountability, and money staked on a party you would hate to fund.
20Opportunity cost
- Opportunity cost is the single best alternative you gave up, not the sum of all of them, and people ignore it unless it is spelled out.
- A $30 lunch that eats 90 minutes really costs closer to $100 once you price the time at your own rate.
- Make the no explicit: name the specific thing you will not be doing every time you say yes to something vague.
21Goodhart's law and Chesterton's fence
- When a measure becomes a target it stops being a good measure, which is why paying for closed tickets produces closed tickets and not solved problems.
- Defend against Goodhart by pairing every target with a countermetric that gets worse when the first one is gamed.
- Before deleting a rule, process or constraint, find out why it exists, because someone probably paid to learn that lesson already.
- If you cannot find the reason, ask the person who built it, check the history, then remove it only if the removal is reversible.
22Bayes in natural frequencies
- Bayesian thinking replaces true or false with probabilities you update, one small piece of evidence at a time.
- Evidence is only evidence if it is more likely under one hypothesis than the other, which is why 'send me the details' tells you nothing.
- Work in odds and multiply rather than adding percentages, so a quick reply, a pricing question and a no-show each move the number proportionally.
23Fermi approximations
- Fermi estimates work by decomposing, estimating within about 3x, and multiplying, with errors in opposite directions cancelling out.
- Round aggressively and accept an order of magnitude, because the point is to stop wasting time on ideas that fail the sniff test.
24Hourly rate and the offload rule
- You need two hourly rates: the average across all working hours, and the marginal rate of your single highest-value activity.
- Offload only when hours times your marginal rate exceeds price plus supervision hours times that rate, and never assume supervision is free.
- Some tasks should never be offloaded regardless of the math, because they are the thing customers are actually paying for.
- For solo operators, offloading often loses to the chain law, since adding a person adds a step whose success rate caps the whole process.
25The planning fallacy multiplier
- Plans are best-case simulations, which is why 9 out of 10 large infrastructure projects overran, averaging 28 percent.
- Apply a fixed planning fallacy multiplier of 1.5 to 3 to your own estimate, scope steps tightly internally, and quote the multiplied number externally.
26Explore versus exploit
- Explore when the time horizon is long or returns are declining, exploit when the horizon is short or you have just found a winner.
- Exploring and exploiting are seasons, not personality traits, and both failure modes have flattering names: focused, or always learning.
27TAPs: trigger action plans
- A trigger action plan is an if-then rule where the trigger is obvious, visceral and reliable, and the action is small and physical.
- Implementation intentions moved goal attainment by 0.65 standard deviations across 94 studies, roughly a jump from the 50th to the 75th percentile.
- Install a TAP by rehearsing it: visualize the trigger in sensory detail, physically run the sequence ten times, and debug only the trigger when it misfires.
- Chain TAPs so each action becomes the next cue, which is how a morning routine becomes a 90-second automatic sequence.
28Noticing: confusion, flinch, should
- Noticing confusion means saying it out loud before your mind buries it, because confusion is the clearest signal your map is wrong.
- A flinch is an opportunity, and sitting with it for ten seconds trains the opposite circuit from avoidance.
- Interrogate every 'should' for whose goal it serves, then either convert it into a concrete next step or drop it.
29Premortems and resolve cycles
- Premortems beat risk brainstorms because past tense makes the brain produce specific stories rather than abstract categories.
- Fifteen minutes and one page of premortem replaces weeks of actual failure, and AI is genuinely good at generating the failure narratives.
- Resolve cycles cap a dreaded task at five minutes, and all three possible outcomes are wins.
- If five minutes is not enough, spend the next five writing the concrete next steps, since a vague next action is what akrasia feeds on.
- Write while you think and define the end state in advance, because 'I have thought about it' is not an end state.
30Using AI without losing your edge
- Models are trained to maximize human ratings, so asking whether your idea is good leaks your hope into the prompt and gets it mirrored back.
- Ask why the idea is bad instead, and use reference-class, premortem and Fermi prompts to force the model into adversarial mode.
- Never let AI write the final decision line, because you bear the consequences and articulating the reasoning is the part that keeps the skill alive.
Terms worth knowing.
- Chunk
- A single meaningful unit of information in working memory. Eight loose letters are eight chunks, but rearranged into one recognizable word they become one, which is why expertise increases apparent memory.
- Choice architecture
- The arrangement of options that influences which one gets picked. Coined in the book Nudge, the core claim is that there is no neutral way to present more than one option.
- Friction
- Any cost sitting between a person and a behavior: an extra click, an extra walk, an extra password. Add it to reduce a behavior, remove it to increase one.
- UARS
- Upper airway resistance syndrome, the quieter cousin of sleep apnea. It fragments sleep dozens of times an hour without waking you fully, and shows up as brain fog rather than snoring.
- REHIT
- Reduced exertion high intensity interval training. A ten-minute session where only 40 to 60 seconds are all-out sprints, used to raise VO2 max with minimal total time.
- Expected value
- The sum of each possible outcome's value multiplied by its probability. It converts a yes-or-no gut call into a number you can write down, share, and revise later.
- Resulting
- Judging the quality of a decision by how it turned out. Named by Annie Duke in Thinking in Bets, it is what makes people abandon correct high-variance bets after one loss.
- Power law
- A distribution where one or two items account for most of the outcome. It makes averages misleading and turns strategy into finding and doubling down on the head.
- Chain law
- The inverse of the power law. System reliability is the product of its steps, not the average, so the worst link caps the whole and every extra link lowers the ceiling.
- Local maximum
- The top of the hill you happen to be on. Every available move from there goes down, even though a taller hill exists nearby, which is why plateaus feel like success.
- Attentional residue
- The portion of attention that stays stuck on a previous task after you switch. Named by Sophie Leroy in 2009; a written resume note before switching shrinks it sharply.
- Akrasia
- A Greek term for knowing the right action and not taking it. It is the gap between what you know and what you do, and it is the usual limiting factor rather than knowledge.
- Map versus territory
- The principle that every model, metric, dashboard or report is a lossy compression of reality. Compression is what makes it useful, and also what makes it wrong at the edges.
- Ulysses contract
- A commitment made while clear-headed that removes your future self's ability to choose badly. Named for Odysseus having himself tied to the mast before the sirens.
- Opportunity cost
- The value of the single best alternative you gave up, not the sum of all alternatives. It is why a $30 lunch that eats 90 minutes can really cost close to $100.
- Goodhart's law
- When a measure becomes a target it stops being a good measure. Proposed by Charles Goodhart in 1975 and visible any time a metric gets gamed instead of improved.
- Chesterton's fence
- The rule that you should understand why a rule, process or constraint exists before removing it, because the expense of putting it there was probably paid by someone who learned something.
- Bayesian prior
- Your starting estimate of how likely something is, before new evidence arrives. Updating it with each piece of evidence is what turns opinions into revisable probabilities.
- Natural frequencies
- Expressing probability as odds like 1 in 99 rather than percentages, so Bayesian updates become simple multiplication instead of conditional probability notation.
- Fermi approximation
- Estimating an unknown quantity by decomposing it into parts, guessing each part, and multiplying. Individual errors tend to cancel, landing the answer within an order of magnitude.
- Marginal hourly rate
- What an hour spent on your single highest-value activity is worth, as opposed to your average hourly rate across all working hours. It is the number that governs delegation.
- Planning fallacy multiplier
- A fixed constant, commonly 1.5 to 3, applied to your own time estimate to correct for the fact that plans are always best-case simulations.
- Explore-exploit tradeoff
- The choice between milking a known-good option and sampling unknown ones. Long time horizons favor exploring; short horizons and declining returns favor exploiting.
- TAP
- A trigger action plan, known in the literature as an implementation intention. A pre-written if-then rule linking an obvious physical cue to one small concrete action.
- Premortem
- A planning exercise where you assume the project already failed and write the story of why. The past tense forces concrete failure modes instead of abstract risks.
- Red teaming
- Deliberately attacking your own plan to find its weaknesses. Applied to AI, it means asking the model why an idea is bad rather than whether it is good.
- Sycophancy
- The tendency of language models, trained to maximize human ratings, to agree with the user. It makes the model a map that tells you you are heading the right way regardless.
Things they pointed at.
46:30bookNudge (book, origin of choice architecture)
1:00:23linkScott Alexander, 2009 essay on trivial inconveniences
2:04:40bookAnnie Duke, Thinking in Bets (2018)
3:31:20linkEliezer Yudkowsky's intuitive explanation of Bayes' theorem
1:30:40productEight Sleep mattress topper
1:28:20productAirthings CO2 monitor
1:36:00linkAllen et al. 2016, CO2 and cognitive function
3:51:00linkFlyvbjerg study of 258 transport projects
4:07:30linkPeter Gollwitzer, implementation intentions
2:34:40linkSophie Leroy, attentional residue (2009)
4:48:00linkAnthropic research on sycophancy in Claude models
Lines you could clip.
13:00
“From a metabolic perspective, your brain is going to be burning those calories anyway. Why not use those calories as efficiently and as effectively as possible?”
33:33
“Think bowling with the guardrails up. Landing in the gutters is simply not one of the possible mistakes.”
37:05
“If one of the steps in your plan is I just won't do that bad thing that I always do, it's not a plan. It is a hope.”
51:55
“If a default is set by a company and you do nothing, you get what they want. But if you set a default and you do nothing, you get what you want.”
1:40:23
“The purpose of business is to be effective, right?”
1:02:00
“From the outside, it all looks like discipline.”
1:50:20
“You are basically paying somebody $150 to scroll TikTok.”
2:25:00
“Most people who are stuck in life or career, they're not at the bottom of the hill. They're at the top of one.”
2:42:00
“It is not the most knowledgeable people that consistently win. It is always the people that act.”
3:10:40
“Every time you say yes to something, you are always saying no to something else.”
3:36:50
“One trustworthy step is worth 12 clever offloaded steps.”
4:50:30
“It's almost like your first opinion, just wearing a lab coat.”
4:27:39
“You are a donut-shaped sack of meat around a hole that metabolizes and poops out food. Why should you call somebody else?”
5:01:00
“The gap between you and a less defined version of yourself does not grow linearly, but exponentially.”
Word for word.
Read-along
Don't just watch it. Burn it in.
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
metaphoranalogystory
So this video is five hours long, and by the end of it, you will understand how to think clearly in a world that explicitly profits out of you not doing so. I'm not a big productivity guy. I am a business owner who shares what I've learned on the internet.
I've made over a million dollars a year in profit for the last three years, and I've made almost five million dollars in the last 12 months. I say all of this because I firmly believe that once you figure out how to think clearly, as well how to act clearly.
And so it is the thinking clearly that is the foundation behind being effective in a lot of real world settings. In my case, you know, entrepreneurship, but in your case, maybe something else. So in my view, over the course of the last maybe two or three years, AI and other algorithms have negatively impacted our ability to think for ourselves.
One of the goals of my course as somebody that talks a lot about AI is to help people take some of their ability to think. back from these models, and hopefully get to a place where you can both think deeply for yourself and achieve the myriad gains that thinking deeply gives you, while also being able to utilize this technology effectively, because I believe it does have the promise to be used very effectively.
Now you guys won't know this if you watch my channel often, but before I went into business, I actually studied behavioral neuroscience for around five years at a research university. And during this time, I worked in a variety of labs. We performed experiments on animals to determine how various things affected their ability to literally think.
In the literature, this is referred to as their cognition. I changed stuff like their eating schedules. I changed who they got to hang out with.
I changed the lighting in their cages. I changed how they were trained and so on and so forth. And so a lot of the principles that I'm going to be talking about in this course.
or taken directly from my time back in university. Lastly, before we get into the course, I want to make it clear that this is not AI slop. I wrote all of this myself using the same principles that I'm going to share with you.
I had Claude Fable generate diagrams based on the data, but all the words are mine. And I poured many, many hours into this because I frankly think it is probably one of the most valuable things that I could share with you. So without further ado, let me show you how to think clearly.
First, a lot of these terms will probably seem kind of senseless to you at face value, and I'll do my best to explain them, but you're going to learn them later on in the course. There's no way that you can't. I'm planning on going in depth into both research and then practical applications of all this stuff.
The whole idea being that you walk away with both deep underlying theory, which will allow you to categorize problems in real life into the bins that I provide you, but also actionables, aka real strategies and immediately implementable techniques that you guys could take immediately after this course to go and solve problems that you're struggling with.
What I'm going to do is front load with the hardware sections of the course. And then finally, we'll chat software before running you guys through those immediately actionable skills. So here is the outline of our course.
I'm going to start with the biological basis of cognition. Then we'll talk about scientifically defining self -control. Afterwards, we'll talk a concept called choice architecture.
Then I'll discuss friction in action and behavior. Afterwards, we'll start optimizing our biology. Then we'll optimize our environment, our workspace, our virtual environment, and so on.
Then I'll cover a concept the ancient Greeks talked about called akrasia, which is where you know something, but you cannot do it, and how to solve that. We'll then run through expected value methodology and how to use it to accomplish tasks. I'll then cover local and global maxime, which is a mental model handed to me by calculus and optimization math.
Afterwards, we'll cover what is called attentional residue. I'll talk about models like the map versus the territory, extremely important. You'll learn about Ulysses or Odysseus contracts for people here that have seen the Odyssey recently.
I'll then talk power law. I'll talk the chain law. I'll teach you guys about opportunity cost.
Incentives, what's called good heart. If you guys haven't heard of Chesterton's fence, which is a strong thinking technique, I'll run you guys through Chesterton's fence. I'll then talk about Bayesian theorem and Bayesian probability theory in natural frequencies.
This is a very strong way to solve problems. You'll learn Fermi approximates, which are ways of estimating magnitudes and sizes. I'll run you guys through hourly rate and offload rules, planning fallacies, and how to multiply your way out of them.
I'll cover the explore -exploit trade -off, which is increasingly relevant with AI agents and models. And then finally, we'll actually cover these immediately actionable techniques, like taps, noticing. I'll show you guys how to run what are called premortems.
I'll then show you guys how to... use five minute timers in a very strategic and tactical way to solve problems using structured decision making. And then finally, at the end of this course, I'm going to show you guys how to use AI intelligently and not be overwhelmed or destroyed cognitively by these models.
Now, that's obviously a lot of information. And because of that, I should take a moment to discuss how best to take this course and actually cram it in your brain. Because as you guys will find, the sort of meta planning, even a few moments of that, significantly improves the ability of your brain to consolidate this information into something useful over time.
So how do you actually take this course? Well, I would recommend you start by watching or listening to the course passively at 1 .5x speed. Now this will obviously take you a lot of time, but not as much time as if you played it at 1x.
After you're done, I would watch the course at 1x speed. So I'd actually slow it down. After every section, you guys will find that I've added these action questions down to the bottom.
Now I like to do that because quizzes, short little questions that you answer, force you to retrieve information from the various corners of your mind that we've just stuffed it in. And what you'll find is encoding or putting information into the system is quite different from retrieval, which is getting information out of the system.
It's important that you experiment and practice both if you really want to consolidate concepts. So I'd recommend after you play it back at 1x speed, answer those action questions if you really want to get the most out of the material. Finally, this entire document is available just in the link in the description, so you guys can download it and then keep it as a companion guide for the short term, for the long term, for a study partner, or whatever it is that you want.
So let's talk about the biological basis of cognition, shall we? For those of you guys that don't know, thinking, it's also referred to as cognition, it is a biological process. And understanding that biological process is very important if you want to learn how to optimize it.
I'm not going to sit here and then pretend to give you guys, you know, a biology or neurobiology 101. So I'm going to spare you most of the details. And instead, I just want to leave you with probably the most useful definition I've found so far.
Back when I was pursuing work in a cognitive sciences lab in university, which I did not end up getting, but this was still useful nonetheless. Now, cognition, your ability to think, is composed of three parts. You have your attention, you have your working memory, and then you have your executive function.
Now put another way, any thought that you have, even the thoughts that you are having right now as you are watching this course, is a byproduct of three things. The first is attention itself. And a good way of picturing attention is like a gatekeeper that lets in a small slice of the incredible amount of sensory information that is constantly flooding in from the outside world.
Think things like the feeling of the headphones in your ear, if you are wearing them, or where your computer is, or your wrists on the desk, or the various background noises outside, wind, flapping of birds, sounds from your family members, whatever it is. All of this is currently flooding into your brain constantly. Yet you're not constantly aware of it.
Why? Well, that's because . Your brain can put that spotlight on the things that it wants you to attend to.
And so attention is a gatekeeper, like that spotlight that chooses what it is in our environment we're actually going to be lighting up. Working memory is the next step. And that is .
If you guys have ever worked with any sort of heavy machinery or tools, you know that that workbench, you know, can typically hold a few things simultaneously. And the whole idea of a workbench is you put something on the workbench, you work on it, you get rid of it, you finish it. and then you put it out away.
You clear the workbench so that somebody else, or let's say you in the future, can put new items on the workbench and then repeat that process again. That is what working memory is, and that's an excellent way of conceptualizing it. The next is executive function.
The best way I've found of conceptualizing executive function is like a foreman or a general contractor, if you guys have experience with construction. These are the people that ultimately decide what goes on the workbench and who works on the workbench. You have all these things that are constantly flooding into your working memory from your attention because it is the gatekeeper that permits things, while the executive function is the foreman that decides what gets thrown out, when you should stop one task and move to another, etc.
They are the project managers of your thinking. Now the problem is, the majority of human beings currently operate with next to no attention, very poor working memory, and a not very functional executive function. In practice, this leads to you feeling fuzzy all the time.
You really struggle focusing on anything that is actually important. You struggle with carrying an idea through to its logical conclusion. A lot of the time, things probably feel confusing or unnecessarily difficult, and things, like maybe this course, just feel like they're happening a little bit too quickly.
This often frustrates you constantly, and it also leads to a relatively emotionally fragile state, which I'll cover more in the course. So how do you actually fix this? I mean, it sounds like a pretty big problem, right?
We have attention, we have working memory, we have executive function. These are obviously deep cognitive science concepts, and you're learning them for the first time. Just like a video game character creator screen, you can actually just pour bonus stats from various hardware and software optimizations that I will teach you in this course directly into these skills of your attention, your working memory, and then your executive function.
You know, we can find hardware improvements through things like health optimization, better nutrition, circadian rhythm optimization, and so on. And then in time, we can distribute these free wins to parts like your attention, working memory, and executive function to allow you to level up each of these. Then we can also take various software wins, which are problem -solving techniques and so on and so forth, to also fundamentally improve your executive function as well.
And that is simply all we are going to be doing today. I'm going to be going through, like a caveman, finding all the extra bonus stats that you're probably leaving on the table. Now we're just going to combine them to make the best video game character you've ever seen.
So let's cover a little bit about the biology behind the brain. Now if you didn't already know, the human brain is extremely expensive. By weight, the brain makes up around 2 % of our total body weight.
But at the same time, it consumes approximately 20 % of our body's energy at rest. That means that per unit weight, the brain is actually the most metabolically expensive tissue in the entire body. And you can see that with this graph right over here.
Despite having a small share of body mass, only around 2 % or so, it includes over 20 % of all resting energy usage. And obviously this is very disproportionate. A very, very light organ, but one that requires a lot of energy to fuel.
Now, because of this incredible amount of energy per unit, there are two major implications and frames that I want you to think about this course through. The first is that if something is expensive energetically, then evolution and other pressures will always force it to minimize costs. And the brain is no exception to that.
That means that the brain will always prefer a habit to a decision. It'll prefer, you know, a guess to a calculation.
to the new, and so on. You can think of this as the brain budgeting its energy, you know, like a spreadsheet or something like that, trying to spend as little energy as humanly possible, assuming energy is a resource. Now, many of the techniques that I'm going to be discussing with you guys, things like chunking or implementation intentions, and so on and so forth, they work because they exploit this inherent stinginess, which is currently operating by default, but which you can actually learn to utilize to think more clearly and be more effective.
Now, much more surprising than that is that the brain's energy consumption is actually fixed. What I mean by this is, throughout the day, there's actually no statistically significant increase in whole brain energy consumption during difficult tasks versus simple and easy tasks. At most, any local energy consumption difference caps out at around 5 % or so.
Now, this has interesting implications because it means that, from a metabolic perspective, you burn just as much energy solving a calculus problem, and really racking your brain with it than you would if you were to just stare out the window. Now, this, I think, is at the core of why it makes sense to improve your ability to think and be effective.
Because the core reality is, from a metabolic perspective, your brain is going to be burning those calories anyway. Why not use those calories as efficiently and as effectively as possible? Part of me feels like, as an animal, I have a moral and ethical duty to use every one of those damn calories.
in a way such that I positively impact the world. And I think that if you guys understand that perspective, it'll probably inherently motivate you to learn how to do that as well. Now, there's this notion of willpower, which is the idea that there's a certain amount of effort you can apply to thinking, and that it is a naturally depletable resource.
Well, think about it from this perspective. Looking out the window and then doing a bunch of extremely economically valuable work consumes the same amount of energy. So where is that depletable resource?
The science doesn't actually check out with that explanation, and I'm going to explain to you guys in a moment how you guys can use a better concept of willpower to actually get things done. Anyway, for now, just know that the brain is expensive. It is always on.
It's also very stingy, and it budgets its energy as much as possible, typically by reducing the total capacity at all times. All of this puts you under a very strong constraint or optimization pressure, which, you know, if you want to think clearly, you have to maneuver. Okay, now that we've covered the brain in general, let's cover probably the most important singular component of the brain that most people would say is responsible for how clearly you think.
That's called your working memory. Now, if I asked you to repeat a sequence of 12 random digits immediately, like if I just gave you a random list of 12 numbers, and then had you walk away from the computer and come back a minute later and then repeat them, would you be able to do so? Statistically, no.
Very few people can. Now, it's not because you're dumb. although that may have something to do with it.
It's because of the incredibly small and natural size of a human being's working memory. I'm going to show you guys how you guys can increase or improve the size of your working memory in a moment. The current consensus in psychology, which is supported by many, many studies, is that you can hold approximately three to five meaningful items in your working memory at any one moment in time.
Because it's a range of three to five, most researchers will round this down to about four. For those not in the know, a meaningful item is also referred to as a chunk. And a chunk is any unit of information that your memory stores.
For instance, how many chunks would these letters be? E -I -L -N -D -D -A -E. Well, if it's not clear, that is eight chunks long, because it's eight letters.
Given the previous cap on three to five items stored, approximately four. Statistically, if you were to walk away from the computer for a second and then come back afterwards, most of you guys would not actually be able to repeat those letters. And the reason why is simply we do not have the working memory capacity.
There are only so many things we can store on our workbench at once. But the magic here is the meaningful part of meaningful items. Now, if you took those letters above, E -I -L -N -D -D -A -E, and then you reorganize them into something meaningful.
you can actually get a word out of them. For instance, if you reorganize E -I -L -N -D -D -A -E, you'll find that, putting all letters together magically, it turns into the word deadline. Now, if I flashed the word deadline on the screen for just a second, and then I asked you guys a moment later, hey, what word flashed on the screen?
Not a very tough question to answer, right? It's the word deadline, obviously. Of course I'd get it right.
So that takes me to an interesting idea. Why can't you remember... eight letters by themselves, like E -I -L -N -D -D -A -E, but when you organize all of those letters into a word, then it becomes really easy.
Well, it's because of meaning. Meaning is a layer that inherently organizes information into higher and higher levels of abstraction. Which means, rather than store the base data of E -I -L -N -D -D -A -E into your working memory, each as an individual chunk, you can actually just zoom out or reorganize the information a little bit into something that makes sense, and then store the entire thing as a single chunk instead.
Now, this continues up the ladder of abstraction, reducing chunk size the entire time, which, as you get good at things, allows you to store more and more information in your working memory at once. Put another way, it allows you to zoom out and see more of the picture at once. And as you guys know, seeing a lot of the picture at once is how you make good decisions and think clearly.
So, for example, let's say I had this sequence here. At face value, if you guys are listening and not watching, we have 22 chunks here, 22 letters. Now, this is obviously a lot of data, and the probability that you'd be able to memorize this whole long string of characters without spending a lot of time and energy invested on it is quite low.
But if I reorganize that information... You'll find that those letters, if put in a specific way, actually mean quarterly tax deadline. And to an average adult, we're no longer at 22 chunks anymore.
Quarterly means something. Tax means something. And deadline means something.
Which means we will have transformed these 22 chunks into just three. Now, if I flash quarterly tax deadline on the screen and walked away, you'd probably be able to memorize that and be able to repeat that back to me in a few moments, right? Of course.
But here's the really interesting part. You can actually go one level up, and that is where a high degree of skill becomes important. If I put the term quarterly tax deadline, not in front of an average adult, but I put it in front of a specialized tax accountant, you know, who has spent years internalizing the concept of a quarterly tax deadline, to them, it's not even three chunks anymore.
It's just one. Why? Because they are now storing quarterly tax deadline.
as a concept in and of itself. A quarterly tax deadline is not just the combination of three separate words, quarterly tax and deadline. It is a thing.
It is a constant reoccurring thing in an average tax accountant's life. It's something they're always looking up to and forward to. They recognize that as a standalone idea.
And that takes me to the following point. People who are exceptional at things can typically remember far more than an average adult, so long as they're in their zone of competence. The reason why is because throughout their lives, they've applied meaning to larger and larger groups of information, which allows them to chunk larger amounts of data in any single time.
A good example of this is a chess master. A world -famous chess master can remember the positions of pieces on a board far better than you or I can. And it's not because their brains are inherently better than ours.
They do not remember any more chunks, so to speak. What they do is they've simply combined all of the information into larger chunks themselves, and then they store those fewer chunks in the same area that you and I have available to us. Whereas you might see 32 distinct pieces on a board, they would see only three or four distinct patterns, and thus can store far more information with just four chunks than we can.
Another example. Can you remember the following number?
1492 -1776 -2001. Well, if I just said that to you guys and then walked away and came back in a minute and asked you to repeat it, the vast majority of you would not be able to do that. The reason why is because you're trying to remember 12 individual chunks.
But, for example, what if I told you that that number was actually just three chunks? 1492, which was the year that Columbus sailed to the New World. 1776, which is the year of the American Revolution, and 2001, which is the year that the Twin Towers fell.
If I asked you to now repeat those numbers back to me, it'd probably be far easier, right? Simply because you now know 1492, right, that's the year Columbus made it to the New World. 1776, right, that's the year of the American Revolution.
And 2001, right, that's the year that the Twin Towers fell. I'm doing this obviously for an American -centric audience. I'm Canadian, but hopefully you'll bear with me.
Okay, and so as we see on this diagram here, we started with 12 items, which was vastly over our capacity. You know, 1492, 1776, 2001. Very, very difficult sort of to remember it as 149217762001.
But because we're now applying meaning to this information, we're organizing it in a way that allows us to chunk them smaller and then store them all in our workbench simultaneously. Now, if I told you I was going to quiz you on this later, you'd probably remember it for a week or more. And studies have shown things like that.
You can be shown a long string of characters or numbers. You can then be told how the meaning organizes that into some sort of concept that you understand. And a week, a month, a year later, people can still understand that information, which is obviously far more efficient, right?
I mean, could you imagine working three times as hard to try and store 12 items in your working memory as somebody else who just knows a little bit more about the underlying patterns and facts and let's say, in this case, history, and can turn them into just three chunks? Why would you work three times as hard if you don't have to?
Well, this is a good example of exploiting that natural energy conservation concept of the brain. Now, this is just a very tiny example in the grand scheme of things. As I discussed with you guys earlier, cognition is composed of three parts.
We have attention, working memory, and executive function. And what we just discussed together, chunking, is just one very small and tactical way to slightly improve the working memory stat. In reality, there are actually dozens, if not hundreds of different ways that you can improve each of these three stats.
High performers and avenues where thinking is very essential, so think things like poker, or public speaking, or business, memory championships, and so on and so forth, they will take all of the techniques that I'm about to show you guys and apply them across the board in a very intelligent way so that they can achieve productivity and thinking gains of well over the 300 % that we just saw with that 1492, 1776, 2001 example.
And that's ultimately where we're getting to. In addition to talking about what they are, I'm also going to show you guys how you can apply them to real problems in your life. Until then, I want to talk to you guys a little bit more about the hardware to fully flesh out this idea of what thinking really means.
Now, as mentioned, I spent about half a decade in behavioral neuroscience. And a lot of that time was spent volunteering and later working in laboratories, whose whole goal was to manipulate the environment of animals, their food, their cage mates, the other animals around them, the time of day that their lights were on or off, their training, and a bunch of other things.
And then immediately after, test their resulting cognitive performance on big batteries of tests. And so you can think about this as... I, at the time, was volunteering and then later paid specifically to analyze what inputs resulted in clearer or less clear thinking.
And so the key takeaway I got from that experience was that counterintuitively, rather than thinking clearly being very hard, thinking clearly is actually quite easy. They're actually extremely simple and very mundane. but very powerful interventions that drastically improve animal performance.
And I guarantee you, you will not have thought about more than half of them. And the most powerful and impressive ones to me were not the ones that required injections or supplementation or various drugs or treatments. The most powerful ones were always things like a few hours difference in the feeding times of the animals.
Or which lights got turned off earlier or later than usual? Or which cage mates, which other rats or mice did they get to play around with before they went to bed? And you'll find that all of these made a massive statistically significant difference.
In many cases, between 30 to 50 % of improved or impacted performance, which turned them into basically different animals. And it all boiled down to simple changes that I don't think anybody here would probably bat their eyes at. These simple changes.
impacted their hardware. Now, obviously, it'd be unethical to perform these experiments at large on people. Okay, I'm sure some people have, unfortunately, but it would be unethical to do this sort of thing, obviously, unless somebody explicitly opted into it.
But with what limited research that, you know, neuroscientists have done, it is clear that human beings are quite similar to the animals studied in most of these experiments. Small changes in what you eat, or the strength of the lights that you are exposed to, or maybe the people in your periphery or your friends. or maybe the colors in your environment.
Believe it or not, all of these variables that you might not have otherwise noticed at all, if you stack them up, can actually make a massive difference to your ability to think clearly over time, for your brain to basically function. I'm going to run you guys through a bunch of examples of those in this section, and then give you a big laundry list of all possible ways that you can optimize your biology or hardware for the purposes of thinking.
And we're only going to do thinking in this course, although obviously you can optimize your biology for anything these days. Afterwards, I'm going to give you guys the programs that I talked about to stretch that hardware out as much as possible. So if you guys didn't know, there's this very famous 1997 experiment where a bunch of adult mice were put into what they called an enriched environment, which is more or less a cage that contains more cage mates, a bunch of running wheels, a bunch of tunnels and toys, and so on and so forth.
When they were compared to control mice, and control just means like the default, the mice that did not have all of this. What they found is that mice in the enriched environments showed massively increased granule cell layer thickness in an area of the brain called the dentate gyrus, which is the area of the brain that is typically responsible for new memory formation.
The cool thing is this was achieved solely through housing. There were no drugs and no training involved, which means quite literally the only difference was that their environment was more stimulating in a positive sense than the mice that did not see these gains. And you can actually see on this graph here that the granule neurons in the dentate gyrus indexed to 100 here.
In the standard cage, we're 100. In the enriched cage, which had all these extra things that I talked about, we hit 115. Now, what's important to note here is that the mice did not have to, you know, hustle and grind for these improvements.
They did not have to wake up at 5, watch a bunch of Instagram inspirational TikToks, and go to the gym. Really, they didn't get more disciplined at all. They also did not have to watch super comprehensive courses on how to improve your hardware and your software.
All that happened was they were, by circumstance, put into better enclosures. And as a result, their brains, their hardware, followed suit and became better. Now, a similar study on human beings.
There were 24 architects, programmers, engineers, managers, just a bunch of knowledge workers, basically, that were put into a fake office for six full days. These had the exact same desks. They had the exact same co -workers as before, and to my knowledge, they were also doing the exact same work.
The only thing that differed in that experiment was the air. How much of it came from outside, how much CO2 was inside of it, and how much off -gassing there was, which is something called volatile organic compounds, which typically are gassed off of paint, carpets, furniture, and things like that, typically when they're purchased new.
Now, every afternoon at 3pm, these 24 knowledge workers took the exact same set of cognitive tests to evaluate their ability to think clearly. Well, the shocking result is that on the days with very clean and well -ventilated air, the scores of these knowledge workers were literally double what they were on their normal office days.
That's not 10 % better. That is literally double. And the specific part of their thinking that improved the most was their strategy, their ability to plan, prioritize, and then sequence actions, all of which are executive function related.
Basically, their ability to think clearly. There wasn't a single person in that room that slept more. Nobody took supplements or nootropics.
Nobody tried harder. All of these people literally just breathed slightly different air and then became twice as good at certain batteries as a result. which is kind of nuts, right?
And as a business owner, that's extremely attractive to me because it means I can achieve 200 % output with the exact same input. We define that as leverage. It's free money.
If you can make 200 % with a minor change that has no real other impact on my day -to -day, why the hell wouldn't I do it? In general terms, you guys might have also heard of this concept as low -hanging fruit. And when you have low -hanging fruit, you should probably pluck it wherever possible.
It's low -hanging, right? That's what I'm going to be doing with you guys next. With zero additional effort, I want you guys to be able to output twice as much as you are currently outputting.
Or put another way, think twice as clearly. So our next few sections, which are on the scientific definitions of willpower, on choice architecture, on friction, on hardware optimization, and on environment optimization, you can think of as basically the human equivalent of designing a better cage. like those workers were in, or like those mice were in, compared to their default counterparts.
Now, it's miraculous to me, as somebody that's worked with modifying animal environments for performance reasons, that human beings have as much control over our own environments as we do. Because if you think about it, you can quite literally design your world. You can, for instance, choose where to live.
You can buy or build your own home. You can surround yourself voluntarily with better cage mates, which are things like family. friends, competitors, your spouse, and so on.
When you do this and you actually design your environment or design your own cage, you'll find that for at least the hardware side of things, thinking clearly becomes way less a question of being really disciplined and using various structures to think and to guide your thinking and more really just a sense of minor engineering actions.
that you only really ever need to make once, which is when you make the decision to move somewhere or, I don't know, buy a new house or associate with a person and not associate with another person. Okay, and that takes me to the first set of action questions in our course. First, what was that 12 -digit number that I asked you to memorize from earlier?
I actually want you to sit with it for a second and think. What was that 12 -digit number that I asked you to memorize from earlier? Second question is given the experiment on knowledge workers and CO2 slash organic compounds, can you guys just off the top of your head, without us going in depth on hardware optimizations, think of any other biological limiters that you yourself might be facing?
And that's just right now without you having any extremely detailed list of steps to optimize. Anything come to mind? Odds are that thing is probably pretty important to fix.
Okay, moving on to willpower. On a similar level, what the hell is willpower? Well, most people think of willpower as a reservoir.
Every time that you use your self -control, they are withdrawing a small amount from their willpower reserve. Like, let's say you didn't hit snooze this morning, even though you really wanted to. Or you didn't check your phone, even though you really wanted to.
Or somebody was rude to you, but you kind of buried it and were polite instead. The way that most people conceptualize willpower is that it consumes a little bit of willpower, a unit here, a unit there.
And then eventually what happens is at the end of the day, your willpower reserve is low, maybe because you've had a, you know, a lot of temptation fighting all day. And then because it's low, you know, if another temptation comes around or something happens that requires self -control, maybe you're more likely to give up.
You know, and at some point you refill your willpower reservoir with sleep and then just refill or redo the process tomorrow. Now what's interesting is, have you ever noticed how some people are capable of staying extremely clear -minded and self -controlled, whereas other people cannot? Why is this?
Do you think it's because they have more willpower? Well, it's not really. The answer, which has been backed by a tremendous amount of science at this point, is not that some people have more willpower than others.
It is that the people that you think have more willpower simply never have to fight the same level of temptation that you do in the first place. What I mean by this is people that maintain clear thinking and strong self -control usually just do not put themselves in situations where they have the opportunity to make a mistake at all.
If you think about it logically, if you cannot make mistakes at all because that is not one of the possible moves available to you, like in a video game or a chessboard, well, newsflash, you're probably going to make fewer mistakes, right? Because you simply can't make the move that leads to a mistake. Think bowling with the guardrails up.
If you bowl with the guardrails up, landing in the gutters is simply not one of the possible mistakes. I mean, I guess it is technically a possible mistake, but damn, you'd have to screw up really hard to make it in the gutter.
What happens when you bowl with the guardrails up is if you bowl and you're just a little bit off to a side, you knock off a guardrail and eventually end up virtually somewhere in the neighborhood of hitting the pin that you want. Now, in a similar vein, it turns out the majority of the effort involved in resisting temptation or avoiding work that you have to do or some sort of similar issue like that.
really just boils down to you having to adjust to do work at times that are maybe less effective or profitable for you. To do things outside of a regular schedule. To do things at times that are closer to the end of the day, when your quote -unquote reserves have fallen low, than at the beginning of the day, when your quote -unquote reserves are still high.
I want to give you guys an example of this. A person that doesn't check their phone on the meeting, for instance, would commonly be referred to as resisting temptation. And that's good for them, right?
But if you think about it logically, a lot of people don't check their phones during meetings. Do you know how most people do it? They just don't have their phones on them in the meeting.
The people that don't check their phones in the meetings that have them in their bags constantly have to fight the temptation of pulling out their phone to check a text message or write something down. Whereas the people that simply do not have phones on them whatsoever, well, they never have to fight that temptation. Their phone is, you know, three doors down.
It's impossible to get it without making a total ass of yourself. Basically, the people... that don't check their phone in this example because they don't have their phone on them.
They have adjusted and built their environment such that temptation never shows up in the first place. And this is a very small example of much larger sweeping lifestyle changes that I'm going to run through with you. Now to be clear, this is probably not what is going on in your brain under the hood.
There probably is not actually a reservoir of willpower. It's not like you're using up energy every time you say no to something. As mentioned, we use approximately the same amount of energy.
when staring out the window or, you know, working on a task. But I want to introduce a concept to you today called asymmetric bets. If you assume that your willpower is limited and then acted according to that philosophy, that would probably lead you to reducing temptations more, right?
So even if you are wrong, you will have built yourself a better environment and probably some extra capacity. Now, similarly, if you assume that your willpower was unlimited, and that you could just rely on it every time because you're a badass MF, right?
Well, you will fail in front of everybody at the worst possible time, simply because that is the modus operandi with which you have lived your life. So logically then, if you have possibility to believe one thing or the other thing, and one thing leads to a universe where you are better off, aka you resist temptation less because you just do not have the opportunity to resist temptation.
and the other one you resist temptation more, which leads you to screwing up every now and then, shouldn't you believe in the story that results in the universe where you do not have to resist temptation at all? Obviously, right? This is a concept known as instrumental rationality, which we'll cover a little bit later.
But basically, it is where you choose the most effective and useful belief system to achieve a goal, even if it slightly differs from reality. The takeaway is, do not build resisting into any of the plans that you come up with. If one of the steps in your proposed plan, which we'll be doing a lot of later on, is, you know, I just won't do that bad thing that I always do.
Well, it's not a plan. It is a hope, and it is quite unrealistic at that. Instead of eliminating X completely, what you should aim to do is just reduce X by reducing your access and opportunity to abuse X.
If you remove some sort of temptation entirely, it will naturally go down for free and you will think more clearly as a result. Now, in terms of an actual, immediately actionable example that you guys can do right now while watching this course to gain an additional 10 to 20 % out of it, you guys probably have your phone in the room right now.
And some of you may actually be listening or watching to this course on their phones. Now, by doing so, you are increasing the probability of picking up your phone for whatever reason and getting distracted by an order of magnitude, which is 10x or more. What I'd recommend you do right now is...
if you are watching me on the computer and your phone is somewhere in your periphery, I'd recommend you actually stop, go grab your phone, pick it up, move it to another room, and then place it somewhere where it isn't directly in your periphery. And if you are listening to me while on your phone, because this is your only access to media, then I would recommend you swipe down from the top, go to the do not disturb button or the equivalent, depends on your operating system, and turn that on.
If you just do that one thing right now, if you pre -commit, you will get 10 to 20 % more out of this course. Imagine if I had two options available to you, kind of a red pill or blue pill, like in the matrix.
And on one hand, I had a dollar. And on another hand, I had a dollar and 10 cents. And I said, hey, brother, which one do you want?
Do you want the dollar? Or do you want the dollar and 10 cents? What would the logical thing be to do in that scenario?
It would obviously be to be like, hey, give me the freaking dollar and 10 cents, obviously. Well, that is what I'm offering you right now. You have the ability to design the future few hours that you will be listening to this and to maximize your engagement and minimize your distractibility during that time period.
And so this is a great example of a bunch of techniques that we're going to line up back to back in order to help you think more clearly and get more done. But I'd recommend that you do so right now. Because if you do so, you won't be serendipitously distracted by like a rogue notification from your breeds.
You won't be randomly distracted by a friend of yours texting you a funny meme on Instagram, which you have because the damn default notifications always blow up your phone. You won't have a little buzzing constantly occurring in the background, which pulls 5 or 6 % of your content capacity away from you, like we'll learn about in the hardware optimization section.
Can you think of a freer win? You're literally picking up a 100 gram object, moving it 10 meters to the right, and then immediately being 10 or 20 % more effective. Okay, another good example of this, which my family would probably hate, is that I used to smoke weed.
mostly in social settings, luckily. And I did not really like what weed did to me. You know, I grew up in a British Columbia, you know, Vancouver.
So I was smoking that BC Bud for anybody in the know. And I grew up in East Van. So there was a lot of it, an abundance, let's say, in my periphery.
So I tried many times after, you know, picking up the habit with my friends at parties and stuff like that to just quit many, many, many times. But I didn't really know half of what I know now. And back then, I tried to quit every time based on my willpower.
And not shockingly, every one of my attempts failed. You know, I would do pretty well for like a week or two saying no and no and no. But then finally, one random night, I'd go with my friends and maybe I'd be a little bit sleep deprived.
Maybe I wouldn't have been entirely at my best state. And somebody would have some weed on them and they'd offer some to me. And then next thing you know, I was smoking.
So eventually, I identified a rule. And the rule was when I hung out with person X or person Y, I would lead to smoking weed. And then when I did not hang out with person X or person Y, I would not.
So logically, me being a, you know, I don't know, bright 18 or 19 year old kid thought, hmm, I wonder, can I use this rule not to smoke weed anymore? So what I did is I just stopped hanging out with the people that smoke weed. And miraculously, I stopped smoking weed.
And it's been a hot minute quite a while since the last time I did. If you think about it. This is a very simple and minor example of a deep part of me not wanting to engage in a habit that I'd picked up just by happenstance that I then felt chained to for quite a long time.
We all have the ability to do so. We simply need to stop and observe and notice, which I'll cover in the software section, various things in our periphery that tend to lead to actions that we do not want. The second that you establish an if -then rule, everything becomes very, very easy.
So the key pattern is, I put the work in before I got tempted. not during the temptation. I could have come up with a variety of things to say.
Maybe like I tell my friends, no, not tonight, guys. I got to wake up early tomorrow. But that works far less effectively than just building a rule in to eliminate the possibility of temptation in the first place.
It sucks that we stopped hanging out, by the way. You know, some of these people were pretty cool, but obviously my life is way better off. And I think that's a core thing to understand.
Like, do you want your life to be better off? Or do you want to be dragged down to the level of everybody else? I don't necessarily want to do things like that.
Maybe if you zoom out on a deep level, I don't want to. Maybe on a superficial level when I'm tempted and so on and so forth, I do. But I obviously like the deep version of myself more and that's ultimately who I am.
I'm looking to better and self -actualize. So if you find yourselves in similar situations, that's a very simple and easy hack to do. Okay, let's think of some action questions.
Think of the last time this week that you lost to some sort of temptation. Maybe it is your phone. Maybe it is your food.
Maybe it is skipping your workout. Whatever it is for your particular challenges. Was there a way that you think you could have reorganized your environment to avoid you doing the thing that you did not want to do?
I'll give you guys an example. I had some junk food in the house because my girlfriend and I bought a bunch of these wonderful creamy all -dressed potato chips from, I think, Miss Vicky's. And these things are incredible.
They are so good. But one of my goals was not to eat these sorts of extremely ultra -processed potato chips. Well, anyway, we had bought it earlier on in the week because we were celebrating something.
I just got hungry one day. I opened up the cabinet and there it was. I ate it.
If I had not had that at home, I obviously logically would not have eaten that bag of chips, right? I would not have just downed 1 ,200 calories worth of probably the most ultra -processed fat known to man. Now, listen, I'm not saying you have to, you know, live your life like a freaking ascetic.
You really don't. You can enjoy your life and have fun. And it's not the first time I had a crazy bag of chips, not the last time.
But the end result is I ended up doing something I didn't want to do, not because of intention. simply because of my environment. So why not only do the things that I actually want to do?
And I have the power to do the things that I want to do. If I did want to have that bag of chips, I could have reconstructed my environment such that I got them when I wanted them, but I did not. So at some level, a dumber, younger version of me, or somebody else, the manufacturer of the chip, and so on and so forth, essentially made me do something that I did not entirely want to do.
So think about an example like that from your own life. Do you have any? If so, consider them now.
Next, find some step in your plans that depends on you resisting something in the moment. Rewrite that step so nothing needs resisting at all. Maybe instead of you, I don't know, coming home from work or something and then having a big fat candy bar, remove the ability that you have to do that in the first place.
So let's talk phones. What is a phone, really? Well, initially, it was a tool used for making phone calls.
Then it expanded into a tool used for both messaging and making phone calls. But eventually, it exploded into a basically general -purpose computer that could do everything from check your GPS route to video conference your friends, to take photos and post them, and more. How about social media?
Social media initially started as a simple tool for messaging your friends. Think things like AOL chat rooms, IRC, and so on. Eventually, in the pursuit of increased functionality, it became a tool used for both messaging and then posting updates.
Think things like MySpace with like your original pages. Eventually, like everything else, it exploded into an everything app where you can play games and you can even buy furniture. Have you ever asked yourself the question, who came up with all of that and why?
Well, the unfortunate reality is virtually all of the decisions around your virtual environment and your physical environment were made by other people. for the express purpose of consuming more of your time, typically in an unproductive way. Think about it logically.
You don't need to be able to play Farm Simulator 2000 on Facebook, do you? No. So why does it exist?
Well, it's because there was some smart person on Facebook's team that noticed that having games embedded in their little message box increases the average user time spent by, you know, 10 % or 17 % or 25 % or something. If you think about it, the vast majority of the time that you do something on the internet and really life in general, it is a decision that somebody made for you.
This is called choice architecture. Now, this term was coined in a book called Nudge. Really good book if you guys haven't heard of it before.
And it basically means the way that you present options or the way that options are presented to you influences which options are chosen. Put another way. There is no single way to neutrally present more than one option.
Anytime that you give somebody a choice between X or Y or A or B or C, in some way, shape, or form, you will always be influencing their behavior logically. A good example of this is a cafeteria manager. A cafeteria manager can actually fundamentally change what people eat, not by taking anything off the menu entirely.
but simply by changing the way in which the food is presented or the story is shaped. For instance, if you have two items, a healthy bowl and then an unhealthy burrito. If a photo of the healthy bowl looks like crap and the photo of the unhealthy bowl, unhealthy burrito is high quality and attractive.
People will eat way more of the unhealthy burrito. And if you think about it, that's a massive responsibility on behalf of the cafeteria manager. Because over 10 years or 20 years or 30 years or even 50 years, that single action could theoretically lead to hundreds of times more heart attacks, strokes, deaths, and so on and so forth.
If this is a really busy cafeteria with several hundred people visiting every day and you've nudged them even 30 % more towards a very unhealthy food option or something like that, obviously people are going to eat way more of it and as a result, screw their bodies up. And that obviously has consequences, right? What I mean to say here is that you are being manipulated constantly.
Even the algorithm that led you to my video here today was in part some sort of manipulation. Now, in rare cases, like the one that you guys currently have, manipulations can be positive and they can end up improving your life. But that is the exception and that is not the rule.
The vast majority of the time, the 99 .999 % of the time, manipulations are actively detracting from your ability to think clearly. Because your ability to think clearly is tied to your ability to set intentions and then reason through those intentions. But if you're being led not by your own intentions, but by somebody else's, obviously you're not being led to your own ends, you're being led to someone others.
A great example of this from like a sociological perspective is organ donation. Some European countries run opt -in systems where you're not a donor unless you actually register to become a donor. And then other ones run opt -out systems, which are where you are automatically a donor unless you register out of it.
Now, they compared the two groups of countries back in 2003 and found an insane difference in total organ essentially defaults. Now, in Germany, which is an opt -in country, only 12 % of people were registered as donors. In Austria, which is right next door and which was an opt -out region, 99 .98 % of people registered to be an organ donor.
Every single opt -in country was around 28 % or less, whereas every single opt -out country, was around 85 % or more. If you think about it, that is one of the clearest examples of how impactful choice architecture can be.
These governments simply made a choice really early on. Should we default people into organ collection? Or should we allow people to make the choice about whether they want their organs collected or not?
You know, basically opt in. The countries that said opt in have virtually no organ collection. The countries that say opt out have virtually all of the organ collection.
And it is a hell of a power law difference here, which I'll cover later on. Now these are societal level choices that are being made for you, and it shows clearly how much a single line on a piece of paper can impact things like your organ availability, or more morosely, what people will do with your body after you die. And that takes me to this concept of default.
If you think about it, autoplay is a default that is selected and defaulted to because it increases your watch time. Notifications on your phone are a default because they are selected to increase the total number of times that you open an app. Pre -selecting an annual plan when you are shopping around a SaaS.
That is done to increase the total amount of money that you spend and your ultimate retention or churn. And putting candy by a checkout is done to maximize the number of people who ultimately impulse buy because they're saying, well, I already got all that healthy stuff. Why don't I just throw in some candy too?
It is not illegal right now to do this. It's just designing an environment in the service of a KPI. The issue is...
It's just not your KPI. Now, I don't mention any of this because I want to freak you the hell out or mislead you. I just want to offer a new lens to look at the world through.
Next time that you're presented with some sort of choice, just ask whose interests are being served by this choice. You know, if you have a choice between A or B, is that a choice you need to make at all? When you get a phone and all of your notifications are automatically defaulted to on, or you get a new Windows PC and you have a bunch of game apps that are auto -pre -installed, why are they auto -pre -installed?
Why is all of that stuff set up the way that it is to begin with? It's not like phones evolved. There's something that a person put together.
Same thing with Windows PCs. The reality is there's a complicated, interconnected web of incentives at play. But the harsh reality is most of those incentives are not for you.
They're against you. Now, you can use this knowledge to design your own life. As mentioned, you guys can build defaults yourselves that overwrite the typical defaults of corporations and big organizations.
And in doing so, you can have a lot more control over what it is that you do on a day -to -day basis. The math is very much in your favor, and it's another clear example of leverage, if you think about it. Because you only ever have to make the decision one time, while very clear -headed.
Maybe on a Sunday morning after you've had your cup of coffee. And despite you making that decision once, it will continue to pay dividends even at every other 6 a .m. morning when you don't get a lot of sleep, you're tired, you're lazy, maybe you're a little less caffeinated, whatever.
The point that I'm making is you can make a default decision right now when you're in a strong state of mind and then force all future instances of you to do the default decision and basically just treat that as how things are. Okay, so if a default is set by a company and you do nothing, you get what they want.
But if you set a default and you do nothing, you get what you want. So maybe your laptop automatically opens the tab that you were on yesterday. And usually before you go to bed, you're on some sort of entertainment website or social media.
What that means is the default is when you go to your laptop in the morning, you are now seeing a bunch of social media. If there were two universes, one where you had that tab open, and then another where you had a tab open with your work on it, in which of the two universes do you think you would do more work?
Obviously the one where your tab is open with the work on it, right? You have the ability to control that. Tabs reopening are not a default setting.
That is something that the people that designed these browsers did because they want to keep you on them for longer. Likewise, maybe your refrigerator is organized so that the worst snacks are at eye level. And so every time you open it, they're literally the very first thing that you see.
Now, that is a default that you yourself probably unknowingly created, but it's also a default that you can change. How about your calendar, your virtual environment? Maybe it's bookable 24 -7.
And the reason is you just never set your working hours. And that results in a variety of annoying -ass notifications that come in throughout the day and a bunch of people that come in and steal your time. Well, you have the ability to control that and change that right now.
So all these are defaults. They take you maybe a few moments. But the cool benefit here is if you change them, you get repeated gains literally every day until you die or until the company claws back that default.
How do you actually identify all the defaults in your life? Since obviously this is a okay list, but it's not an exhaustive one. Well, I use the following diagnostic tool to do this.
For every transition in my day, I just ask myself, what happens if I do nothing? So for instance, at 5 a .m. in the morning when I access my computer, What happens by default if I do nothing?
So a good example of, you know, environment design is what I have done to my computer. At 5 a .m. when I access my computer, I open it up.
There are no additional applications on my screen except for Chrome. So I will double click on Chrome. I then look at my bookmarks bar, which includes links to all of the work laid out in serial capacity.
Literally one, two, three, four, five. I then click one and then it opens the page that I have to start doing my work on. If I try and access any sites outside of these, my browser will actually block them until 3 p .m.
So what I mean by this is that that's a transition state from my day, right? I went from waking to working. What happens by default if I do nothing in this situation?
Well, I work. I physically can't not work. I basically have, if you think about it, gone onto like some sort of or something.
And, you know, I've gone up to like the slide and I've entered the slide and... The slide takes me to the end destination and there's nothing I can do about it. The cool thing is I'm the one that architected that slide and I'm the one that made it how I wanted it to be.
As opposed to entering somebody else's slide, which maybe had a lot of entertainment on it or distractions or things that reduced my capacity for attention, working memory or executive function, I got to do all of this myself. Similarly, since my phone is not in my room, you know, if I do nothing, I cannot check my phone easily without physically getting up and going somewhere else.
Which, if you think about it, means that my default if I do nothing is I have very low usage of my phone. This graph here is quite important because if you notice that when you do nothing, you achieve an outcome that is beneficial for somebody else, like a company, odds are that default is something that you need to change.
Now, if you do nothing and then you achieve a positive outcome that you like, odds are you have in some way, shape, or form shaped that default, and that is okay. So some action questions. Look at the last 24 hours and list every point at which something happened because you did nothing.
For instance, what did you open? What did you eat? What did you check?
Ask yourself, which of those defaults did you set? And which ones did a company or an organization that is incentivized to make money off of you and your time set instead? Next, pick the single default that costs you the most each week and write down the one time change that would change that.
When will you make that change? By the way, I'll run you through a bunch of great ways to make changes in a moment. Okay, one final point before we begin the optimization of our hardware.
The golden rule of this whole course, if you think about it on a meta level, is friction. I've been dancing around that word for the last maybe 30 minutes or so. I say 15 here, but it took a little longer than I was expecting.
But it's worth us clearly defining it before we move on. The way I define friction is it's a cost between a person and a behavior. This typically means an extra click.
It means an extra step, means an extra 10 -foot walk down a hallway, means having to put in an extra password, and so on and so forth. A great example of this is if you, a person, wants to go to the gym, which is a behavior, friction is anything that makes it harder for you to go to the gym. For instance, your distance from the gym, like where you live versus where the gym is, that is friction.
If you go to sign up to the gym, but the sign -up portal bugs out, that is friction. If there is a physical cost, which is a little bit higher than you were anticipating, that is friction. In general, if you want to decrease a behavior, what you do is you add friction.
If you want to increase a behavior, what you do is you remove friction. Because your brain, as mentioned, is extraordinarily energy conservative, and it wants to conserve energy wherever possible because it consumes 20 % of the entire body's energy reserves, it will do the rest. You simply need to make something a little bit harder.
and make the other thing, presumably the thing you want to do, a little bit easier, and your environment will start shaping your choices without you even noticing. The most important point about friction is that it is proportionate. For instance, the size of the friction scales typically with the size of your opportunity to engage in the behavior.
So if you make something maybe three times as hard to do, on average you will do it maybe three times less. If going to the gym is now three times easier, because maybe the gym has moved three times closer to you, or maybe because, I don't know, you have a friend that picks you up in the morning and actually drives you and drops you off there like I used to have, then you will do it far more often.
A good example of this is friction in food consumption. If you guys think back to our cafeteria example, where we had a good -looking photo and then a bad -looking photo of a bowl and a burrito, and that impacted the uptake of food. Lots of studies have been done on the exact same idea, except instead, all about applying friction.
What I mean by that is making something easier or making something a little bit harder. So in one of these studies, they changed the distance of three foods in a cafeteria. There was a bowl of cherry tomatoes, there was a bowl of peas, and then there was a bowl of carrots.
Aside from that, they didn't change the signage or move any of the food at all. Here are some results. When they moved the cherry tomatoes 10 inches further away, consumption of all cherry tomatoes dropped 9%.
That means that by moving the cherry tomatoes literally less than one foot further from where they were before, people consumed approximately 9 % less. When the peas were moved 10 inches away, consumption dropped 13%. When the carrots were moved 10 inches away, consumption dropped 9%.
Then the researchers went as far as to replace the serving utensils for the peas from a spoon, which is quite easy to use to grab peas, to tongs, which are hard to use and you can't really get too many. Total consumption of peas dropped 16 .5%. Now, this is a massive effect.
Not to mention, all of this took only a few seconds to actually implement, right? If you guys look at this graph here, you can see food taken on baseline. The spoon at reach, obviously, allowed for baseline consumption.
When you move something 10 inches, it reduced it by 13 .4%, 8 .9%. And then when you replaced your spoon with tongs, it reduced it by 16 .5%. All of this is pretty crazy to me.
Here's a quote from an essay that was written back in 2009 by really smart guy Scott Alexander, and I personally think about this quote all the time. The point is that you don't have to make things impossible to make them not happen. You just have to make them inconvenient.
Even a trivial inconvenience, one that takes a second or two, is enough to stop the majority of people the majority of the time. So what does this mean? Have you guys ever been in the following situation?
where you decided not to open an application because it signed you out? Have you ever not read some sort of report because maybe it was attached as an attachment or PDF rather than in the body of an email or an easy to click link? Or maybe have you ever not worked out because you went to your bag and then found that you just didn't have the stuff inside of the bag or you didn't have your shorts on or something like that?
Realistically, none of these are real barriers. Like would this stuff have stopped Kangas Khan from conquering the known world? No, probably not.
But they all still worked on you. A good model to internalize is that your mind is always calculating the startup cost to do anything in your environment, also known as the friction. So if you add or remove even a second of friction from a process, you can, over a long enough time, change your life completely.
Like we talked about in the first section on the biological basis of cognition, executive function is very, very expensive. Your working memory is also quite small, and you guys only really have enough space to attend to a couple of things in your environment at any one point in time. What that means is your brain, which is always looking for a way to minimize costs, will always default to not now as the most efficient option, as opposed to the let me figure out how to take the first step option.
The crazy thing is from the outside, it all looks like discipline. You know, when I used to be really poor, I would look at people that were making 10k a month. or 20k a month, or 15k a month, or maybe even 100k per month.
And I would always think, you know, these people, they're a different species. They're just built different, and I'll never be able to catch up. I looked at how much time and effort it took me to make $2 ,000, and then I looked at the guy making $10 ,000, and then I went, there's no way.
I mean, that's five times the work. That's impossible. The harsh reality is, those people are not doing five times the work.
In fact, many of them were probably doing half of the work that I was doing, maybe even a quarter of the work that I was doing. The main difference was they had designed their environment to be able to do the things that actually make them money really easily and automatically. Whereas I, I was a door -to -door salesman at the time, had to get up and push through a slog every single day.
Many of the times when I would have to do this, I would not actually carry through with my core foundational revenue generating activity simply because of how hard I had made it for myself. So that takes me to a few action questions. The first is name one behavior that you want to do more of this month.
How many steps currently sit between the thought and then the action? And what single change can cut down the number of steps so that instead of going to the gym requiring 10 things where you got to put your shorts in your bag and get your shoes and get the basketball and walk out the house and take the bus, instead of all of those steps, what changes can you make that would bring that to just one?
The other question is name the behavior that ate the vast majority of your last week. What one step could put you in front of it? that would force a conscious decision every time.
It is now time to actually optimize our hardware. Before we get started, none of this is probably going to seem very fancy to you guys. I mean, if you think about it, we're going to talk about some very elementary and basic things here.
So things like food, sleep, exercise, and so on and so forth. And these are all things that people have been realistically discussing for decades. But if you optimize them in the way that I will show you, most people that are taking this course can expect 40 to 80 % improvements in cognitive function.
And that is not a typo. I mean, we did the math. That is literally 1 .5 extra performance on a wide cognitive battery.
To make it clearer, I'm not a doctor, and none of this is medical advice. What I have done is I've gone through all the currently available up -to -date literature, the meta -analyses, and the randomized controlled trials on various biological interventions. I've also ranked them according to EV, or expected value, which I will teach you guys how to do later, for the median person.
And so if you're unsure about any of this, I would encourage you guys to talk to a qualified doctor or your LLM. With that out of the way, I've seen people spend tens of thousands of dollars on biohacking and cold plunges and crazy supplement stacks and nootropics and all this other stuff, ultimately to eke out maybe 1 -2 % improvement in cognitive capacity.
These same people, when asked, hey bro, how much sleep did you get last night, will confidently turn around and tell you, I don't know, maybe like 5 or 6 hours. If you're one of those people, If you are currently focusing on trendy and super shiny objects, or maybe all these biohacks, without actually having maxed out your health, the biological gold, which if you think about it is available to us all, very low hanging fruit, you're almost certainly wasting your time and I recommend you stop it.
I think about it like winning a minor skirmish and then losing a generational war, or maybe dodging a bullet just to get hit by a nuke. You guys remember the basis of cognition? It was attention, working memory, and then executive function.
What I'm going to do is I'm going to show you guys how to use biological interventions to improve each of these in turn. To begin, I'm going to start by going through the three major drivers of clear thinking, which is sleep, nutrition, and exercise. And we're going to do that one at a time.
For every one of these, I'm going to explain how it works, what you guys can do about it, and then also my own routine. And if something is not backed by hard evidence and it is just a suggestion, I will be sure to tell you because I don't want to mislead anyone or have you do things that are not necessarily scientifically backed.
So this graph shows where the effective sleep, nutrition, and exercise, these tend to be approximately 70 to 80 % of all possible gains, although there are some others that I'll cover too. Unfortunately, the attention currently goes towards flashy things like nootropics, cold plunges, supplement stacks, productivity apps, morning routines, biohacks, all that stuff.
While some of these are important, they're typically more important in concert with the former three. And it's also a good point to remember that the whole point of this is none of these should really require tremendous mental willpower. That's the whole point, right?
My goal is really to get most of these gains done without really adjusting, by only really adjusting your schedule. Things like your light, your diet, your exercise, and some other stuff. And the plan is to make habits that clarify your thinking and turn those into defaults.
And then take those that ruin your thinking and make them, like we saw with the European examples, opt -ins. The format typically is the title with the actional tip plus expected cognitive improvement. Any expected cognitive improvement is just what I've reasoned through first principles, as well as, you know, summarizing the literature and meta -analyses on the subject.
I also want you to know that these are not additive. So you're not just going to add the 10 to 15 % with the 8 % with the 14%. These are all multiplicative.
So think about it not as 10 % plus 10 % equals 20%, but 1 .1 times 1 .1 equals, I don't know, 1 .11 or something. Okay, so the very first is to rule out a sleep disorder. Now, logically, you will sleep every night for the rest of your entire life, right?
Given how often you sleep, doesn't it make sense to ensure that you guys are sleeping well? If you think about it, that's a huge return on investment for just a few moments of your consideration, right? And the unfortunate news is a lot of people haven't spent more than five minutes thinking about something they're literally going to be forced to do for about a third of their life.
So first, before proceeding to optimize anything, because I'm going to give you guys a big laundry list of sleep improvement techniques which have direct and measurable impacts on cognition, just make sure you don't have a sleep disorder. Because if you do, no amount of the next few tips are going to help you given that the problem is mostly on the hardware level and not actually on anything else.
The main one to look for, which unfortunately around 30 % of men in the United States are expected to have, is some form of sleep disordered breathing. This is an umbrella term that includes sleep apnea, as well as its less audible cousin called UARS. As mentioned, 30 % of men in the US are suspected to have one of these two.
And it's simply mostly a byproduct of weight at this point. You know, the more you weigh, the more compression occurs around your trachea, the thing you use to breathe. And as a result, that compression leads you to not being able to breathe fully in the middle of the night when all of your muscles collapse.
As a result, you typically have disordered sleeping. Now, essentially, this triggers the brain to wake you up partially dozens of times per hour to fix the issue. You have to wake up a little bit to take a really deep breath, or you start snoring a lot, which kind of wakes you up as well.
And you don't remember any of it, but it still reduces your sleep effectiveness significantly. A very common thing to do is to spend around eight hours in bed and then only get the equivalent of maybe five hours of actual sleep. The symptoms are also quite vague.
They're often misattributed as personality traits, which I find funny. You know, brain fog, or waking up tired after nine hours in bed, or always being the sleepy person, or the person that loves their coffee, and so on and so forth. All of this stuff is associated with sleep disordered breathing, and it's not something that I would recommend.
The whole reason why I spent a solid two minutes talking about this, by the way, is because I actually have that. And for many years, I attributed my grogginess in the mornings to a lack of discipline and a lack of routine. I also tried a bunch of different sleep hygiene protocols and medications.
None of those worked. Until now. So if you have any of these, snore, wake up tired, breathe through your mouth at night, you have a narrow jaw, maybe you drink an incredible amount of coffee or energy drinks.
I would recommend getting a sleep study and ruling it out. If you find that you have one and you do treat it, not only will you live many additional years, but you will also think far more clearly. You'll also instantaneously, basically, not actually, but almost, become fitter, healthier, more vibrant.
And it also costs you very little in the grand scheme of things. Think about that right now. If you had one and you figured it out, you've just gained 10%.
Like the matrix pill analogy, you will have just taken the $1 .10. Next is to sleep somewhere between 7 to 9 hours for another 10 to 20%. Sleep deprivation is super common.
I would wager more than half of the people that are currently watching this course have some form of sleep deprivation, knowingly or otherwise. The major issue is it's not actually the sleep deprivation itself that's so dangerous, but the fact that it is invisible to us as humans. Because you don't know that you're suffering.
And science has shown this many times. In this example, 48 healthy adults were monitored in a lab for two weeks, 24 -7. and they were restricted to either four, six, or eight hours in bed, so moderate, mild, or not at all sleep -deprived.
Results were, over two weeks, that people in the six -hour group accumulated as much of a sleep deficit, aka impacted performance, as an entire one or two whole nights without sleeping at all. Yet their perceived sleepiness, how they thought they were, did not change after the first few initial days. Which is interesting, because the researchers found that subjects were unable to reliably detect their own impairment.
You might be able to think, hey, you know, I can survive off six hours. No problem. I'm special.
But logically speaking, the very instrument that you would use to determine that is the thing that is ultimately impaired, right? So how the hell are you going to know? You can't.
Our brain is ultimately the thing that we use to think through whether or not we are sleep deprived. Now, there are obviously a few people on Earth with a genetic mutation that allow them to sleep much less on average. They're pretty cool, and you should look into that if you wanted to.
Unfortunately, statistically, you are almost certainly not one of them. So as a sensible person, try and fix this. What's really interesting is, you know, if you are sleeping six hours a night, as this graph shows right over here, you could see that after, I don't know, maybe let's say a week or so, you say that you're sleepier.
You say that you're less sleepy than you actually are. You essentially have more confidence than reality. And I don't mean that to mean not sleeping makes you super confident you should go and do it.
I mean, you think you are more capable than you realistically are. And as a result, you make a lot of strategic and tactical errors that you otherwise would not. I have a couple of really good friends that have probably been sleeping on somewhere between five to six hours for better or worse, three or four years.
And I hate to say it, but I've watched them measurably degrade in their cognitive and tactical capacities throughout that time. Despite repeated urgings, they continuously say things like, dude, it doesn't matter. I can't.
No, man, I'm totally fine. I don't feel bad at all. Well, they're definitely on this side of the graph, if you catch my drift.
Now, worth noting, any deficit here is dose dependent. So six hours, for instance, is worse than eight hours. If you sleep four hours, it's worse than six hours, etc.
My recommendation is 7 to 9. I would consider that non -negotiable if you really are serious about improving this. And you can do this easily, by the way.
All you have to do is set a bedtime alarm and then a wake -up alarm at the same time every evening and morning for maybe a month or so. So let's say you go to bed at 10 and then you wake up at 6. You can actually stop right now and set alarms either on your phone or if you have one of those physical sort of wall clock alarms, do that as well.
The first few times, your mornings and evenings are almost certainly going to be rough. But that's simply because you're not used to it. Over time, you'll begin getting tired right around the time that you're supposed to, and that's just because like any animal, including the rats that were in the labs that I managed, we thrive on regularity.
Another easy 6 -12 % improvement is through improving your wake time. So this is a very powerful sleep tactic, and it also costs you virtually nothing if you think about it from an ROI perspective. So I used to work with a lab that dealt with the neurobiology of sleep.
We changed rat feeding, waking, and light -dark schedules pretty often. And to give you guys a rundown of the main concept there, it's called the circadian clock. Your circadian clock is just a system internal to your body that injects you with various chemicals at various parts of the day.
So cortisol, adrenaline, and other wake up chemicals, basically, usually as you wake up. The idea behind that evolutionarily was you're waking up. You better get to it, right?
You're going to have to go out, find food, fend off saber tooth tigers, do whatever the hell else. So cortisol and adrenaline, naturally, when they're applied at the right times, are extremely valuable, and they're also correlated with major improvements in cognition. I know that all the TikTok girlies are talking about reducing cortisol so their face looks less bloated or whatever, but cortisol is not inherently bad.
Obviously, we have all of the chemicals we have in our body for a reason. So if you've ever had a morning where you just feel on, a lot of those mornings are going to be because you just so happen to align your internal circadian clock with the time at which you are waking up. It's an easy 6 -12 % improvement to the clarity and rigor of your thinking, virtually for free.
Now the issue for most people is that our circadian rhythms, you know, they need a couple of weeks of consistent sleep or wake cycles to lock on to a time. And unfortunately, most people do not actually wake up at the same time for many days in a row. Instead, they will wake up early Monday to Friday, but then they'll ruin it all on Saturday or Sunday on their days off, assuming they work a 9 -5 or something.
The best and easiest way to fix this is simply to pick a consistent wake time and then stick to it every day of the week. That way your body will naturally self -correct the energy boost period to occur right at the time that you wake up, or at least within an hour or so of that time. And this doesn't just improve your short -term or acute cognitive capacity during that hour period, it will improve that throughout the day as well.
The next is to cut caffeine such that you are not taking it 10 hours before bed. When you do this, you can expect a 3 % to 8 % improvement in your sleep, which will obviously roll over to your cognition as well. And it's another very easy hardware gain.
The reality is most people who sleep at 10 p .m. or so will drink coffee after 12 p .m. What I mean by that is since caffeine has a very long half -life, aka how long it stays in somebody's system, doing this will directly and immediately impact the quality of your sleep.
Because logically, caffeine will still be floating around in your body, right? Lots of studies show that caffeine right before sleep impacts how much you actually get, which leads to the same problem that you would otherwise get with sleep deprivation. So in this study, people were given 400 milligrams of caffeine six hours before their sleep.
These folks literally lost a whole hour of measured sleep, and they didn't even know it, meaning that they were, like those people in that study I talked about earlier, accumulating impacts on their cognition, and they were completely unaware of it the entire time. The major rule is if you go to bed at 10 p .m., you should stop caffeine consumption at noon and no later.
Otherwise, you could see that despite the fact that you drink it at 12, if you drink your, I don't know, let's say 200 at about 10 p .m. or so, the average or typical person will still have approximately 50 milligrams in their system, which is about a quarter of that coffee. For those of you guys that haven't had 50 milligrams of caffeine, that is pretty sizable.
And to have that running around your system in the middle of the night, which is the direct opposite sort of thing you want in your system in the middle of the night, mind you, is obviously not very good from a long -term health perspective. Next, get bright mornings and then dim evenings so that they're dark for an additional 3 -8%.
What I mean by that is circadian rhythms drive a lot of our behavior and our performance, right? Well, have you ever asked yourself how these things are set? The reality is they're set via light.
So there are tiny brain regions in your head which are activated by light, which later entrain the rest of your brain. And they also entrain those chemicals that we talked about earlier. So if you want a consistent boost of energy, as well as a far more predictable schedule, you have to get...
consistent morning light exposure. In general, what this looks like is about 10 minutes of natural sunlight within an hour or so of waking up. Even a cloudy day typically has a lot more light than, you know, your kitchen or an office.
And so my usual recommendation is unless you have a really big window or you have some sort of artificial light, which I'll show you guys in a moment, you're going to want to go outside. And at night, you want the exact opposite. A couple of hours before bed, you want some sort of dim or warm or low intensity light.
And that will, you know, reverse that whole cortisol adrenaline thing that I talked about earlier and start bringing it down. A note on blue light, by the way, since I think many of us will probably have heard about blue light from pop science shows or YouTube channels or whatever. The idea of blue light making you unable to sleep.
is probably less of an issue than the actual content on the screen itself. Most of the time, if you're scrolling around really late at night, you're doing so on social media platforms or some sort of feeds that facilitate doom scrolls. And that, if you think about it, is incentivized to show you guys highly engaging and controversial content just to sort of keep you going.
These algorithms are tuned and they know your time zone. So they will, they will do stuff like this on purpose. I don't mean the algorithms themselves are evil.
I just mean like their goal is to keep you on. Right. So they will show you these terrible things in order to keep you awake and keep you scrolling.
Obviously the, there's nobody, uh, there's no Facebook user, like a person that can't fall asleep, uh, until one or 2 AM in the morning. Right. So despite the fact that, you know, I think blue light at night is important.
I think it's a lot less important than people make it out to be. And I think blue light blockers and stuff like that, while they help, they probably don't help in the capacity of, hey, I can wear blue light blockers and keep scrolling through gore and whatever the hell else on my ex account. You definitely should not be doing that regardless of what it is that you are doing at night.
My rule is I just stop looking at socials every night. I do so like a couple hours before I go to bed. And honestly, I try and not look at socials at all because they're just inherently terrible for you.
Anywho. Another quick and easy way to gain somewhere between 2 % to 6 % to your cognitive capacity is to sleep in a cold and a dark environment. So in general, you want your bedroom to be cold, dark, and quiet.
The majority of human beings sleep best at temperatures in the high teens. I'm Canadian, so in Celsius, and I do not know what that is for my freedom -loving brothers down south, unfortunately. But I know for us, it's somewhere between like 18 to 20 degrees Celsius.
Personally, the way that I do it is I have a mattress topper called an 8Sleep. And I'm not affiliated, but these guys are pretty cool. Sort of like water cooled.
And what it does is right before I fall asleep, it reduces the temperature of the bed to get me close to that 18, 19 degree range. And then as I sleep, it modulates and mediates the temperature. So it keeps me basically, according to known science, in like the optimal space of comfort to be able to sleep for longer.
This is not cheap. I think this costs four grand or five grand or something like that. But if you can afford it, I would highly recommend it, again, because of the ROI.
You're doing this every day for the rest of your life. You might as well get as much of an ROI as possible. Now, for those of you guys in maybe warmer climates, you know, I'm up here in Canada, so it gets pretty cold.
You can use an AC, fan, blackout curtains, and so on and so forth. If you're somewhere really north, you might need a heater as well to get to that point. Another really easy hack is just to understand whether or not you have any nutrient deficiencies or anything like that.
Now, I understand this is not the most accessible to a lot of people, especially those in commonwealth areas where maybe health is public. But I would recommend at some point over a course of maybe a year or two, you should probably get some sort of blood test. If you think about it, your brain is fed by blood.
It's the most metabolically active organ per unit mass, around 2 % of the mass and 20 % of the energy. But it stores very little energy of its own, meaning that the quality of your blood is a very important part of your nutrition. Unfortunately, because our diets have all gotten very efficient.
we eat ultra -processed foods, or we eat food of which natural minerals and so on and so forth have been stripped from, a lot of people end up deficient in one or more of these nutrients, which means they're sacrificing breadth of diversity of nutrients for the convenience of maybe the same thing every day. I'm guilty of this as well, by the way.
If you have a deficiency, a lot of the time it will manifest as things that look like, you know, a laziness or a lack of focus or unclear thinking, etc., often to the tune of 10 to 15 % or more if you have a nutrient deficiency. You know, as mentioned, I'm not a doctor and I can't give you clear medical advice on this.
So do this if this sounds like it may or may not be something that you are experiencing. But some very common ones that a lot of people have are things like iron, especially in women, ferritin, B12, especially if you don't eat a lot of meat or on a vegetarian or vegan diet, a vitamin D, if you live in Canada or other northern countries, northern latitudes, and it's between October and April, you'll probably have this unless you supplement, you know, underactive thyroid problems and so on and so forth.
These are all things that are worth at least thinking about and then cluing into. to see whether or not there's something that you might realistically have. Because all of them essentially unlock you a good 5 or 10%.
I would also get some form of metabolic marker, just to see how your blood sugar is doing. A lot of people probably eat these ultra -processed foods and then result in chronically high blood sugar that leads to clarity of thought issues long before you develop things like diabetes and so on and so forth. And yeah, you guys can do this through a doctor's referral if you're in a public healthcare place like I am, or you can pay out of pocket for insurance if you're somewhere private, like the US or other places.
The harsh reality is this is probably going to be a few hundred dollars or maybe a couple hours of your time max, but it will pay dividends for the rest of your life as you get proactive about managing your health. Okay, so I have a standard panel over here that I would recommend you guys get at some point in time. All of this is super easy to see.
Just get the results, then you can take them to, I don't know, your own LLM. You can take them to any doctor that you want to discuss these things, people that you trust, and so on and so forth for, as mentioned, proactive health purposes. Another big thing I do is I cut alcohol.
People that do typically see somewhere between a 4 % to a 10 % improvement on a variety of these batteries. If you're going to get anything done, you might as well finish with booze. And I don't say you have to abstain completely.
I just mean if you want the best performance and you really want to maximize the clarity of your thinking, alcohol just is not for you. I want you to think about this from the perspective that I talked earlier. Everybody's like, oh, just have a drink.
Why not? We didn't evolve with alcohol. it's not like alcohol like alcohol was created you know what i mean obviously it was created through a bunch of natural fermentation processes that led people to you know smelling a bunch of very off smelling barley or whatever and then thinking oh maybe i'll put this in my mouth for whatever reason prehistoric people man but the point that i'm making is this is a default that was given to you the whole idea of like going out to the bar or whatever that was a default that was given to you You can certainly enjoy it and you can like it.
And if you find that you like it, you can turn it into something with intention. But so many people live lives of no intention whatsoever or they think they have to simply because of the people that they're surrounded with or the environments that they happen to grow up with. Take the permission from me.
You don't have to do any of the stuff if you really don't want to. Most people don't actually even enjoy the taste of booze. It's one of those hidden secrets that nobody really likes talking about.
So anybody that's young that takes a sip of, I don't know, some super hard liquor and goes. That smells and tastes like crap. You're not wrong.
It does. It's just people tend to grow used to it and then eventually find joy in the suffering. Now, the main issue for us cognition interested types is not actually like things like liver disease and so on and so forth.
What it is, is it's sleep. Alcohol lets you fall asleep faster, which makes sense since it's a central nervous system depressant, right? But it does so at the cost of actually disrupting the second half of your night, usually, which is where you're supposed to be in what's called REM sleep or rapid eye movement sleep.
This is critical for making you feel more rested the next morning. And since alcohol disrupts that, it interferes with your ability to feel rested and then also consolidate things like memory and so on and so forth. Basically take things that were in your working memory during the day and then turn those and encode those into memories.
So even one or two drinks can massively impact your next day. I personally keep my alcohol consumption at zero. In the last eight years, I've probably had less than a dozen drinks and virtually every one of those was at some...
quote unquote, once in a lifetime opportunity, say half a beer at my best friend's wedding or something of that nature. And despite that being relatively small in the grand scheme of things, I still feel it the next morning. Okay.
Like alcohol, I'd highly recommend that you guys cut any sort of drug that may or may not be something that has been medically prescribed to you. A good example of that is cannabis or weed for two to 6%. So if you're an active cannabis user, it's a very similar problem to alcohol, although it works through slightly different mechanisms.
It can help people fall asleep faster, at least feel like they're falling asleep faster, but a lot of the time it does this at the expense of sleep quality. When I stopped smoking, I felt almost immediate improvements in just mental clarity, maybe two or three days, and this is an advantage I've never once let go of since then.
Also, if you're at the point where maybe you need this to fall asleep, or at least you think you do, you really should urgently fix this, because relying on any sort of drug to sleep is terrible, it's obviously bad, and it's not where you want to go. Eating more protein and fiber for maybe 2 -4 % each is extremely important.
Protein helps stabilize blood glucose levels as well as prolongs feelings of fullness, which is a side effect and can help you lose weight if you guys are overweight. Also beneficial to cognition and then also has a tie -in to the sleep -disordered breathing loop. Fiber slows down the absorption of sugar from other foods and also feeds your gut microbiome, which is strongly associated with mood and cognitive function.
And while all of this is very early on, I mean, the research that I'm talking about here has really only come out in the last couple of years, it is not insignificant. And there are a lot of suggestions that, you know, some major depressive disorders or, you know, a variety of like mental or disordered states of thinking and so on and so forth are a result of a fractured gut microbiome.
So you can fix that. Now, I used to blow at fiber. I would eat maybe 5 to 10 grams on a good day.
And just so you know, the recommended intake for somebody about my size is probably closer to 40 grams or so. When I corrected that by eating more fiber, I noticed an immediate improvement in sustained energy levels throughout the day, as well as just my mood on a day -to -day basis. The next is to take creatine for around a 1 to 3 % improvement in cognition.
More if you are sleep deprived, because creatine acts as a buffer for sleep deprivation. Now creatine is really the only supplement that has any real evidence support in the literature. They've done lots of systematic reviews of randomized trials, and it shows that Creatine may improve reasoning in short -term memory, specifically in people that do not have lots of creatine floating around in their bodies.
Creatine comes from meat, typically red meat, so steaks and very fatty sort of gamey meats as well, meaning vegetarians, vegans, and so on and so forth, or people that have more pescatarian -style diets typically do not get as much creatine. So when you supplement with creatine in that micropopulation, you typically have a more pronounced result than if you supplement with creatine in, let's say, a population that eats a lot of beef, but it's still present.
It also seems to impact stress management for the better, which I think is important. Now, if you're going to take a supplement, and again, supplements are such a small portion of the total productivity and clarity of thought pie that I just want to make it clear that like they're good, but you shouldn't rely on them. You should probably take creatine.
I'd recommend three to five grams a day. costs around 20 cents or so, and there's virtually zero risk. Not completely zero risk.
Obviously, if you are concerned, talk about it with your doctor. But if you lift weights or have had a poor night's sleep, all of this is a no -brainer. It's basically as close to the limitless pill as you'll ever get, which is disappointing, but, you know, still pretty good.
Now, for a massive upgrade, raise your VO2 max and get 10 to 20 % better cognitive capacity. Now, exercise improves cognitive performance in a lot of ways. Cardio improves blood flow to the brain.
It increases expression of what's called BDNF, which is a brain -derived chemical promoting neurogenesis, and the hippocampus, which is a region of the brain involved in memory. It also improves your insulin sensitivity and resistance, and then it improves sleep as well, which we've talked about. Now, we can't directly measure all of these effects in humans clearly, since it is way easier to figure all this stuff out if you can just euthanize the animal and chop its brain in half.
Scientists have been seeing this sort of thing in rodents for many decades, and there's also very strong evidence that all of this is directly causative, meaning that there is a if -then loop behind exercise and then sustained cognitive performance and focus. If you literally just give animals access to a running wheel, and then they tend to run more because it's in their environment, they get more neurons created in the memory centers of the brain.
They also learn mazes much faster than sedentary, which means controls that don't move a lot. So if you're going to track just one part of your exercise, my recommendation is to track your VO2 max.
And this is something that I'm currently in the process of doing, having started from relatively low VO2 max due to both some biological disadvantages I was given and then just a lack of cardiovascular exercise. And I'm seeing massive improvements. It's also the gold standard for measuring cardiovascular fitness.
And it's also, interestingly enough, a much better predictor of longevity than almost any other risk factor. So for cognitive performance, it's a good proxy for your brain's ability to receive and then metabolize oxygen, because the whole idea behind VO2max is how much oxygen you can actually take in per unit time. Given that, as we know, your brain is the most oxygen -demanding and energy -demanding organ of the body, it kind of makes sense that becoming better at the process of taking in energy would improve your brain's ability to function, right?
It's also a really good KPI, so you can measure this in a lab once every quarter, costs a few hundred bucks, or you can use an approximate estimate for maybe a smartwatch, a Wootband, an Apple Watch, etc. Or you can also use a formula. There are a variety of them, like the Cooper equation, which helps you understand the direction of your health over time.
Now another 5 -10 % improvement here in using the ReHIT protocol. ReHIT is a way to improve your VO2 max in surprisingly little time. Now because this is technically linked to the VO2max, I'm not saying that these two stack necessarily, but this is a cheap and easy way for you to improve that VO2max marker for virtually no effort.
The way it works is, REHIT, which stands for Reduced Exertion High Intensity Interval Training, is based off a study where a bunch of sedentary adults were given a bike and asked to pedal for three 10 -minute sessions per week. Twice per session, they pedaled as hard as they could for 10 to 20 seconds, in total exerting maximum capacity for 40 to 60 seconds per 10 -minute session.
After six weeks, their VO2 max improved by 15 % in men and 12 % in women. We also saw an improvement in insulin sensitivity of 28 % in men. To be clear, these people did just one minute of hard work every time they worked out.
The other nine minutes... was easy, slow pedaling, which means that if you do the math on three times a week times six weeks, they spent 18 minutes in total working very hard over six weeks to improve their VO2 max by 15%. Every session essentially improved their VO2 max by almost 1%.
Keeping in mind that these were sedentary adults, aka they probably had poor VO2 maxes to begin with, if you are young, healthy, and somewhat fit, you're unlikely to get 1 % improvement to VO2 max per re -hit protocol, and obviously you have a massive decrease in the impact of this marginally as you do more, but it's still insane.
So I'm starting to do this, and I have been doing this now for maybe a few weeks in total, because I've been traveling. This is usually two 20 -second sprints. I will do this just after every workout.
So my protocol is a little bit different. But if you think about the total capacity and total impact of this, it's now 40 seconds, six days a week. You know, it is essentially equivalent to that same protocol in terms of total number of minutes of like extremely difficult work.
I'd recommend you just tie this into like after your cardio, after your exercise, if you're doing strength training or something like that. Just do not allow yourself the decision about whether or not to do this. Just do it.
I mean, a 1 % improvement in VO2 max every workout for at least maybe the first six to eight weeks is probably the number one thing you can do for your cognition right now, aside from taking care of, you know, major health issues and sleep disordered breathing and nutrient deficiencies. Finally, I'd recommend you stack your defaults.
What I mean by this is when you lock in even a few of these habits, you guys can eventually stack them together for far more efficiency. For instance, instead of just working out in a gym, which has limited light and which sometimes will make you feel pressured to do this in the afternoons or the evenings since it's an artificial environment, why don't you exercise outside in the mornings?
This would give you immediate exposure to bright light, should knock that out. You'd have a consistent wake -up time as well, should knock that out. And it would also give you, of course, your exercise, so you'd knock that out.
Even better if you pair it with a friend, because now you also have the equivalent of a cage mate, which is a very important part of socialization and cognitive improvement. Alright, let's run through some action questions before talking about workspace design. Which of these three levers, sleep, nutrition, or exercise, is most broken for you right now?
And what is the one probably pretty boring test, maybe a home sleep study or a blood panel or a VO2 max measurement, that you can do this week to find out for sure? Next, add up all of the percentages from your personal hardware low -hanging fruit. What gains would you say you are currently leaving on the table?
And to be clear, I don't mean add as in the literal sense, I mean multiply them all together. All right, let's now talk about workspace design. You have an extremely low -hanging fruit by doing things like managing, as mentioned, your air quality, by managing your light intake, and so on and so forth.
And I wanted to give this section the time and energy that it deserved because I think most people are currently operating in suboptimal environments. Now, a common part of your environment that most people just do not think about for the life of them is their air. Our metabolism currently runs on oxygen and output CO2.
That may change when we get accepted into the organism, but for now, we run on oxygen and output CO2, and so you should live with that in mind. What that means is, when you're in a closed room, you are actively consuming all of the O2 in your room, and then you're pumping out, in response, the CO2. If your room ventilation is very poor, over time, what will occur is the balance of O2 and CO2 in that room will shift.
And many studies have shown that a poor O2 or CO2 ratio starts negatively impacting cognition right around when CO2 gets to about 800 parts per million. As a rule of thumb, air outside, like outside of your window, is 420 parts per million approximately. Whereas the air that you exhale is around 40 ,000 parts per million.
That means that in a small and relatively closed office, one that does not have any outdoor ventilation, CO2 levels can climb to as much as over 1 ,000 parts per million in just an hour or two, unless you get some fresh air in there, especially when there are multiple people. Now, the worst thing is we don't notice this.
Lots of studies have been performed on your ability to sense minor performance degradations like this, and the brain just virtually cannot. So your brain thinks this is normal. You don't have a headache.
You don't have any obvious sign that you are just slowly becoming a worse version of yourself. I like to think of it as . It's just now instead of water, it is CO2.
And you can actually see performance here. If this is your baseline reference at about 550 parts per million, at some hypothetical office, this is when Allen et al. in 2016 was studying a big group of office workers.
When CO2 went up to 945 parts per million, it's almost double, performance went down 15%. When it was at 1 ,400 parts per million, where many people that are currently listening to this right now are currently living and existing, performance went down 50%. There are way too many people listening to this course right now that are probably in highly CO2 -ed out environments that are significantly impacting their ability to perform certain cognitive function scores.
Obviously, to be clear, a cognitive function score does not mean everything that you do. It's not going to impact your ability to, I don't know, let's say like... pay attention to things or whatever, hear things, but it will impact your ability to think multiple levels deeper than you probably would assume was possible in a CO2 environment.
So all this aside, you should get the air moving. My recommendation is to buy a $200 CO2 monitor, which sounds like a lot. And I'm going to do a product placement here.
I wasn't paid for this, but you know, this is what I have. And what this does is it says more or less what the current CO2 levels in the room are. And you guys could see mine are currently a little too high.
I have 763 parts per million of CO2. And if I go any further, I'm going to have a little sad face on here. Now, this is pretty expensive.
You don't have to get the specific one. I think this one was the air things or something like that. But there are a variety of different ones that you can get.
As long as it measures CO2 in parts per million, you know, do a quick little Google review or ask your agent, you're probably fine. Alternatively, just always have a little window cracked open to the outside world. Like a fan is not enough because you're just recirculating the air.
So you can also prop your door open between work periods. That's something that I used to do when I would work in a closet. Thanks, Matt Larson.
I'd also turn on the AC. Some buildings have like, you know, circulation that pulls air from the outside and then circulates it. I personally prefer working with my door closed because of the noise.
So that means that if I don't ventilate it a couple times a day, like I open the door like once every four or five hours, this invariably leads to CO2 levels above 1200 parts per million before EOD, which if you think about it is now capable of resulting in like a... probably a 30 to 40 % decrease in cognitive function. Keep in mind that CO2 is also not distributed everywhere.
Depending on its weight density, depending on the other things in the room, CO2 could sit in different places. But this is a fair proxy, I would say, for how good your brain is doing. I wish I could actually reprogram this so that I showed an additional metric, and maybe I can, that shows like percentage cognitive impact.
Maybe, you know, as it rises to 763, I'm already at 3 or 4%, which, you know, if presented to me in this manner, would alert my brain to being like, holy crap, dude, you're doing something bad. Here's another really easy gain. Remove any intelligible speech in your environment for an improvement of 2 to 6%.
Do not listen to podcasts, music with lyrics, or any family, friends, partners, or colleagues while you work. It is okay to have instrumental music or brown noise or fans or rain or other masking sounds, but the key is anything that is intelligible will negatively impact your ability to think clearly over long periods of time.
This is because when you talk, your brain will automatically try to figure out the meaning of what those words are. If I am hearing a bunch of other people say stuff, I will constantly be doing that background processing because attention as a gatekeeper is only so good. We're always going to be letting in stuff that we're not directly putting the spotlight on simply by virtue of how attention grabbing it is.
And if you think about it logically, you know, when you do this, it consumes a lot of cognitive resources. Even when speech is mostly unintelligible, even if it's just a bunch of murmuring through a door or something like that, you still see performance decreases. And the second you follow a conversation, you actually have a major penalty to your reasoning, which empirically ends up being somewhere between 44 to 45 % or so, which is obviously a wide range, but it just depends on what task you're doing.
Tasks that involve verbalizing and reasoning through words tend to do way worse than those that maybe are more spatially based. Okay? So how exactly does this work?
Well, speech tends to get impacted by intelligibility, not loudness, as we see here. So one time, a client of mine was discussing low performance of their writing team. And I love these guys to death.
So if any one of them are listening to me, just know that you guys are the best. I visited their office, and I noticed that all of the writers on their writing team sat next to each other on a big open floor plan. And they also often spoke to each other.
And it was what you would call a tight -knit team. You know, people really liked each other, and it was fun to go to work. However, I noticed that when a single person in the office spoke, everybody else could hear what that person was saying.
Now, logically, and I didn't know what I know at the time, but I knew enough. This meant that all other performers in the office had a tax on their ability to write between four to 45%. And I would guess probably closer to 45 % because writing is obviously directly reasoning based.
So what that means is, obviously, everybody in their office was constantly underperforming. Why? Because, you know, if it's not, let's say it was an office of 10 people.
If one person talks 10 % of the time. That is virtually, literally the entire time that people are working. They're going to have somebody saying something, which will be impacting their ability to reason verbally.
So what I did is I suggested they move their best writers, the people that were performing the most, to closed offices for a few weeks. Just a few weeks. And they obliged.
And wouldn't you know it, all of their KPIs massively skyrocketed. Like, night and day. And I mean, sure, some of it probably made the environment a little bit sadder, because now we didn't have a bunch of people together cracking jokes and shooting the shit.
But I'll be honest, the purpose of business is to be effective, right? They weren't currently not being very effective despite being all close -knit and stuff like that. So, you know, instead, they just hung out during lunch and had a blast then.
I would not recommend you pay the verbal reasoning tax on intelligible speech in your environment. It is completely needless. Personally, I use a combination of foam earplugs and a closed door.
And I know this sounds a little antisocial, but significantly improves my clarity of thought, and that is ultimately what we're going for here. So yeah, I mean, I don't have the earplugs with me. I would have shown them to you, but they're just bright orange and I stick them in my ear before I start working.
Another easy 2 -5 % improvement is working next to a window. As I mentioned, your circadian rhythm depends on your eyes receiving light. So to optimize this in your own workspace, my recommendation is have your desk near a window of some kind.
So in my case, for instance, I have a big floor -to -ceiling window immediately to my left. Not only does this function as nice lighting for my face, it also improves my cognition. Now moving your desk, if you think about it, is free.
absolutely free leverage. But if you don't have a window, let's say you're working in one of those closet -style, cave -style offices, what you can do is use a powerful lamp on your desk every morning, use it as an alternative, sit near that. For instance, I have a 10 ,000 lux one right over here, which I don't know if I'll be able to pull in.
Yes, I don't think you guys can see this really. It's just out of frame here. There it is.
Anyway, it's a very big and very powerful light. It costs me 200 bucks or so, and I use it in the winters before I wake up. I'll talk a little bit more about that later, but basically in northern environments, just due to the way the earth is tilted, you don't get anywhere near as much light in your average day until maybe 9 or 10 a .m., and most of the time I'm awake before that, so if I want my circadian rhythm to get jump -started, that's what I have to do.
Next up, reduce clutter for 1 -3%. Now, tension, as mentioned, is a spotlight, but it doesn't just illuminate things that are underneath that spotlight. The gatekeeper sometimes allows additional things into your working memory, even if you didn't want to.
And unfortunately, as we know, the working memory is quite limited chunk -wise, so you can only keep a certain number of things in that puppy at any one time. Meaning, if you have a bunch of stuff in your environment, even clutter on your desk, some of that will enter your attention, and thus your working memory, without you wanting to.
Now, just like intelligible speech, every object in your visual field is technically competing for some small share of your attention. If you have to constantly fight all of that off, it is consuming your resources. When you focus, your visual system has spent effort suppressing all of the clutter in your environment, which is called biased competition.
Meaning if your desk is filled with papers, cables, mail, and other miscellaneous objects, your brain actually spends energy every time you look at it, slightly classifying their location, their position, the way that they're oriented, and so on and so forth. So there's just no need to suffer that. And if you've ever wondered why really rich people's homes are typically quite clean, I think part of it is this.
They likely don't know it. at face value but it is likely one of the major reasons they just feel like they can focus more when they don't have clutter on their desks or on their kitchens or so on and so forth the next easy three to six percent has to do less with physical workspaces and more with virtual and digital workspaces and this one is batching your notifications for an easy three to six percent in case you didn't know phones and computers are designed to capture and hold your attention It's an inherent part of the economic value model of companies that make them.
The more time you spend on your phones and computers, the more money these companies will make. So they set their defaults as high levels of notifications. These play into their incentives and then make them tons of money.
What I'm going to do here is help you change those settings a single time and then gain clear thinking as a result and then be done with them for the rest of your life. to easily and immediately pick three to six percent on the cognitive capacity side, but probably far more than that on the general life happiness side. Now first, every notification forces your brain to temporarily switch tasks between what it was doing and then the notification.
As mentioned, your attention and working memory is quite limited, so when something salient in your environment occurs, like maybe a phone buzzing or maybe like, I don't know, your screen turning on in the corner, your mind will actually attend to it whether you want to or not. This leads to a byproduct called attentional residue.
where even if you shift back to work, okay, if I see the notice and then I go back to my screen, a part of my attention, a part of my spotlight is just impossible to wrench away from my phone. That means I'm basically keeping track of it in case it buzzes again or in case it lights up again. Because, you know, evolutionarily, I'm constantly keeping track of all the things in my environment that move in case it's a snake coiled up ready to attack me or something like that.
Now, in the literature, this can cost you... 10 to 15 % of your cognitive capacity for as long as 20 freaking minutes. And many of you guys, unfortunately, get disrupted and distracted by notifications more than once every 20 minutes.
Some people here probably get distracted by notes like every three minutes. Put it another way, you are basically, like in my office example, permanently hobbling your ability to think clearly because you always have somebody saying something in the background. The good news is the fix is really easy.
You just go through your phone, turn off all of the banners, bubbles, and sounds notice. And instead, you schedule your notifications to be delivered in two to three batches per day using the built -in and iOS feature called Scheduled Summary. Whereas on Android, you can just turn off all notifications and check your phone at specific times, like 11 or 4 p .m.
The way you do that is just by setting an alarm. So you guys hearing me right? Instead of checking your phone randomly over the course of the day as follows, okay, 24 -hour chunk, you're constantly just looking, looking, looking, looking every couple of hours, and then facing that 20 minutes of additional residue afterwards.
What you do is you choose the times to batch and check. Maybe you do two or three throughout the day. And now instead of constantly being in a state of 20 % reduced performance for maybe 50 % of the day, so 10 % across the board, maybe you only have to suffer two or 3 % while still getting the upsides of using this really cool technology to stay in touch with people.
So I run all my social media channels personally using the same principle. If you think about it, one major revenue driver of mine right now is I run Maker School, which is a community that guarantees your first customer for an AI or automation service in 90 days, or I give you your money back. So there's lots of people in that freaking community.
I mean, almost 2 ,500 right now. And they make posts all the time. Just like a social media notification, I could respond to their post every day, all day, throughout the day.
But instead, what I do is I batch it. What I'll do is I will check in at one or two intervals every day. I will then respond in bulk and then I do not let it eat up the rest of my day.
So rather than me being constantly pulled away from what I'm doing and then having to go back to this thing, you know, I can just focus on the thing and get it done. This is obviously far more efficient than me checking in 40 times a day. And in addition to helping me think clearer throughout the rest of my day, it also probably saves you a lot of time because you're not constantly being pulled away and then working in an inefficient manner, meaning you have both the direct gains in terms of performance improvements and you also have the indirect gains in terms of time improvements as well.
Another really quick and easy hack is to single task your screen for three to 8%. Now, similarly to the clutter rule, right? We spend a lot of time in a virtual environment, which is basically like our physical environment.
Now the screen almost extends into three dimensions. I could see far past it. So similar to that rule, you should minimize clutter on your environment.
The way that I do it is I go full screen with no other apps or tabs visible. This minimizes the attentional drag and forces me to actually just sit and focus on one task. For instance, while recording this video, I only have one window open.
I do not have 45 different monitors blinking stock tickers at me and whatever the hell else. All I have is this. Therefore, all I can do is this.
A couple of exceptions, obviously, if you guys are doing explicit design heavy work. if you guys are crazy active day traders or whatever the hell, if you need to constantly compare things between one screen and another, you know, there can be some value in having a few windows open, a few tabs, the visual ease of dragging and dropping stuff and looking and so on and so forth probably does outweigh the cognitive impact.
However, all other times and in all other situations, I would close your email, your chat apps, and all of the 31 tabs you have open, unless you're explicitly working on getting those through. I'd also remove your URL bar, your taskbar, and give yourself nothing but the thing in front of you. What do I mean by URL bar?
I mean, I would literally like go to the setting in the top right or top left -hand corner of Chrome or whatever operating system you're on and go into the view tab and saying, turn off URL bar. I'd go down to the bottom. I'd say, turn off taskbar.
I'd make it so that when I look at my screen, I only have the thing that I'm currently working on. It just makes you far more efficient. All right, that's it for the hardware portion.
If you guys have followed along with everything up until now, you're probably looking at a stacked multiplicative improvement in the clarity of your thinking and your focus, right? So combine those to the tune of about 150 to 200 % or so. And obviously, you know, a lot of these are estimates.
Of course, I just took what the science said and then gave them to you and told you to multiply them. I think you have to infer the actual end impact on your own. stats, if you want to call them that, of attention, working memory, and executive function through the lens of your own problems and the own things that you struggle with.
For instance, I think a lot of people here probably suffer from way worse than the 800 parts per million CO2 issue. I think a lot of people probably have ambient problems with air quality, especially in Southeast Asia, Bangladesh, India, and so on and so forth. I think there are different things that...
everybody gets to pluck out of that big list. But so long as you plucked out a couple of them, I think you'll probably be at least, you know, 20 to 30 % better, if not 50 % or more. And because all of those scale with the magnitude of your starting place, if you're sleep disordered, for instance, fixing your sleep will lead to a significant improvement in your hardware that, you know, a non sleep disordered person would probably feel.
I think, you know, we all sort of carve out our own story there. So anyway, not bad for a bunch of low hanging fruit, huh? That's about as free leverage as you could possibly find.
From here on out, what I'm going to be doing is pivoting to software and creating a toolbox of processes and algorithms that help you think more clearly. Very interesting thing. Software is infinitely more malleable and growable than hardware.
Whereas our hardware, basically our attention, working memory, executive function are fairly fixed and can only be nudged or expanded slightly. Better software can provide algorithmic improvements to the tune of many thousands of percentage points. And I don't say that lightly.
I think great thinkers are genuinely orders of magnitude faster and better than poor thinkers are. And it's all because they just apply the tools that I'm about to teach you. So some action questions just to wrap that up.
Before we go into software, answer these two questions. Take an inventory of the room that you are in right now and apply the workspace design principles to it. What is your air like?
Can you hear the people around you? What is in your direct field of view? And how can you improve it?
Next, open your phone. What notifications are asking for your attention? Turn them all off unless absolutely necessary.
So on software, obviously the human mind of the computer are different things, and you can't analogize them one -to -one. You know, one is silicon, the other is meat. But bear with me.
Between 1988 and 2003, computers got around a thousand times faster via hardware. Better memory, better processors, better architecture, and so on and so forth. And so if you think about it, that's pretty incredible, right?
But do you know what is even more incredible? During that same time period, software algorithms got 10 ,000 times faster. Now, clearly, software is a different beast entirely.
Whereas you could extract significant and fixed gains from improving your own hardware, having the right program to run on is an order of magnitude as powerful, since it also allows you to consider problems from a fundamentally different lens. So imagine, if you will, that you were only ever taught addition, you know, no multiplication, no nothing.
If I asked you, hey, can you add the number 10 to itself 100 times, aka 10 plus 10 plus 10 plus 10 all the way up to 100, how long would that take you? It would probably take you forever, right? It doesn't matter how fast you are at mental addition because fundamentally, the solution to adding a bunch of numbers together is slow and it's linear.
And it'd probably take you three or four minutes to actually go through the process of adding 10 to 10 to 10 to 10 100 times to get to 1 ,000. There's also, in a situation like that, a much higher probability of error, because maybe you just missed one or two of those tens, and now you think it's 980, whereas the actual answer is obviously 1 ,000.
But imagine giving that same problem to somebody who has the multiplication program installed on their brain. Whereas person A might take four minutes to do it, person B could literally do it in a second. Not only would they do it in a second, the error rate would be far, far lower.
That's exactly what I intend to show you guys today. I intend to show you a set of operators that you could use to better understand, conceptualize, and then think through all of the problems in your own life. And the first is expected value.
Now, most decisions always get made based off of gut feeling. This is type one decision making. In the average day -to -day life of a human, this is usually fine.
Because the decisions that you get are things like, what coffee did you want today? Eh, I'm feeling like a latte. Holy crap, is that my bus?
Let me run over towards it. And so on and so forth. But you're here because you want to learn more than that.
You want to learn how to think through decisions that ultimately impact the rest of your life. So that's things like, should I launch X product? Or should I spend the next six months of my life on Y pursuit?
And these are nuanced questions that have nuanced answers. You know, gut feelings are ultimately a terrible way to make high risk, high impact decisions like that, because gut feelings are always just yes or no. They're also not accountable at all, and they're also not interpretable.
The whole go with your gut thing sounds nice in practice, but when you can't look back and say, so why the hell did I do that thing to begin with? And your only answer is because my gut told me so. Obviously, that is not a very rigorous decision making process that you can rely on for larger things.
The solution to all this is expected value. It is the single most useful piece of arithmetic that I know, and it'll take you literally just one minute once you've done it a couple of times. Now, expected value, defined in as simple terms as possible, is when you sum each possible outcome's value multiplied by its probability.
And I know that sounds really mathematical, it's probably pretty intimidating to you, but it's actually very simple. Let's pretend that you and I are playing this game where we flip a coin, and if it comes up heads, you get $3, and if it comes up tails, you lose $1. Now, logically, the coin flip has two outcomes, heads or tails, right?
The probability of heads is 50 % and the probability of tails is 50%. So there's nothing super crazy here just yet. Keep following along with me.
In expected value terms, what you do to figure out the value of whether or not I should play this game is you multiply the probabilities of each outcome by their respected values. So for instance, a probability of 50 % is equivalent to 0 .5 mathematically. So that's our probability.
And we already know the value of the outcome, which is $3 for heads, right? Which means our equations look like this. If we wanted to figure out the outcome of heads, well, heads is 0 .5 times the outcome, which is $3, means that the EV of heads is actually $1 .50.
Tails is 0 .5 as well, but this has a value of minus one, because that's where you lose a dollar, which means that the EV of tails is 0 .5. Now, if you wanted to get the expected value, not of a specific head or tails, but the whole game itself, then what you would do is you would take the value of heads and add it to the value of tails.
The value of heads is $1 .50. The value of tails is minus 50 cents, meaning the expected value is a dollar. What this means then is that the expected value of playing this game every time is a dollar, meaning that if I play the game a thousand times and I zoomed out, and ask myself how much money would I make if I played this game a thousand times, I would make approximately a thousand dollars.
Because every time I play, statistically, I win a dollar. Now, while it's true that you probably won't make exactly a thousand dollars, because of statistical chance, you know, I might have a lucky winning streak, or I might make a bit over that, maybe a thousand five dollars, or 943, or something. The more you zoom out, if it's 10 ,000 runs, or 100 ,000 runs, or a million runs, the closer I will get to that actual reality.
Okay, so that is the hypothetical coin flip example, which I think people understand intuitively. Let's apply this to something that's a bit more real life, shall we? Let's say I run a small service company right now and I install solar panels.
I've seen a bunch of cool new developments on Twitter, and I'm trying to decide on whether I should start a big SaaS company, which is new and exciting, or I should keep my small service agency running, which is small, old, and familiar. Instead of me just being like, what does my gut tell me? Man, I should go big or go home.
I'm going to start the SaaS. We can actually compute the expected values of both of these options using this new framework. So let's say you do a bunch of research, and you determine that the industry that you're trying to start a SaaS in is willing to buy a company like yours for a billion dollars.
Maybe that's the average sale value. Meanwhile, you also notice that the majority of companies that are started in this space fail miserably, meaning that your probability of success is very low. Maybe it's 0 .1 % or 1 in 1 ,000.
What that means is you have a 1 in 1 ,000 chance of making a billion dollars. Kind of how that works, right? So keep that in mind.
On the other hand, let's think about the SaaS business, the service business. Your current service business makes $500 ,000 a year. You do some market research and think about your business and you say, you know what, this is probably going to run for another five years minimum.
And I give it a 50 % chance that I'm right there. Maybe 50 % of the time it'll run less than five years, maybe 50 % of the time it'll run more than five years. Okay, so that means the expected value of starting a SaaS versus your current company is as follows.
The EV of the SAS is, again, equal to the value times the probability. So the EV of the SAS is, again, $1 billion times 1 ,000th of odds. So $1 billion times 1 ,000th is $1 million.
That's just how that works. What that means is the expected value of doing the SAS is if you choose to make the SAS, it will make you $1 million statistically. If you could zoom out and start a billion SASes on net, every one of those assets would actually only make you a million dollars.
Why? Because the vast majority of them would fail, only a couple of them would win, and the ones that win on average would win approximately this much. Now, how about the expected value of your current company?
The EV of this small business is again equal to the value times the probability. Well, I added an additional step here because I wanted the value to be non -trivial to compute. So in this case, I said the value was $500 ,000, but how many years would it run?
Five. So we actually have to multiply these two numbers together. which means the EV of the small business is actually 2 .5 million times our 50 % probability that we'd actually still maintain the business for that five years.
What we get is 2 .5 times 0 .5 or 1 .25 million, which if you think about it logically means that the small business makes us 1 .25 million. The SaaS makes us 1 million. So in this example, the expected value of the SaaS is actually lower than the expected value of the small business.
Put in simple terms, the value of starting this big fancy company is not as high as the value of just continuing to work on your SMB. Now, I know what you're thinking here as I go through this example.
You're thinking, Nick, how the hell am I supposed to know the precise dollar amount or probability of every single decision that we're considering making? If you do know this somehow magically, give me a call. I have some work for you.
The reality is you will not know these numbers, and that is totally normal. The point of expected value is not meant to be taken literally. The point is, it's meant to be a tool that you use to objectively evaluate multiple different objects.
objects or options in concert. Rather than just going off of somebody's gut feeling, which is always the same circuit that was used by your animal ancestors, which is only yes or no, this actually lets you upgrade your software and apply a stronger, more rigorous decision -making process to the art of making decisions. And so instead of you just being like, I'm going to do this or I'm not going to do this, you now actually say, well, I predict that if I do this, I'll make this much money.
That is now a falsifiable prediction. You can actually take that prediction, store it somewhere like in a journal, in a year from now come back and update it based off of your actual experience.
Likewise, you could share that prediction with other people, have them follow the same rigorous standardized thinking process, and in that way think far more clearly about whether or not something is likely to actually work. And the cool thing is in the case of the SAS example, we see how that an option that may seem sexy, exciting, and perhaps like the right gut feeling for most people, obviously I want to go big or go home, that sounds fun, actually underperforms compared to, in this case, the very boring and very safe one.
I want to talk about poker for a second. Professional poker players that are with me right now will probably have nodded along this whole time and been like, okay, get to the good part. The reason why is they already know expected value, and they do that math in their heads constantly.
I'd say that's actually one of the core skills of playing poker really well. Good poker players are always computing the expected values of their hands versus their competitors' hands at all possible times. And then the entire way you play is you just look for games where you have a slight edge, meaning the EV of you doing the, I don't know, you know, folding or something is a little bit higher than the EV of you going all in.
And as long as the expected values are positive, they will play around. And if the expected values are not positive, they'll stay out. So since the expected value is all about averaging out outcomes over a long period of time, poker players that even have a slight edge, like they make like an expected value of like a dollar per hand or something, will typically come out on top over the long run.
The issue is it won't be abundantly clear in your first few hands because you need to play enough times for the law of large numbers to work out in your favor, for statistics to kick in. And that's why having enough of a betting pile comes in handy. Because the more money you have to bet, the longer you can stay in the game enough for your fundamental expected value bets to pay off.
This is the same thing, by the way, that underlies why people that have a lot of money tend to be a lot less risk averse. If you have a lot of money, if you have a big betting pile, and you're confident that at some level the expected value of your decisions are right, you're okay wasting a bunch of that money on bets that may not pay off in the short term, but will pay off in the long run.
That's because your model of expected value is high. If you've ever wondered why some people go all in on stuff or spend tons of money on small things or, I don't know, venture capitalists give $50 ,000 to 100 different companies, that's because they compute the expected value of every one of these bets as just a little bit bigger than zero.
Maybe it's, you know, $50 ,000 or something. And so in that way, they know they just need to do enough of these. Eventually, they'll make $50 ,000 and their $50 ,000 doubling their ROI.
I mean, VCs in particular, they want way more than a double, but you get my point. So all of that is to say your probabilities are estimates. You won't know them perfectly, and that's totally okay.
The key is just write them down. When you write 20%, that is something that is concrete and it is modifiable. If you just feel, nah, it's risky, I don't want to do it, that is not concrete, modifiable, and you cannot use it to make good decisions.
I want to talk for a sec about the resulting trap. Now, when people start learning about EV, they will often fall prey to the trap of resulting. The term was initially invented by Annie Duke, who was a former professional poker player back in a 2018 book called Thinking in Bets.
You guys might have read it. It's very popular when it came out. The core idea is resulting is the tendency to evaluate the quality of a decision based on the decision's outcome.
For instance, let's say you use expected value on a hypothetical decision between A or B. Here is the expected value of A. It is 2 times 0 .8.
equals 1 .6. The expected value of B is 10 times 0 .2, which is equal to 2. If you think about it in our formula, 2 here is the value, 0 .8 is the probability.
So for A, there is a 80 % chance you will win a 2, meaning the value of this play is 1 .6. For B, there is a 20 % chance you'll win a 10, meaning the value of this play is 2. So logically speaking, you know, A is a lot safer because there's a much higher probability you'll win.
With B, it's less likely you'll win, but when you do, you win big. And so that's how you sort of think about expected values. You think about it in terms of probabilities and outcomes.
So in an environment like this, where EVB is equal to 2 and EVA is equal to 1 .6, logically you should probably pick B, right? Because EVB is higher than EVA. But keep in mind the probability of 0 .2 or 20%.
Despite B being technically the right answer according to our formula, it'll actually pay off 20 % of the time. So if you choose B and then you don't get that 10 payout, you will naturally consider it a dumb decision. You'll be like, man, I know it said the EV was two for B and that's why I picked it, but I just lost all my money.
Or I don't know, I didn't make the two. What the hell? This is logically incorrect because B is in fact the right choice.
It's just the right choice happens to lose 80 % of the time. The thing is, when you get the win, it'll pay out enough to reverse that trend 20 % of the time. So real life is kind of like that.
Because in real life, you'll have a lot of decisions that you make that don't immediately pay off. You need to hit the net constantly in order to make them pay off. But if you start evaluating your decision making based off the results you get on a small sample size, you will make incorrect inferences that lead obviously to incorrect conclusions.
If you penalized yourself for every single decision you made that did not immediately pay off, you would probably quickly learn never to take a B sort of bet again, right? But that would be a major mistake. Likewise, a lot of people find the opposite.
They might choose A, and then they'll get like a little payout of two every time. And then they'll say, I'm a genius. Every time I do this, I win, I win, I win.
So then you'll start logically thinking, well, it's smarter for me to continue taking A as opposed to B. Despite A paying a little bit less, that's fine. I'm making way more money.
This is the same thing that happens in poker. If you fold in a hand based on correct odds and then the winning hand comes in anyway, you have not made a mistake. You had the right idea.
The fold was the right move despite losing. The point is you just need to zoom out and understand that this one time that I won or this one time that I lost is not actually representative of the whole thing. I see this, you know, B play a lot in business, specifically in cold outreach, because there is about a 20 % chance that most people will see any sort of outreach that you send.
And so if I'm attempting to sell my product or service and I send an email to somebody, let's say, and they only had a 20 % chance of looking at it, that means 80 % of the time they see nothing. So if I am new to this and I send out outreach and four times out of five people don't see it, I might think, this sucks. I don't get anything out of it.
It's simply because I do not have enough of a sample size to draw from for me to understand that, well, when you do get something that pays off, boy, it pays off really big. If you understood this on a visceral level, which you do after you usually get a couple of wins, the probability of you staying in the ring for longer goes way up.
So zoom out, play a thousand hands or make a thousand decisions or start a hundred businesses or whatever. When you do this, you will easily be able to tell the difference between a good decision and a bad decision. Okay, so here's a little decision matrix.
If you had a good decision, but it was a bad outcome, that means that you just had bad luck and you should keep doing it. If you had a good decision and a good outcome, you obviously deserve the win. If you had a bad decision and a bad outcome, you obviously deserve the loss.
But if you made a bad decision and got a good outcome, you got lucky, and you should count that as a win. Here's another example for EV. If you guys were considering hiring a salesperson at a cost of $6K per month, and you believe there's a 60 % chance they'll succeed and generate $15K in gross profit per month after a three -month ramp, and a 40 % chance that they will fail and generate $0 in gross profit before you fire them at the end of that three -month ramp, resulting in an $18 ,000 loss, should you hire the salesperson?
Well, you can actually do expected value math by subtracting the two. So in the case of should I do X or should I do Y, or rather just should I do X in general, where X has two outcomes, the expected value of the decision is expected value of success, assuming everything goes well, minus the expected value of fail, assuming things do not go well.
So in our case here, and I'm just going to bury the mathematics here, because I know some people are probably listening as a podcast, not necessarily watching it, EV success is equal to 9 ,000. EV fail is equal to $7 ,200. That means that EV success minus EV fail is equal to $1 ,800.
That means that the value of making this decision is $1 ,800 over that three months. That means basically, if you make this decision from an expected value standpoint, you will gain $600 a month every month for those three months, which is obviously positive, right? Who doesn't want $600 for free?
Now, I do a rough version of this expected value framework in my head for small stuff, and then I'll actually take it on paper, just like I did here, for anything that costs me more than a few days or maybe $10 ,000 or so. What I have found is that writing the probability down influences my later decision -making far more than the actual arithmetic that I'm doing.
It forces me to try my best at quantifying the odds, which is most of the value, literally just articulating how likely do I think this actually is. So speaking of actual, let's do some action questions. Write an expected value calculation for a decision you've made in the last year with probabilities and payoffs.
Then decide whether it was a good decision or a bad decision based on your expected value. Okay, let's talk power laws for a second. Power laws, like expected value, are a cool framework that allow you to peek a little bit behind the curtain at probability theory and what I should probably be doing with my time.
And they're at the core of what I would consider to be effective, especially business decision making in insert current year. So to make good decisions, you must understand how the world is distributed. Most people have poor intuitions here.
They will assume that things in life are distributed normally, like height. Okay, for those of you guys that don't know, if you were to plot the height of a million random people, you would get a normal distribution or a bell curve distribution, where the majority of people are going to be of average height, and there'd be very few super short or super tall people.
And for those of you guys that have never seen this before, what that distribution would look like would be like this. You know, around here, zero, you would have the average height, which I don't know. I don't know the average height, but maybe for males, it's like five, seven or five, eight or so.
On the right hand side of the graph, you'd have a bunch of people that are taller and the left hand side of the graph, you'd have a bunch of people that are shorter. And so because human beings tend to intuitively understand that, you know, height is normally distributed, that most people are going to be here. Like this is frequency, right?
Number of people. And then this is the standard deviation. They're going to assume everything in life is like this.
but actually very few things in life are like this. And that's the key. Because human intuition often assumes that distributions of like clients, videos, decisions, finances, and tasks are normally distributed, they end up making a lot of bad mistakes.
The reality is you cannot make good decisions merely by distributing your time or effort equally among candidate options. What you need to do is instead of normal distribution, look for a power distribution. And this is where the power law, comes into play.
So what is a power law? Well, in mathematics, a power law is a relationship between two variables where one is proportional to the other raised to some power. This is what it looks like mathematically.
X is equal to Y raised to the N, let's say. This equation is less important than the graph. So when graphed, it looks something like this.
Generally speaking, almost all power law graphs are going to look like this massive descending exponential. And the reason why is because on the left -hand side of the exponential, you have the outcome or impact of a thing, typically the lowest frequency thing. What I mean by this is, let me pretend for a moment that we are running a videography agency, and we currently have 20 clients on our belt.
You know, we have client number one, client number two, client number three, client number four, all the way up to number 20. And on the left -hand side of this graph, we have the amount of money that every one of those clients has generated us. If you were to sort all of those clients based off of how much money they've paid you, the result would almost invariably look like this.
And this is just a universal law of money and client relationships. One of those clients would probably have paid you the vast majority of your wealth. The second client, descending order, would have paid you another big chunk.
The third client would have paid you another big chunk, and the fourth client would have paid you the last big chunk. But all of the other 16 clients would have paid you virtually nothing. What that means is literally just client number one is probably equivalent to the bottom 16 or so.
Another way of conceptualizing this is having a few items carry the total. Now, the general rule is in life, aside from things like height and some biological principles, one item is typically responsible for a disproportionate share of whatever the outcome of interest is. For example, in business, a power law might look like a few products driving most of your revenue, a few clients comprising most of your MRR, or in investing, a few stocks driving most of your portfolio return.
Many people here will probably have heard of this idea of this Pareto principle. It's a great example of the same concept where 80 % of results are brought to you by 20 % of causes. Pareto came up with this while studying the distribution of land in Italy way back in the day.
I believe he noticed that 80 % of the land was owned by 20 % of the people, but I might be mistaken there. A more modern look at that might be companies like Microsoft. Because companies like Microsoft understand power laws and the Pareto principle, what they do is they allocate the resources differently than if everything were normally distributed.
For instance, Microsoft will identify the 20 % of bugs in their software that cause 80 % of customer complaints. They will then spend only 20 % of their time on those bugs. Or rather, 100 % of the time on the 20 % of the bugs.
They will not distribute their time equally across all of the bugs, because logically speaking, it is only a small portion of the bugs that actually lead to the most customer experience. What I mean by that is, by focusing on the 20 % of the bugs, they could reduce the number of crashes by 80%. This number, 20%, is about four times smaller than this number, 80%, which means they can spend four times fewer resources to affect a result that is equivalent, or they could affect a result that is equivalent that is four times as large as they could if they were allocating the resources uniformly.
So what that means is they just get to do a lot for very little. And in business, that is the whole game. It's literally all leverage.
So the specific numbers will vary. It's not always 20, 80, 80, 20, but the shape will be the same. I always encourage you guys to think of power law as a graph where the left -hand side is really, really big and it sort of goes inverse exponentially down.
Now, why do power laws matter? Well, they matter because they make the concept of averages basically completely useless. You know, if you guys ran a business and your clients were all normally distributed, that means that the highest paying client of yours might only be 30 or 40 % taller or bigger than average.
So instead of maybe, let's say you're selling a $5 ,000 a month product, maybe they would pay you $6 ,500, right? That would have a major impact on strategy, wouldn't it? If most clients tend to pay you the same amount, your primary way of growing your business is likely just getting as many clients as possible.
And you would be happy to do so even if quantity comes at the expense of quality. But if you understand the way that clients are really distributed, aka through a power law, you'd understand that in that world, 90 % of your clients probably pay you very little money, maybe $2 ,000 or $3 ,000, whereas a single client might pay you way more, like $30 ,000, $40 ,000, or $50 ,000.
That's just how it works. What that means is, in this world, clearly, your goal is actually to avoid low -paying $2 ,000 a month clients and instead orient your business and strategy such that you can find more of those $50 ,000 a month bangers, since one of those clients are literally equal to 25 of the others. Now, this is true in my own business.
One of my own portfolio companies makes about one to one point two million dollars a year. At least that's their run rate right now. And single client makes up over eight hundred thousand dollars of that.
That is literally the power law in action. And I can think about all of the other businesses I run and you will see that same idea applied, although probably less starkly in a variety of ways. And so this is not optimal for a variety of reasons.
In business, it's called key man risk or key client risk. If you have most of your earnings coming from one or two clients, obviously it's risky because if that client disappears, you are then kind of screwed, right? I would only have $400 ,000 a year or $40 ,000 a month versus about $100 ,000 a month.
But it is what it is, and that's just how it is. Similarly, in my media company, I have a small handful of courses that I make that drive the vast majority of my views. And most of the other videos, including many videos that I worked very hard on, are literal rounding errors beside them, right?
One of my courses drives over 10 ,000 views every 48 hours. A variety of my other ones, which I invested way more time and effort into, make me maybe 500. Simply, it is power law.
You will always have one or two major levers that do most of the heavy lifting. You can think about this from technology as well. Your phone is a similar idea.
A few apps eat the vast majority of your screen time, meaning that blocking just one or two specific apps can achieve perhaps 80 % of the screen reduction time if you just spent that time and energy finding that out. You can also think of it more philosophically in terms of how your days, months, and years go. Realistically, it's just a couple of key decisions that occur throughout the year that dictate the vast majority of how the year goes, right?
It's like, should I have gone on this trip with this person? Should I have stayed with this person or broken up with this person? Should I have started a family with this person?
Should I have called my grand uncle at that special time that I probably should have or something, right? All of these things here are the power law of your decisions. So the idea is strategically, You want to find the 20 % of decisions that you make that dictate the rest of your year and then spend all your time and energy making decisions like that.
So how do you use the power law to actually make better decisions? As mentioned, first we find the items that make up that 20%. And this is much easier than most people think.
You just have to look at your pre -existing distribution. So for whatever skill or business or avenue you are currently taking this course to improve on, I would enumerate all of my outputs. What I mean is all the things that we have produced.
The 20 % will always be at the top of the distribution if you sort it in descending order based off results. So on YouTube, for instance, this would be the videos with the most views. In our time example, these would be the apps with the most usage.
And in our life example, these would be the most impactful decisions that you've made in the last year, which would be really cool if you had on a journal or something, that way you could review them. If you were EVing all of the decisions that you made, you would have that written down on a piece of paper, and you could very easily go back and maybe filter and sort them through to determine which ones actually had large impacts.
Now, once you identify the 20%, you have two choices. The first is you can double down on the winning 20%. For instance, if 80 % of my views come from just three video formats, well, I could just make more of those three formats rather than innovating on a fourth, right?
My second choice is to stop investing in the losing 80%. Logically, the losing 80 % makes me just a few percentage points of the actual outcome. Why don't I just cut them off entirely or just stop adding resources to them?
Doesn't really make much of a difference to me either way, right? Now this connects directly to the idea of opportunity cost and the hourly rate rule, which I'm going to cover in detail later. They're both very strong thinking frameworks because every hour that you spend on something that is in the tail of the power law, the tail is the really long skinny part at the end where there's very little outcome, is an hour that you could have spent on something in the head, which is that really tall part, aka in the client example, the client that pays you $50 ,000 a month.
I am going to continue producing long multi -hour videos like this one, even though they require many times the effort of an average video, they are the head of my distribution. And the numbers tell me to focus my efforts on the head because I'm attempting to proceed logically via power law. If one does not pop for me, that is totally okay.
I trust that the expected value of the comprehensive course program is way better than the expected value of the shitty course or shitty short -term video program that a lot of other people on YouTube are doing. Another example is inverting the power law. So in addition to there being a power law, there's also something called the chain law.
And you can think of the chain law as the inverse of the power law. The power law says that the most valuable item in a list is orders of magnitude better than everything else. But in a chain law, it's sort of like a chain is only as strong as its weakest link.
Every single link matters. So the strength of the entire system is going to be contingent on the worst performer in that system. In mathematical terms, the chain law says the reliability of a system is not the average of its links, it is the product.
Now, if that doesn't make sense for you, consider a chain of steps, like in a business process. Maybe you're making some sort of metal widget or something, a fidget spinner, let's say, and you have a factory that does so. If there are three steps in the fabrication of this product, and step one has a success rate of 100%, step two has a success rate of 100%, but step three has a success rate of just 20%, the overall success of this product is only going to be 20%.
Logically speaking, you know, all the products that pass through step one are going to be fine. All the products that pass through step three are going to be fine. But despite that, only 20 % of the ones that pass through step three are going to make it.
So from here, we can infer that some systems, typically not the power law outcomes that I've talked about earlier, and I'll talk about which ones divide into which camp, are actually very fragile, and they don't do well when their power law distributed at all. The way that you improve systems like this is you significantly increase finish rate if instead of 100%, 100%, and 20%, you actually provide equal emphasis on improving each step.
So for instance, let's say instead of 100, 120, we got this down to 80, 80, and 80, which is now 0 .8 times 0 .8 times 0 .8. If you do the math there, the total accuracy is now 0 .512 or 51 .2%. That's a 2 .5 % improvement on what it was before, which was 20%.
Another major point to make in chain -based systems, before I tell you what makes a chain -based system versus a power law -based system, is it's important to keep the number of links in that chain to a minimum. Because the more links in a chain, the more opportunities you have for one of those chain links to suck. If you're running a process where each independent step has a 90 % success rate, logically a two -step process would have a 0 .9 times 0 .9 or 81 % success rate.
A three -step process would have a 0 .9 times 0 .9 times 0 .9, or a 72 .9 % success rate. And we would just repeat that process as you go. The more steps, the more numbers that we're multiplying together, and ultimately the lower the success rate of the entire process becomes.
This is also known as chains compound down. And so the more steps you have to a process, as you guys could see on this graph here, the lower the probability of success gets. So ultimately and ideally, what you would do is you would just...
go to the top here. You would have one step in a chain or one link in a chain, and you would just make that link very strong. And there are ways that you can combine, at least in workflow automation and process automation and optimization, multiple steps into one to achieve the outcome of, you know, chains compounding down.
Now, you can also drive a disproportionate amount of value out of spending your time on the worst, lowest quality step. And this is why it's the inverse chain rule. Let's say I had 100 hours to invest to improve a system that looked like this.
95 % step one. 70 % step two, 95 % step three. How should I allocate my time?
Well, because as we know, it is the weakest link in the chain that breaks the chain. I should spend all 100 of those hours on that 70 % until it gets to 95 % or above. When it gets to 95 % or above, I can then start equally distributing my time between each of these steps.
Okay, so which is which? Naturally, if you've made it to this point, you're probably wondering, well, why would you just tell me the power law and then tell me about a law that is the exact opposite immediately after? It's because now you guys understand that different classes of problems tend to get solved in different ways.
So the way you solve power law problems. So power laws typically apply to outcomes where you're examining the outputs of a system to determine which system you should use. So if I had a bunch of systems in front of me, okay, and then I looked at all of their outputs over the course of the last week, maybe a bunch of factories in front of me, and I looked at which factories made the most widgets.
I would pick the factories that made the most widgets, and then I would double down on those because of the power law. And I would expect to find that one or two or three of those factories probably disproportionately outproduced the vast majority of the others. Now that's the power law.
The chain law doesn't apply to outcomes of lots of different processes. What the chain law applies to is an individual process. So the chain law makes sense when you're optimizing the performance of just one of those factories.
Like for instance, if my task was not to pick the best fast factory and then double down on the winner, if my task was instead, hey, I have a factory and I have to optimize the product and work of that factory, well, then I would apply the chain law. I would find the weakest link in the process. I would ask myself, how do I connect steps together to minimize the number of total steps or links in this chain?
And then how do I harden the weakest links to bring them up to speed? So some action questions here. What have you guys produced?
Revenue, views, clients, whatever it is that you measure over the last 12 months. And of that output, what would you consider the top three contributors? What would happen if you only did those top three contributors?
What about a process that you rely on? Take one, count the number of steps involved, and then estimate how often each one of these processes work. Maybe it's a process in academia, if you're in school.
Maybe it's a process in business, if you're in business, or something at your workplace, if you are working in corporate. Take those steps. and multiply their probabilities together.
Is that result a number that you would happily bet money on? If the answer to that question is no, you should probably apply the chain law. Next idea is of local and global maxima.
Now, local and global maxima are important concepts in what's called optimization math. I first learned about them while I was in research, and I've applied them to most of my life ever since. In short, human beings always want to go up, right?
We want to climb. Our goal is to make it to the peak wherever possible. So what human beings will do at any point in time is we will typically, logically, we will evaluate all possible options and decisions, and then we'll choose which one gives us the largest rise or the largest climb.
But what happens when you make it to a peak, even a local peak? Well, if you're already on top of a peak, think about yourself on the top of a mountain, all possible decisions, aka all possible places you could go, go down, right? So what happens?
Well, then you get stuck. Well, this occurs in math just like it occurs with people. Now here, for instance, I am climbing and climbing and climbing.
I constantly evaluate by looking back and then I look forward and I say, what direction moves me up and which direction moves me down? Well, obviously that direction moves me up, so I should keep moving. But eventually you make it to the top.
But if you notice and you zoom out, you realize that isn't actually the top that you could get to. That's just the top of the area that you're currently in. The top of the area that you're currently in is called a local maximum.
Whereas the top of all possible areas is called a global maximum. The issue is, okay, when you make it to the top of a local maximum, the only move is to go down. And most people, because human brains just always work in a hill climbing way, we're always looking to grow or improve in some capacity.
Lots of people will never ever want to make that first step down. Because as you see, we got stuck here. If we only kicked ourselves out, went down a bit, we would eventually be able to climb all the way up to that global maximum.
But it's going to require a loss of some kind. It's going to require like a step back, so to speak. Now, most people who are stuck in life or career, they're not at the bottom of the hill.
They're at the top of one. Unfortunately, that hill is usually quite small. And there's often another larger hill that would happily host them if only they came down from their little hill first.
So that's what a local maximum is. It's a peak, and from the inside, a lot of the time it feels like you're successful. But the issue is it's not the biggest peak, and because you're already at the top of your current hill, it's very difficult to depart.
A good example from my own life was in 2024. I was running a service company that implemented automated workflows for agencies called LeftClick. Then I tried really hard, but I could not grow past $50 ,000 per month, no matter what I did.
My main lead generation channels at the time, which were Cold Email and Upwork, they were inherently difficult for me to scale past that. So instead of stay there at this local maximum, which was a great maximum, but it was not the maximum I thought I could theoretically have eventually climbed to, I started looking around and I saw people at a variety of these peaks that were far higher and greater than where I was at.
You know, they were at peaks of $100 ,000 a month. They were making $500 ,000 a month and so on. And I noticed that many of them were using an organic brand as their funnel and not cold email and Upwork.
So what I decided to do was I decided I would take a short -term hit. of figuring this stuff out, meaning I would step down from my local maxima. And eventually, after doing so, I immediately found the path to the global maxima, which has now allowed me to hill climb in a different direction that has since paid many multiples.
Now that I'm here, I'm zooming out and I'm realizing, well, hey, there's actually even bigger maxima. And I'm asking myself, okay, what sorts of decisions do I have to make to step back in order to get to the same place that I was at before, right? Grow on a new hill.
So how do you actually tell whether or not you're at a local maximum? Human beings don't just have nice graphs. Obviously, life is a little more complicated than that.
So instead, you need some form of sign. And there are some reliable ones here that I always go by. The first is if your improvements are shrinking.
If every change that you do is buying you lower return on investment than the last. Well, that's a sign that you're nearing the top. And that kind of makes sense just looking at the graph, right?
Like over here, me moving in the direction of the top moved me a lot vertically. But over here, me nudging myself just a little bit hasn't really moved me all that much vertically. The slope of that line has changed a lot.
Another reliable one I go by is if the input is not producing more output, aka if I'm spending a bunch of extra time, effort, hires, hours, whatever, and I'm not generating any returns at all. The last one is if someone has less skill than you on maybe a different hill doing something similar and they're beating you. And so, for instance, that's what I saw when I looked over at a bunch of people with organic brands that were absolutely crushing it financially.
I was like, I feel like I know way more about all this business stuff than they do. What's going on? So I'm really big on optimizations if it's not clear, meaning this was pretty hard pill for me to swallow.
I always wanted to look for ways to do better. And what I had to come to realize was the only way to do better sometimes is to take a step back and actually do worse. But me embracing this has resulted in far greater clarity of thought, and it's also made me way more money, which is a win -win.
So there is a cost of stepping down, as mentioned. To go from a local maximum to a global one means you have to cross this valley. That typically means you have to earn less money, or have worse numbers, or be perceived by others as doing worse for some period of time.
The question is, how do you bear this, and how do you choose to do it anyway? The simplest way is by quantifying what that valley looks like. Instead of it being some vague or scary thing, you should make it a concrete number and then accept it.
Now, for example, I concluded when I made the transition I was telling you guys about earlier that I was open to making $25 ,000 a month less for the next three months while trying out the brand thing. Craziest thing is I didn't actually end up shrinking my income at all, but I was prepared to. The second thing is not jumping off the cliff into the valley all at once.
What I mean by this is you can actually allocate a portion of your time, maybe 10 or 20 % of all working hours to this new endeavor while you guys are still working on your current thing. So in my case, what I was doing was I was still sending lots of cold email and Upwork applications. I just reduced the time I was investing on that by 60 to 90 minutes a day and said, okay, I'm going to redirect this to my brand.
So the idea is I was still maybe spending 80 % of all of my hours on the thing that I knew was growing my business. The only 20 % of my time was being spent to explore this new opportunity, which I had not carved out a return on investment for. The downside is obvious.
It's very certain. And the upside is sort of like unclear and it's vague, which makes sense why so many people get scared to take the leap into that valley. But you should.
You really should if you are in one of those local maxima. And if you're not in a local maxima, keep climbing until you are in a local maxima. It tends to be pretty obvious when you get there.
Okay, so let's answer two questions before the next section. The first is for you to name one thing in your life or relationships or your business that has stopped improving no matter what you do. Is that a local maximum?
Think about that. If you step down from this local maximum or changed your approach in the pursuit of climbing something taller, what do you think would suffer? By how much and for how long?
Here's another concept that'll massively improve the quality of your thinking. It's called attentional residue. This is coined by a woman, Sophie Leroy, in 2009, and it refers to the phenomenon where when you switch your attention from task A to task B, a portion of your attention will remain stuck.
sort of like a gooey residue on task A, regardless of how much time and energy you try and use to force yourself to focus on task B. Now, the larger the intentional residue, the more impactful, negatively, it is to your next task. We talked about this a little bit earlier, but I wanted to really go in detail here because I think it's at the core of what makes people so unproductive these days.
Most of the time, we're switching tasks because of some interruption these days, which makes it a very common problem. especially with AI models, which I think inherently incentivize having multiple windows open, sending requests, working on something else, and then doubling back. My current consensus on interruptions, and one that is backed by science, is it's occurring once every three to five minutes or so, which is insane.
If you are anything like the statistical average, that means that once every maybe four or five minutes or so, you guys are getting interrupted by something, and that interruption is causing a residue which is negatively impairing your ability to think clearly about your next task that is always slowing you down. Now, a bunch of studies have been done on just how bad the impact magnitude of attentional residue is.
And they vary based on the type of task and the length of task. Some are quite high, but I estimate it as a little bit lower than sort of like worst case scenario, maybe five or 10%. But another way to put this is if you fix your attentional residue, you guys immediately gain back five to 10 % to your cognitive capacity, which is insane, right?
So basically what happens is, you know, if you are doing task A, and then you move over to task B for a second, you'll notice that part of you is still on task A for 20 minutes before making it all the way to task B. Well, that is the attentional residue. No matter how hard you try, you will have this like 10 or 15%, or I mean, maybe five, let's say five to 10 % impact on task completion until you can sort of finish this timer.
Now imagine hypothetically that you were just doing this all the time. Task A, task B, task A, task B, task A, task B. You would always be under this red overlay, right?
Well, that's the problem. Now the good news is you don't have to be forced to suffer this forever. Even if you have to work in a fractured state, there are ways to minimize the impact.
There are a bunch of studies that Sophie Leroy continued doing that found that the difficulty of detaching was the highest when people were expected to return to a task in a time -pressured and rushed state. Which means that if you want to mitigate attentional residue, and we just accept that this is something that will continue happening to us no matter how hard we try, the simplest and easiest way is to make what's called a resume plan before you start switching.
Okay, so what is a resume plan? It's just a minute where you take note of where you're at now with your project and the next step you have to take when you get back to it. So in the example of switching from task A to task B, what you would do is...
task A, 20 minutes in, you decide to switch. Right before you switch, maybe that last minute or so, you'd say, okay, here's where I'm at and here's my next action. And then you'd switch to task B and the attentional residue would last maybe a third of how long it lasts here.
And it's very easy to do. If you have a to -do list or a project manager or something, you can immediately eliminate most of the attentional residue. Obviously, the best way is just abstain from moving over tasks entirely.
But this simple habit can significantly reduce the attention residue that's left behind. Instead of 5 -10 % of that sticky, tar -like substance, maybe it'll bring you down to 1 -2%. I have a personal rule, which is no more than two open tasks personally.
If I have two tasks open that I'm working on, nothing else can be open until I'm done with at least one. And so in my case right now, if I'm writing an essay and I'm waiting for an email, I will not open a message tab or have a community feed open or open a calculator or a spreadsheet to quickly check a number or whatever.
I will literally just stay focused on that essay, even if it feels mind -numbingly slow and even if every part of my body just like is begging me to change. The harsh reality is I am going to be way more effective finishing that off if I just stay on that screen than if I move around, even if it doesn't feel that way. So I either finish that essay.
Or I close it properly and I leave a note on what I will do when I come back. And that's where those task managers are quite valuable. And I know that this sounds inflexible.
It is. Unfortunately, the human brain just does not work well juggling multiple tasks. We blow at multitasking.
And if you've ever wondered why some people can just blaze through their work twice as fast as others while you may struggle to make heads or tails of what it is that you're doing, it is likely in part because you are the one that is fragmenting your attention constantly, whereas they, for whatever reason, maybe a God -given habit or maybe something their family instilled in them or something they learned from school or a course like mine, tend not to.
Consider, if you were writing a piece of software, would you design it so that it could have 40 browser tabs open at once, each tab using a portion of the computer's memory to slow down the rest of the system? No, of course not. You design it in such a way where, you know, the tabs that are active are the ones that will be getting most of your focus, but you would probably still have some sort of background leech going on with the other ones just in case you needed to move it on, right?
That's how your brain works. So most of us run our brains like this all day long, constantly task and context switching between something else. Do not be one of those people.
I realize I wrote two questions here, but I only had one. Which task did you leave unfinished today that is still running in the back of your head? What would you write if you had 30 seconds to create a resume note?
Akrasia is a term from an ancient Greek writer that literally means a lack of self -control. More generally, it refers to the state of knowing what the right action is to take, but still not doing it. Now, Aristotle wrote about it in the Nicomachean Ethics.
A bit of a slug if I do say so myself. You've probably experienced it at some point today. And I mean that literally.
Accretion is probably the most common thing that people these days have to deal with from a knowledge work perspective. It answers the I really should do X statement. If you've ever caught yourself saying that you have in one way or another totally experienced accretion.
So the following is an excerpt from an essay that I read back in April 2009 by a guy called Scott Alexander. It was originally posted on this website called LessWrong called Rationality is Not That Useful. For those of you guys that aren't aware, Less Wrong is a wonderful blog where I learned probably more than half of what I've been talking about in this course.
It's a really cool resource, especially their concepts to practical section. There's a lot of talk on like AGI and discourse and stuff like that, which can be valuable, but I do find it overly technical and not 100 % applicable. But the stuff on practical ways to utilize rationality to improve your life is just insane.
Probably some of the highest ROI I've ever had reading an internet resource. Anyway. So Scott looks at a group of people, people who have studied thinking clearly or seriously than perhaps any other group in existence, and then he asks if they're actually good at it.
His answer is frank, not really. He concludes that the correlation between consciously practicing rationality and actually succeeding in life is probably closer to 0 .1, not 1. He attributes this not to the inadequacy of the tools, but to akrasia.
This is the bottleneck. And his conclusion is this, it doesn't matter how good your decisions are if you don't have the willpower to act on them. Okay, so basically what this shows us is bottlenecks or capping.
We have a variety of different things that are being mixed in kind of a chemical container right now. But you'll notice that only a few of these actually get used. And the three things that we're mixing are what you know, what you do, and then what you get.
So despite you adding a massive amount of knowledge to this container, the only thing that actually activates that knowledge is what you do. And so what you do is sort of the limiting reagent of this whole chemical reaction. And because what you do is directly correlated with what you get, you cannot increase what you get without increasing what you do.
The unfortunate reality is most people are actually already at this point of their knowledge. They already have more than this orange line in terms of knowledge surplus, what they know. Their issue is almost always what they're doing.
And in very small cases, it's how they're transforming what they're doing into what they're getting. But the bottleneck is almost always just like how much they are currently doing. And I know this because I've run a community where I get to see day by day what everybody works on for, I think, almost two years now.
And today, I think there have been over 1 ,400 people in that community. Sorry, 14 ,000 people in that community posting literal daily. Hey, here's what I did today.
Here's what I was thinking today. Here's what I learned today. And here's how much money I've made.
So I basically have like a sociological study. of all of the ways that you can win and you can lose through Maker School, which is incredible. Now, since people are incentivized to create value and post their wins, there's something like 45 ,500 threads, each of which often have a dozen posts or engagements or more.
And the answer to me, because I did a bunch of analysis using Fable the other day, is clear as day. It is not the most knowledgeable people, the people that write at the highest writing levels, the people that are most well -educated. or the people that have done the most quote -unquote inner work that consistently win and make lots of money.
It is always the people that act, the people that send the emails or build the rough versions of the MVP, or the people that call and screw up. It is like the 18 to 19 -year -olds that come into this with zero knowledge about what the hell to do, who've never talked to a client or picked up the phone before, that don't even know enough to screw themselves up, that just start.
They are the sorts of people that have their what they do in excess and what they know sort of below them. And if what they do is already quite high, then teaching them information is no problem whatsoever. It's the people that don't move forward that tend to consume the most content but do nothing with it.
And so I understand this is content that you are currently watching, and I really want to drive that point home. I feel like it would be dishonest of me not to, and a lot of people that make long -form courses like this just want to keep you in a perpetual state of learning. But I would like you to think clearly and then do more with your life.
If you're just constantly sitting around watching content like mine, you're not really getting anything done. Good news, I have a bunch of actionable techniques that employ and apply principles like, you know, Akrasia Solutions. I'll be talking about this in a sec.
So this is more or less what I did in my very first business. I was walking through the streets of Vancouver in a tattered button up, trying to sell marketing packages at 22 for 250 bucks. That was my offer, 250 bucks to put you on the map.
I had zero sales experience. I knew absolutely nothing about how to sell. But my partner and I walked into between 50 to 80 businesses a day to give them a pitch.
And that was maybe eight hours a day, 48 hours a day, I'd say for over a year. So looking back, I won not because of all the courses I took, because I hadn't taken any courses there, just because I executed relentlessly. And I made sure that the In the chemical example, the reagent that was doing was way higher than the reagent that was learning.
So how do you actually solve acracia? You won't solve it just by learning more about it, obviously. I just wanted to give it a name because now that you have a name, you've labeled it, you can become aware of it.
The real solution is to make the next action so small, so simple, and so well -defined that doing it is trivial. Once you make that next action really easy, what you do is you attach that action to a trigger. Now we're going to do this in the taps section, which is an implementable technique that you guys can use today.
What happens with Acrasia is, you know, usually it's just that vague or ill -defined first step. The step where like, you know what to do, but you're not really sure exactly how to do it. You know, you got to go, you know, pick up the phone and call a number or whatever, but there's like some small little blocker.
So you reduce that down to the smallest possible step. And then you just make it an if -then rule where it's like, if I get back to my desk, I pick up the phone and I dial. So this is how you do it.
This is how you actually avoid that persistent feeling of, man, I got to do this. I just, I still, I'm putting it off for some reason. Why am I putting it off?
Once you start talking in rules, as opposed to like willpower and motivation, your environment will actually do most of the work for you. And that's why we set up the hardware half of our system beforehand, because it was important for you to understand how environments work. So a good example of a rule might be, I will make at least one sales call on the phone every time I get back from a bathroom break.
What that means is when you get back from the bathroom to your desk, you sit down, click on a Google sheet, and then you just call the next person on that list and boom, you're done. This is an actual rule that I used to use in my own life. And that was before I even knew about implementation intentions and taps.
But it was great. And what's really cool is when I get back and I make the dial, usually I'm like, oh, that really wasn't that hard. Let me make a couple more just in case.
And now I go to the bathroom four or five times in an eight hour session. Okay. Talk to my doctor about that.
I don't know if that's a kid or bad, but I go to the doc, you know, I go to the bathroom four or five times. Now I'm doing, you know, five dials each, which is an additional 20, 30, maybe dials that I otherwise wouldn't have done. Obviously that momentum spills over in other areas of my life.
So that's it. I'm going to go way more in depth on taps later, including defining the acronym and so on and so forth. But for now, just know that making actions super small, super simple, and defining them well is the key to overcoming the I know I should do X thing, but I cannot.
Problem that I think a lot of people suffer from, also known as akrasia. Okay, some action questions here. First, what is something you know you should do right now?
Second, what is the simplest possible first action towards completing that task? I think both of those questions have answers that probably came up immediately. Let's talk map versus territory.
The map is not the territory is a phrase I first encountered eight years ago on a random blog website. And if I'm being honest, More value has been extracted from that one line than the vast majority of the books that I've read in my entire life.
And the crazy thing is this phrase was coined back in 1931. My earnest belief is this is a mental model that every human being on earth should know. And if we did, we would probably be far better off for it.
GDP would skyrocket, unity of nations, and so on and so forth. Okay, so here's the crux of what map is not the territory means. You know what a map is, right?
It's obviously something you look at to move you through a territory. Maps are not territories themselves. If they were territories, they would be really big, right?
And then they would cease to be useful. So inherently, the very fact that a map is a compressed version of reality makes it useful. The smaller a map is, eventually it stops being useful because it's so small you probably can't see it.
But the smaller a map is, the more portable it is. probably the better it is, despite the fact that the map gets further away from the territory. It's getting more and more different from the territory.
For instance, there's a big lake over there, there's a park over here, and there's some mountains in the background. I'm in Kelowna. So if I'm looking at my map and there's a big lake, and then there's the mountains, and then there's the grass over here, if this ends up growing so much that it is actually the size of the lake, the mountains, and the park, obviously it ceases to have all value.
The whole value of the map is in that it's small, right? So inherently, .
That's just how it is. Unfortunately, a lot of people see maps as territories. They take a map of something, which is a compressed version of reality, and they assume it is equivalent to the underlying territory.
Maps compress the world. That's exactly why they're useful. But despite the utility, a map is always a version of the truth.
Objective reality will always be different. Just like how a 2D map on Google Images or something is different from our 3D world. The map is not the territory.
The unfortunate truth is everything in life that you believe is in a way a map. Your ideas about what your customers want is a map. Your understanding of how your body processes drugs like caffeine is a map.
Your new understanding of how cognition works and attention, working memory, and executive function is a map. And the story that you tell yourself about why some product that you launched last week failed is also a map. All of these are lossy compressions of the truth.
They're compressions of the truth that actually lose some detail in doing so. none of them are the actual thing. In business, we get a lot of maps.
So you see dashboards with click -through rates, retention, monthly recurring revenue, pipeline values, CRMs, etc. But all of these are equivalent to the instrument readouts in an airplane panel. They might be a good map of truth.
They might tell you sort of where things are, what direction you should walk in. But at the end of the day, these analytics, this MRR, retention, click -through rates, etc., these are just a number. The problem is that maps and territories often diverge without you noticing, but you assume that they're the same thing.
So if you're steering by the map and the map is off, you may never realize it until it's too late. A good example of this is one I see quite often. You know, you have a CRM, which is a customer relationship management service, which tracks customers as they move through your pipeline.
And maybe it says that you have six figures or over $100 ,000 in your pipeline. Well, you would not believe the number of people who have CRMs who I double check and audit as part of our work that have, I have over six figures in my CRM, but those leads have all been gone or dead for the last four or five months. Or maybe, you know, retention for your product looks really, really great.
And you're like, oh yeah, my churn's really low. But if you think about it, the members that are leaving are all the best members because they're very disappointed in maybe what you're doing. These are all just short -term compressed maps, which guide you.
But they aren't the actual thing underlying itself. So here's the territory. Here's what actually happens.
I always do things in business analogies because I care most about business and being economically impactful. So I see customers, sales, calls, etc. This is what you look at, which is a much more compressed version of what that is.
A major chunk of all of the problems that people face in business and in life just comes from assuming that their map of the world is actually the territory. So in a struggling company, Rather than going out on the field to see what's going on himself, a CEO might kick back, relax, and be content with information from his direct reports.
You know, the direct reports say, ah, things are fine. Yeah, people love our product. No, everything's going well on the marketing end, and so on and so forth.
And the CEO goes, yeah, okay, that sounds good, and then resolves to spend his time on other things. What the CEO has forgotten here is something crucial. His direct reports, all of them, are just little maps of reality.
As a consequence, every one of these maps is assessing information in their own highly compressed and lossy ways. For instance, these people may have their own incentives, which could be providing positive information. They could also have their own biases, which could be not understanding that their own maps are not perfect representations of information.
Or maybe they have skills in one area, but they lack it elsewhere, etc. The point that I'm making is, despite all the metrics looking great on paper, this CEO's company is failing. he doesn't understand why and if you don't have good maps or if you don't understand the divergence between maps and the territories you tend to suffer situations just like this this is why the fastest and best growing companies of the last couple of decades tend to be very flat The people in charge get to see exactly what's going on because they're not at the top of some big, long organizational hierarchy that's as tall as the Empire State Building.
No, they're actually on the ground. They can literally look and see the customers at their feet communicating with the business, and they can talk with them themselves and ask them questions like, why didn't you buy our product? This is how you do things in reality when you want to be very effective, both in thinking clearly and then making clear decisions.
Another great story about how this manifests in lab settings, which I learned myself when I was working in labs, to measure cognition, or the ability of a rat or a mice to think, researchers often time how long it takes them to solve a maze. So they'll put together this really cool maze, and then they'll see how long it takes them to solve it, and then they'll see how long they can memorize it and stuff like that, and then they'll usually move the maze around when they introduce a new task.
In one such experiment, they changed the feeding schedule, which is the times that the rats were fed. And then that miraculously resulted in the rats solving the mazes way faster. The researchers rejoiced.
My God, we found a way to make rats smarter. What if this extends to humans too? We could have a revolution on our hands.
Everybody could be 30 % smarter. And yeah, on paper, the rats were indeed solving the mazes faster, right? The map was telling them that rats were actually doing better and getting smarter.
Now, were they actually getting smarter? No. In reality, they were just getting hungrier.
And so... rat getting hungrier is different from a rat getting smarter, even though both of them will probably try and solve a maze really, really quickly. Unless you were actually there at the experiment, you probably would not have known the difference.
So if you were constantly just taking in data, maybe looking at the studies or something like that, there'd be something that you'd miss. So why does this matter? Because the territories ultimately provides you clarity.
The map is just a simplification. And that's not to say the map is not useful. It is useful because it is not the underlying information.
Imagine a map blown out to be the size of the region it represents. Not exactly valuable, right? So the two will never be 100 % the same, and that's okay.
But when they do differ, you should remember that the territory is always right, and you should aim to actually look outside and see the territory for yourself. The fundamental rule, and one that I recommend you feature prominently in your own life, if you ever find yourself in a situation where your map and your territory are at odds, the territory wins every time.
So the next time you feel that something is off. or maybe if you get a vague sense that something must be wrong, do not check your map. Check the territory.
Look for the underlying reality behind the representation of or abstraction of reality, and you will get far more out of it. Do not take people's opinions at face value. Do not take stats at face value.
Do not always defer to experts or consultants or maps that supposedly solve the problem for you. Actually go and look at the territory for yourself a couple of times. And I don't mean this to say you should not take stock in what people that are better than you at a thing.
say, obviously. What I'm saying is think about it for yourself and then go and look for the ground truth data that underlies all of that stuff. Do not listen to me saying, hey, you got to do XYZ thing for your business or whatever.
Actually try and test it in the market and see. And if the market likes it, great, you've got the answer. And the answer is probably better than I am as a map to guide you towards things anyway.
So if you guys want to go deeper, I've linked an article here called Map and Territory. For the rest of you, I would just start trying to observe the rat in the previous experiment. Now, I personally work with AI agents quite often, right?
That's where most of my brand at this point is. That's where I make most of my money implementing in companies and so on and so forth. And the outputs of agents are, by definition, maps of a map.
They're summarized transcripts, digests of my inbox, and so on and so forth. And so I actually have two layers removed from reality here. Though it costs me a lot of speed, I will almost always check the source material before making really expensive decisions based on AI data.
You know, A summarized transcript oftentimes will tell people that the meeting that they just had went great. But if you open the call recording, you're actually able to see the client expressing concerns or hesitations or the lack of alignment or maybe something on their face suggesting that they don't really like what they're hearing, etc.
So rather than focus and trust maps or stories like, oh, the client left because of pricing, I would actually every now and then go and see the underlying data and just verify, you know, is this map still representative? Two more questions. Which map do you look at most in your business or your life?
And when did you last look at the underlying territory behind that map? Next, pick one story you tell about why something is not working. Like maybe people don't want my product or I'm really, really bad at studying.
What would you have to observe directly to actually test this? Now, have you guys seen the Odyssey? I want to talk for a second about Ulysses contracts.
High willpower people do not try and fight temptation. What they do is they just avoid temptation in the first place. But what if you can't avoid it?
What if your temptation is in your pocket all day, like your smartphone? Or maybe if it's in your house, like your fridge, which contains juicy goods. And for whatever reason, you just always have to have them around you.
Well, that's where a Ulysses contract or a Ulysses pact comes in. A Ulysses pact is a choice you make today that restricts your options tomorrow. For those of you guys that don't know, Ulysses is just the Roman name for Odysseus, which is like the Greek thing.
So if you guys have watched the Odyssey, this should stand out to you. Long ago, Odysseus, or Ulysses, was sailing back from the Trojan War. On the way, he encountered the Sirens, who sang a song so seductive and overwhelming that anyone who heard that song was compelled to steer their ship onto the rocks in an attempt to reach them.
Now Odysseus knew that while he waited and wanted to hear the siren song, he also understood that when he did, his future self would beg to steer the ship into the rocks, thus killing his whole crew. So before he approached the sirens, he took precautions. He had his crew seal their own ears with wax and also ordered the crew to tie him securely to the ship's mast.
Then he commanded his crew to ignore all of his pleas to be freed no matter how strong until they passed the sirens. When the siren sang, Odysseus begged and pleaded for his crew to set him free, but they did not. They simply continued rowing, and Odysseus survived.
So, why do I bring this up? The important thing here is not the rope, but it's the timing. Odysseus made his decision while he was still rational, far before he was tempted by the siren's song in the first place.
The way that Ulysses' contracts, or Odysseus' contracts, whatever you want to call it, work, are typically, you make a contract or a pact with yourself while you are sane, and very high performing. And then when you are, you know, temptable and weak, maybe later on in the day, let's say, or when you're put in front of a situation like in front of your fridge or something, you execute that contract.
So for instance, you know, here's a quick graph. At 7am, let's say, you have next to no temptation and your resolve is very strong. But maybe at the end of the day or something like that, you know, your temptation is very high and your resolve is small.
What a Ulysses contract is, is it takes your resolve at the beginning of the day when it's really strong, and then it just applies it across the rest of the day. Basically fixing or holding concrete your resolve or your willpower. Okay, let me give you an example of how almost every willpower failure you've ever had has probably played out.
You know, in the morning, you were clear headed and really fresh, you knew the right choice. So you just avoided the bad thing, whatever it was. But at 9pm, you know, you got home late, you're exhausted, you're tired, you've been faced with temptation all day.
So you see the stimulus, and then you go, Couldn't hurt any given. When I put it this way, the solution is pretty clear, right?
The way to solve this completely is just removing the 9 p .m. version of you's ability to make the decision to do that thing in the first place. Maybe you need it at home.
That's fine. But you can still find and implement ways to minimize the amount of freedom you have later on in the day. Maybe what you do is in the morning, you know, there's a bag of chips or something like that in the eyeline.
When you wake up, you take that bag of chips, you literally put it in front of, you know, or inside of a chest or a safe, scramble the combination lock and give it to your wife or your friend or something. So that later on at night when you come home, you can't do it. Sounds really dramatic and overplayed, right?
But stuff like this is legitimately how some very high performers get things done. And it's how they take advantage of clarity of thinking early on in the day and then apply it. robustly across the rest of the day.
Okay, so what are some practical Ulysses contracts? Here are three easy examples you guys could probably implement today. The first is putting your phone in another room.
By forcing yourself to physically walk to get the phone, you are adding a layer of friction, which I will cover more of, between you and the phone. And you will also buy a rational, more sane version of yourself 30 seconds to stop walking towards the phone. That's why I personally like doing this.
Accountability to others. Now, I record videos every day. And on some of my channels, I actually tell my audience, hey guys, I publish a video here every single day.
If you think about it, this leads many thousands of people expecting a video from me every 24 hours. The fact that I've signed on to this massive expectation, almost like this informal contract with people that watch my content, is a big motivator for me. Because even when I feel like I have zero motivation to record a video or, you know, I don't really want to be sitting here for another two hours finishing this puppy up.
I still do it because I have an implicit contract that I've signed on to with my audience to do something that I signed on to when I was very clear -minded, which maybe I don't want to do now, but that's okay because I'm copying and using the resolution from earlier, you know, despite the fact that I'm tempted to stop. So you can do this too.
You know, if you guys create media, just announce your publishing or work schedule with your audience. Super easy. If you guys are, you know, working out or something, meet a couple people in the gym and tell them, hey, I'm going to be here tomorrow.
That way, you'll feel held accountable and you will want to do that so that, you know, they don't shame you when you walk in three days later. A last way of doing this is through financial stakes. What you can do, and what I've done before, is prepaying somebody and having them donate your money to a direct competitor or some sort of charity that you don't like.
If you think about it, what is more annoying than having to give money to a direct competitor of a company of yours? That is a massive pain in my ass. I don't want to do that at all.
Still, I did, and it motivated me to counterintuitively... Run that competitor basically out of business. So yeah, anyway, a good Ulysses contract can include all three of these.
I use Ulysses contracts not only to improve my cognition, but I treat it as like a loan of high quality cognition from before to now. I'm basically taking money and then I'm setting it aside and then I use that money in the situation when I have to. So it's almost like a savings account.
Some action questions here. Name a behavior that you keep deciding against, but the tempted version of you continues to do. Can you come up with a Ulysses contract that solves this for you?
Let's talk a little about opportunity cost. There's a saying that I like a lot, and that is, what is seen is seen, and what is not seen is not seen. This comes from Frédéric Bastiat, who was an economist in the 1800s.
He was telling a story about a shopkeeper's broken window. He lived in, I think, a village or a town, and basically a shopkeeper had had their store broken into, and so the window was broken. Now, it came to the attention of the town.
And a bunch of people were asked their opinions on this and what to do. And interestingly, a bunch of people in this town believe that since a broken window meant that there was more work for the local window maker in the town, I think they were called a glazier or something, that implied that all else held equal, a broken window may have impacted the shopkeeper, but then it made the window maker's business a little bit better, meaning it was net neutral.
The logic goes something like this. Window gets broken, shopkeeper loses some money, but... the money goes to the window maker.
So the window maker makes some money, meaning the net money has not changed, despite the fact that the guy had his window broken. So this was the prevailing notion of economics at the time in the 1800s or so in Frederick Bastiat's town. Very advanced stuff, clearly.
The groundbreaking discovery was that Frederick found that since the shopkeeper had to pay for a window repair, he was now unable to buy shoes or books or whatever else he was planning on spending that money on otherwise. This is the unseen part of the transaction. And it's what led Frederick to deciding that every choice actually has two sides, the seen side and the unseen side, and realizing that human beings' brains only record the seen.
That's where he gets this quote, what is seen is seen and what is not seen is not seen. And so in business and in life, this is a problem. Because if we were only ever understand the seen, we would fail to account for a notion of opportunity cost.
Now, put simply, opportunity cost is the value of the next best alternative that you could have taken in a situation. It's not just the sum of all other alternatives. It's just the one thing that you could have done otherwise that would have delivered you the biggest result.
So if you spent your Saturday building a landing page for a website of yours, the opportunity cost is whatever the next most valuable use of your time would have been. So maybe what you could have done with that time instead is you could have spent it on a sales call. or you could have spent it working out, or you could have spent it taking a nap.
If you think about it, all of these have some non -zero expected value to your life, right? But you didn't. Instead, you decided to work on your landing page.
And despite the fact that you worked on your landing page, it's not just that you worked on your landing page. It's that you worked on your landing page, but then you said no to a bunch of other stuff. So in this way, the sticker price of things is rarely their true cost.
A $30 lunch, for example, is not just 30 bucks. It's also 90 minutes. So it's not enough to model the cost of the $30 lunch as price, which is 30 bucks.
You must model the cost as the price plus the number of hours times your hourly rate that you could have made doing the other thing. This and only this will give you the true cost. So let me give you an example of this.
You know, if you were naive and you didn't understand opportunity cost, you would look at the lunch and you would say, oh, that cost me 30 bucks. Well, I guess that's not that bad. But if you took an enriched view with opportunity cost included, you would add the $30 price of the lunch.
to the amount of time you spent on lunch, which would be 1 .5 hours, and then multiply that by the hourly rate that you could have made if you spent it on some beneficial activity like prospecting. And then here, you would have gotten a formula that looked less like $30 total cost and more like $30 plus 1 .5 hours times 45 is equal to $97 .50.
So from an opportunity cost standpoint, if you think about it, your decision to go and take this lunch was actually... three times the amount. And that brings me to an important point.
You know, most of the poor decisions that we make are always found in the discrepancy between price and cost. And this was demonstrated in a 2009 piece of consumer research as well. People tend not to account for opportunity cost unless it is explicitly pointed out to them.
When asked to buy a really cheap item in this study, people were more likely to say no if the don't buy option was instead described as save that money for something else. So literally just changing again, the name of the freaking thing, you know, in this case, the button was enough to reduce the total number of people that said yes, which obviously has dividend can pay dividends if you are the one that is controlling that decision, because you just do the opposite.
So despite the fact that the money was the exact same, just the way that was framed, just the whole like, wait a second, you could actually spend that money on something else was enough to make people buy less of it. These people aren't stupid. They're just not accounting for the alternative because the alternative is never top of mind.
And because, you know, human beings did not evolve with a built -in notion of opportunity cost, right? There's something that Mr. Frederick Bastiat had to figure out himself.
To that end, human beings routinely neglect opportunity costs unless you spell them out explicitly. It's probably one of our most major biases, and it's worth just constantly reminding yourself and keeping in mind. So the cost of a choice is the road not taken.
Let's say you do something on Saturday. and it's uh you know or 500 bucks or whatever what you see is the thing that you chose to do you spent your saturday on some activity but you forgot to count in the equation is the best alternative which is not seen so every time you say yes to something you are always saying no to something else nothing in life comes without a sacrifice as unfortunate as it is to frame that way Now, for instance, in my own life, I was walking into 50 to 80 stores a day to pitch a marketing package in Vancouver when I was 22.
And that looked really productive, looked like I was doing a lot. My choice was like, well, I could go prospect or not prospect. So obviously I'm going to prospect, right?
And if you look at the price, my work was nearly free. I mean, I did a few bucks in gas per day. Thanks, Grindr.
Maybe. And then lunch from time to time. But the real cost was the hours I was not spending on developing a more scalable outreach mechanism that could do the same job thousands of times faster.
And that is an opportunity cost I did not understand for many years. And I'm not saying that I regret it, obviously. I'm very happy to have spent all that time prospecting.
I gained a tremendous amount of grit and resilience that was probably worth it. But I do wish I'd picked up more scalable lead generation mechanisms a little bit earlier. I think I probably would have gotten a lot farther.
So every yes is a no. Once you guys see that cost, anytime you say yes to one thing, you're actually saying no to something else. Your life will probably change.
For example, every time you say yes to some sort of mindless internal meeting at work means you are saying no to the $200 you could have made by working on perhaps your sales. Likewise, saying yes to client number one at a really low MRR logically means saying no to all the other clients you could have gotten. Saying yes to a side project means saying no to the time you could have spent on your main project, and so on and so forth.
In practice, the no is always there, which should just change the way and the manner in which you price. Also, in this way, cheap things can end up being incredibly expensive. For example, scrolling on social media for an hour might be free.
But if you do that every day during your peak production time, it's costing you hundreds, thousands of dollars a week. And at the same time, expensive things can actually end up being bargains. Like maybe paying $200 for somebody to come and clean your house might seem pricey.
But if the $200 you make, you get back makes you over $200. Rather, if the two hours that you get back makes you over $200, it's a great deal, right? So again, no more specific numbers during the section on hourly rates, but this is the general idea.
How do you actually use it? It's simple. The next time you say yes to anything that's ill -defined or consumes your time, you should explicitly state not only, yes, I will go do the thing, you should state what you are saying no to.
So I don't mean time in a vague sense. I mean, you should explicitly state the specific thing you will not be doing or could not be doing because you said yes to this other thing. A good example, if you're about to say yes to a 10am call from some team member that wants some bullshit clarification about something that could have been an email, just say, I'm saying yes to this 10am call.
And in doing so, I'm saying no to writing my money making script or whatever the hell at 10am. That is a simple and easy way to just ground you in reality and constantly remind you of the cost of your decisions. So some action questions for you.
The first is what is the last thing you said yes to? The next is, what was the best alternative use of that time, specifically? Would you make the same call understanding opportunity cost now?
Next up, I want to talk about Goodhart's Law. Charlie Munger, the billionaire investor who was also Warren Buffett's right -hand man, said, show me the incentive and I'll show you the outcome. The funny story is, for three months I thought it was my podcast co -host Jack Roberts that came up with that, so fool me once, buddy.
But there's no better predictor of human behavior than our incentives. They are far more important than any values that anybody says they hold or their individual personality traits. And this all makes sense when you think about the purpose of this course, which is that people are animals and animals act typically in ways that are rewarded like the rats in my lab.
If you give a rat a button to press, and if you reward it with food every time it does so, you know, it'll happily press the lever. It doesn't really care what your intent is or why you're doing it. The only internal incentive is, press button, I will get tasty food.
Once you understand this, you'll also understand if the rat doesn't do what you want, it's not really the rat's fault, is it? It's just responding to its incentives. As you are the more cognitively capable one, it's not the rat's fault, it's yours.
Because you knew its incentives and you used them wrong. So humans are the same, we just use words. An example from business.
If you pay your support team based on how many tickets they close, logically they'll close more tickets. But they'll do that even if the tickets themselves are poorly done. If you pay your sales team based on how many calls they'll book, they'll probably book more calls, right?
But they'll start doing so even if they're unqualified. This doesn't make your support team or your sales team bad people. They're just reacting to the environment that you place them in.
And that is the heart of a law called Goodhart's Law. So Goodhart was proposed initially by Charles Goodhart. He's an economist in 1975.
He said, when a measure becomes a target, it ceases to be a good measure. So some examples of what we've been discussing, a really twisted one that I think you guys can keep in your mind as you proceed. In Hanoi, in the Vietnam sewers in 1902, you know, the French had taken over.
And so the French colonial government at the time had a massive rat infestation problem. So what they started doing is offering one cent per rat tail to hunters who would go out, cut the tails off rats, presumably kill them and then bring them in for money. Predictably, lots of rat tails flooded in.
But despite all of the rat tails, the rat population didn't go down. Why? Because hunters began to cut off the rat tails, but they would release the rats back into the sewer to breed.
And some would even start breeding rats on purpose. So what does that mean? The tail counts were a good measure of dead rats, but when you tied them to an incentive, they stopped being valuable.
So basically, you know, when the measure becomes a target, the measure is removed from the target. If the black is what you actually wanted here, And if at some point you decide to incentivize it, eventually what will happen is people will care more about achieving the measure or the metric than they will about the thing that you actually wanted.
And as a result, the thing you actually wanted tends to go back down. That's what that means when a measure becomes a target, it ceases to be a good measure. So understanding this will help you think far more clearly.
It'll also put, I think, into words a lot of incentive -based behavior you've probably seen. For example, the phone that you're using, dozens of features are good examples of Goodhart's Law here. Like notifications are on by default or autoplaying videos or infinite scrolling or streaks or red notification bubbles or whatever the hell.
None of these are effectively beneficial. What happened is a product manager probably just had a metric put in front of them that they're optimizing for, like daily active users or time in app. They go to a meeting once a week and they get rewarded based off of their time in app or daily active users.
And so they wanted to win. What they did is they set all these defaults to be the way of measuring and manipulating your attention. The end result is you now have to deal with a lot of stupid bullshit on your phone.
A good personal example for me is views. Views are a proxy for the goal of my YouTube channel, because if you think about it, the more views I make, the more impact, and ultimately my goal is impact. So in this way, my view metric is a map of an underlying territory.
But I constantly need to remind myself that views are not the end goal myself. My real goal is for you to watch the video until the end and actually use it, right? The issue is when I focus too much on views and fall prey to good -hearting myself, which I've done before, I end up making decisions that optimize for superficial shitty views at the cost of depth, and so my watch time tends to go down, which is not a good thing.
Despite my best intentions, every few months view maximization behavior creeps up because the measure is becoming a target, and I always have to quash it down. And if you think about it, that's coming from me, who's somebody that has studied this for almost a decade now. Imagine how stupidly someone who has never heard of good hearts law or has never even reflected on their own behavior is probably being driven by this right now.
So how do you avoid getting good hearted or good hearting people? First, if somebody does something you don't understand, just ask yourself who benefits from this. Usually, if you ask yourself who benefits from this and you see who benefits from this, the answer becomes clear.
Next, before you put a target on a metric. which is useful for project management. First, say what it is a proxy for, and then connect it to a countermetric that gets worse when the first one gets better.
So this doesn't make sense to you. It's a little hack. Consider, how about speed versus error rate?
Or tickets closed versus the number of tickets from those closed ones that get reopened because they didn't actually work? Or how about views versus, you know, retention or watch rate or something? If you think about it, if one goes up and the other goes down, it's a sign that you're gaming the system and the underlying metric probably isn't improving.
How about like in my case, impact? I want views and I want watch time, right? But if my views go up and my watch time goes down, what it means is I'm making superficial content that chases views at the cost of my watch time.
So my impact has gone down. So remember map and territory? When the number itself and reality disagrees, reality always wins.
Okay, some more action questions for you. What number do you guys check most often and why? Has the number itself become a target rather than the thing it represents?
What counter metric could you add to show that it is being gamed? There's this lovely idea of Chesterton's fence. And it was, I think, initially introduced in 1929.
And it goes like this. Two people, reformers in this case, are walking down a road. They come to a fence.
One of the reformers wants to tear the fence down because he doesn't see the point of it. But the other suggests that he finds the point of the fence out and only then actually tear it down. I presume one of these fellows was called Chesterton because this parable has been taken to be called Chesterton's fence.
As a systems person, Chesterton's fence is simultaneously one of the most interesting and also most annoying things I've ever witnessed. At its core, it suggests the following. For every fence, which if you think about it in business terms, is a rule, an order, a legislation, a law, a workflow, a process, an SOP, anything like that.
There must have been an investment at some point in time of time, energy, or money into the thing. People in general conserve all these things, right? So there has to be a reason why a fence was created.
And it really just says, before tearing something down, you should understand why it exists. For example, in business, if you have weird steps in a spreadsheet or a process, if you have a strange client approval requirement checklist, if you have inexplicable steps that you do not understand in an SOP, as a newcomer to a system, You know, you come in and you're like, well, why the hell are we doing this?
It's obviously very tempting to cut through all the red tape, right? And then remove all the stuff you don't understand. It makes you feel decisive, you know, like you're improving things or making moves.
There's a lot of situations like that where a new hot chat manager or CEO comes into some like pre -existing team and then immediately starts demolishing all the existing structures because they think, well, you don't need this, you don't need that. But if you think about it, a lot of these weird steps were the result of rather expensive deliberating.
And this occurred far beyond before you came into the picture, right? When you remove these steps, what you're doing is you're forcefully rerunning an experiment that led to all of these conditions occurring in the first place. Except now you are the one who is suffering the consequences.
No, not saying you 100 % have to respect every tradition, rule, SOP, or thing that your predecessors did.
But what I am saying is treat the fence like a map. Just like a map is a compressed and often useful example of a territory, a fence is a compressed and often useful example of a process. Think about it in your own life.
How many weird fences or things that you do seem inexplicable or strange to other people? Probably a lot. A good example is all the places you put all your different kinds of belongings.
So how about your food, your clothing, your keys, your pantry, your goods, etc. You know, if another person were to come into your home and look at all these things with no context, do you really think they wouldn't spot at least one thing that would make them go, what the hell did you do that for? Like, for instance, why are all the plates up on the upper cabinet and not on the more convenient lower cabinet?
That seems weird. You should move them to the lower cabinet. Well, actually, there's a reason for that.
When you stack more than four plates, they start catching the door, you see? In this example, which is actually real, in my own damn house, frickin' cupboards, you are the one who put up the fence, and you did so immediately after an event. In our case, closing the cabinet.
That taught you a lesson, which was sound, pain, etc., which is why you did it, right? So how do you actually deal with it? Here is the workflow.
If there's a rule you want to remove, figure out why it's there. If you figured it out, you can now decide on the merits and either keep it or remove it. If you do not know why it's there, first, ask the person who built it.
That is ideal. If you can't ask the person who built it, check the history. Then, if you still don't know whether or not it is required, try removing it and then see what happens.
If it's not irreversible, you unfortunately have to leave it. So that's it. Just don't take anything down until you understand why it was there in the first place.
Doesn't mean you have to keep the fence forever. It just means that if you want to reason clearly through how best to optimize the process, the fixed order of operations is understand the fence first, decide what to do after. Some ways you can figure this out.
Ask the person who built the fence. Look through the document history. So Google Docs, Drive, etc.
is great for this in modern corporate environments. Check the Git log if you guys are doing code. go through old email threads of SOPs or processes.
You can also just think hard what would happen if the fence wasn't there for like a week or a month or a year. And the crazy thing is most of the time you can do everything I just talked about in like 10 minutes. These inexplicable processes that you do not understand, you can actually understand.
And a lot of the time you'll find that, you know, the fence, whatever it is, was necessary back in 2021, maybe no longer required today because of some new technology. Once you're done with that, you can confidently remove the fence without any guilt and with minimal slowdown and without like really making people hate you while also eliminating risk.
This is something very important that I had to learn the hard way as a trigger happy young man who started basically branding himself as like an operations consultant and then eventually like a fractional COO who would remove a lot of fences that probably did not need to be removed. So hopefully this helps you like it helped me.
Some action questions for you. The first is. Which rule, step, or habit have you been meaning to delete because it looks pointless?
First, write one sentence on why it was put there. If you cannot write one sentence, who do you know that could tell you, and how can you ask them? Next, which fence did past you build for yourself that present you has stopped respecting?
And what was the specific event that made you actually build this thing in the first place? Next up, I want to talk beige frequencies.
No, let's not do that. Next up, I want to talk Bayes' theorem. It's one of my favorite thinking tools, and at its core, it's a very simple idea, which is that you start with a rough guess about how likely something is, which is called a prior probability, and then you update that guess as you get new information.
Now, it is really simple conceptually. Rather than deciding that things are true or false, What you do is you treat everything as a probability, and you say that has a 95 % chance of occurring this way, and maybe a 5 % chance of occurring that way.
Nothing is 100 % or 0 % anymore, it's all a realm of probabilities. For example, using Bayes' theorem, you can expect a very high probability that the sun will go up two days from now. How?
Well, you'll start with a rough guess about how likely it is, which is the prior probability. Given that the sun has come up virtually every day of your life, probably every day of your life, the probability is really high, right? It's like 99 .999%.
Then after that, you update it based off new information. Maybe you'll wait a day and you'll see if the sun comes up again. If so, now you can update your probability from 99 .9 % to 99 .99999 % or something.
Unfortunately, most of you will never have heard of it. In practice, the reason why is because it's presented in a very annoying mathematical format that I think just like most people would never understand unless somebody put a gun to their heads, which is ridiculous, if I'm honest, because it's probably one of the simplest, most generally powerful reasoning tools and is inaccessible to most of the population because a bunch of mathy bullshit.
For instance, my girlfriend's a doctor. They discussed Bayes' theorem with her in school. I got really excited when I heard about this, thinking, hey, we'll get to estimate probabilities for all sorts of things.
But when I asked her about it, she gave me basically a long ass formula based on conditional probabilities, which sucked. And it didn't really explain any of the beauty or intuition behind it. Again, this is something that I personally think should be probably taught in like grade school.
Unfortunately, nobody's born with the ability to intuitively calculate conditional probabilities, which is what Bayes' theorem is all about. Even after five years in research labs, I don't think I'm any better than you are at doing them in my head. So I'm not going to even talk the formula here.
The point is more intuitively. that you need to understand that everything in life is a series of probabilities. So before we look at a specific example of Bayes' theorem, I want to instill a core concept that we will use to determine whether a piece of evidence is important.
What is the likelihood that this evidence would occur if X were true versus if X were false? This is the question that is probably the most effective one that you can answer. It's very important to wrap your head around, so let's think about that sentence deeply.
What is the likelihood that this evidence would occur if X were true versus if X were false? For instance, imagine you are selling something and at the end of the call, the prospect says, this looks great. Can you send me over the details?
Most people assume this is strong evidence that the deal will close, right? And it feels good. So they'll hang up and they'll tell their partner, this went really well.
But I want you to think about the question. What's the likelihood a prospect says, this looks great. Send me the details.
if they're actually going to buy. Well, it's very high, right? Almost all buyers are going to say something like that, maybe 0 .9.
But what's the likelihood they'll also say it if they're never going to buy? Guess what? It's also really high, maybe 0 .9.
Because if you think about it, that is the politest possible way to end a call, and it probably matches your own buyer behavior as well. And that's the real crux of it. Since the sentence shows up in both of these hypothetical worlds at roughly the same rate, simply hearing this piece of information will tell you absolutely nothing.
For instance, it is very weak evidence of a sale, and it is merely evidence that your call is ending in all reality. Now, compare this to somebody that says, interesting, are you able to invoice us quarterly instead of monthly? If you think about it, how often does that happen when somebody's going to buy?
Probably pretty often because it's discussing payment terms, right? Maybe 0 .9. How often does it happen when they aren't going to buy?
Well, pretty much never because nobody cares about negotiating payment terms on a product they have no intention of purchasing, right? I mean, that's just like behavior 101. It's extra work and we'd all rather just get on with our lives and I'd rather end the call and go, you know, bury my head in the sand.
So maybe the odds of that are 0 .1. So what that means is if your prior probability for a close was 0 .25 or just 25%, because that's just how much you guys normally do. If you hear the former, which is that, you know, hey, do you guys invoice quarterly instead of monthly?
You can immediately update that to like 0 .9, which is now 90%. And if you think about it logically, that probability will now change how you interact with that prospect. And in doing so, become a self -fulfilling prophecy where you're much more likely to actually close that client because you are interacting with them differently.
You can also do things like project income, project pipeline value, and so on and so forth. I love Bayes' theorem. It is so hard for me to explain how impactful Bayes' theorem has been to my life and my career.
So let's do a couple more examples. First, a word of advice from somebody that's used and abused Bayes' law for the last few years. update in small steps, and ideally you would do so one small piece of evidence at the time.
Good Bayesian updating is all about chaining multiple steps together, not just one. And so to do so, instead of percentages, we use odds. Odds are very simple.
10 % probability just means your odds are 1 in 9. So it's not 1 in 10, which I know sounds weird, but it's 1 in 9. 1 chance of it being true, 9 chances of it being false.
10 % just means there's a 1 in 10 chance, right? Well, it's not. It's like there's a 1 to 9 odds.
So we can chain our previous example together like this. Let's say you have a 1 in 100 odds of closing a deal on any random prospect. Realizing here, I should probably say 1 in 99, considering I absolutely explicitly just explained how that works.
So there's a 1 % chance, 1 % chance or 1 in 99 odds of closing a deal with any random prospect. That means that your prior odds, if you think about it, are 1 to 99. Okay.
Basically in one out of a hundred cases, you will make money, right? Which means out of a hundred, one of these, that one will occur. Now what happens is you get a very quick reply from one of these people, which makes your odds 2 .7 times more likely.
If you do the math on this, okay, multiplying them out, that's 2 .7 to 99 now, which if you do the math on that is 2 .7 divided by 99 or 0 .27%, which I guess is still the same as it was before. Now, let's say you get a pricing question from the same prospect. That makes you two times more likely after that.
This puts you at 5 .4%, rounded to five. Now, unfortunately, let's say that same prospect eventually no -showed a meeting. That makes them 4x less likely, or maybe a quarter.
So now the odds are 1 .35 to 99. This is 1 .35 % approximately. Notice how we're not just adding or subtracting.
What we're doing is we're multiplying. This allows us to scale our odds up or down quickly while also still maintaining a strong degree of nuance. The value here is Bayes' Law is excellent for sales pipelines, for probabilities, for projects, and so on and so forth.
What you can do is you can get the average order value of a product, maybe $5 ,000, and then just get the conditional probabilities of every pipeline step based off historical closing. Then you can just multiply them all together to give you the expected value of your pipeline. In that way, you can get very granular about how much money your company is actually capable of making.
And as mentioned, I mean, this is just one of like many, many different things that you can do. Base theorem is a very small chunk, I would say, of all, sorry, CRMs are a very small chunk of, I would say, all possible ways you can apply base theorem. So for instance, updating in small steps, the base rate might be, you know, 10, well, you know, I said one in 99, this is 10, but bear with me here because obviously I generate these graphics.
you know this moved up to 2 .4 x this moved up to 3 .7 x let's move back down to 1 .3 or whatever but ultimately the idea is we're updating in steps we're not just updating all at once now if you guys didn't get the explanations behind base theorem don't worry about it there are lots of great resources and many of them are from less wrong or this guy eliza yudkowsky who's just wonderful at explaining these rationality things so you guys can just click this link and check out more intuitive explanations of base theorem from the ladder Okay, one big question before we move on.
Pick an idea you have about your business or your career, like this product's going to kick ass or this hire is awesome. Then write down the Bayesian prior for it and just ask yourself how often you would see it if the belief were false. Congratulations, you just did some Bayesian reasoning.
Let's talk a little bit about Fermi approximations. And for those of you guys that don't know, Fermi was a renowned physicist. At the Trinity nuke test site in 1945, he had a really cool experiment where he dropped a few scraps of paper during the progression of that big blast wave.
And then he just watched how far they traveled. And then using his knowledge of physics, just did some very rough like back of napkin math. He ended up estimating the yield of the nuclear bomb to about 10 kilotons of TNT.
What's really interesting is the actual accepted figure was just about twice of that, meaning he was off by a factor of two. Which means, like to most people, he was wrong, right? But to physicists, they would look at that and they would go like, are you kidding?
You were able to estimate the size of that nuclear bomb within an order of magnitude, aka 10 times, and you just did it by dropping a couple of random scraps of paper and then doing back a napkin math? That's extremely impressive. And the reality is, this idea of Fermi approximates, of being able to approximate things like Fermi, is an extremely powerful skill that I personally use virtually every single day as somebody in business, and that once you learn, you will probably want to use very often as well.
So I figured we'd probably spend a few moments covering it. So how do you do it? Well, it's easy.
First, you decompose, then you estimate, and then you multiply. Okay, so the way that it works is, you know, you have some sort of problem of some kind. I don't know what the problem might be.
Let's say the number of intersections in your city. What you do is you decompose that problem into parts. You ask yourself, okay, so part number one, you know, an intersection is like a four -way stop, right?
There needs to be people crossing. So how many intersections are there realistically? Well, I guess I need to figure out how many streets there are going north -south, or yeah, north -south, and maybe how many avenues there are going east -west.
So how many streets are there? Well... I guess it would probably be like 50.
And how many avenues are there? I don't know, maybe like another 100 or something like that. So it's like, okay, well, if I just multiply the two out, how many am I getting 100 times 50?
So I'm getting what 5000 or so. All right, that seems pretty reasonable. Cool.
Multiplying them together, I get 5000 intersections or something like that. So that's a very simple and naive example, right? But hopefully you guys understand what I mean, you know, extrapolated, I'm sure you guys could see how you could start with something really bad and then just get reasonably within, you know, probably one order of magnitude of the truth.
And the reason why that's valuable is because it allows you without having to take everything in your working memory and then put it on a piece of paper to like quickly reason through things in your head very impressively. So here's an example I use all the time. How many cold emails do you need to get five clients?
Okay. Well, maybe you get a positive reply rate at about 0 .5%. That means one in 200 or so.
maybe you get a reply to call rate of about 25%. Maybe after that, you get a call to close rate of about 20%. And maybe, you know, if we combine all that, the positive reply rate is 0 .005, the reply to call rate is 0 .25, the call to close rate is 0 .2, maybe in total it's 0 .00025, which if you think about it mathematically is one in 4 ,000.
That means you need 4 ,000 cold emails to get one client. If you want to get five clients, you would just multiply these two to get 20 ,000. You know, the crazy thing about this, I didn't actually use like my historical numbers from maker school for any of this stuff.
But then I went back and found a bunch of threads in maker school. And I looked at like how many emails the person had sent when they said that they acquired their first client. And it was roughly on the order of maybe 2000 to like 8000, I would say.
Which if you think about it, like this isn't. this isn't perfect, it's off by a little bit, but it is quite close to the mathematical average of just some cursory actual information that I gather. And I remember just being like, wow, that is wild.
I mean, obviously I know about the industry, just like Fermi knows about physics and nukes, but he was actually able to get within a few percentage points of the answer.
Fermi approximations are foundational to the other two calculations that we've learned. I mean, they provide probabilities and payoffs and so on and so forth, and that's great for computing expected value. They can also help us estimate the probabilities for theorems like Bayes.
So I think they're an excellent way of reducing a lot of the inherent uncertainty involved in estimating probabilities and impact figures for both. When you don't know, just always decompose, estimate, and multiply. It's a very strong clarity of thought technique.
Here's another example in real life. Let's say, hypothetically, you're considering selling appointment booking software to dental clinics in Canada, which is where I live. Canada's a population of about 40 million, right?
So first, if you wanted to run a Fermi approximation to figure out whether or not it makes sense to sell appointment booking software to dental clinics in Canada, a reasonable thing to do might be to think, okay, so how many people and how many dental clinics for how many people? Like, how many people per dental clinic? Maybe there's like a dental clinic every 5 ,000 people or so.
Well, 40 mil divided by 5 ,000 is 8 ,000 clinics in Canada. Okay, that's fair. Maybe about a fifth of these clinics have a problem bad enough to actually want to pay for a solution.
Well, 8 ,000 divided by 5 is 1 ,600. Maybe I could reach and close a small percentage of those in the first year. Let's just say, you know, 3%, which is 48, which I'll call 50.
And let's say, you know, if I went to market really well and had a big team behind me, I could get each of those 50 clinics to pay me three grand a month for the software. That's 150k a month. Hey, that's actually pretty feasible, right?
Now that we've done the Fermi estimate, maybe we can actually go test the waters by sending some emails or calling some people. What I will do constantly while I'm evaluating business models is I will just very, very quickly do back of napkin Fermi approximates and math. And I'll just determine like, does this check out?
Is there a chance? You know, I'll be conservative across the board, but I will always just be like. Okay, let's just run the numbers really quick in my head.
This sounds reasonable. It's probably within, you know, an order of like two or three. This sounds reasonable.
And this is probably within an order of two or three. And then this sounds reasonable. And this is probably within an order of two or three.
So this sounds reasonable. It's probably within an order of two or three. If I made 50 ,000 with this, that'd be cool.
If I made 300 ,000 with this, that'd be cool. It justifies the next step, which is obviously doing a little bit more research. Okay, so Key takeaways here.
The biggest thing to remember with Fermi approximates is they're just a tool to prevent you from wasting time. So don't waste a lot of time on them. I round really aggressively.
If a number is like a seven, I usually make it a 10. If it's 350, I make it 300. My goal is not to be precise.
It is simply to, you know, figure it out within an order of magnitude pretty quickly. The value to Fermi's is each of your individual estimates are probably going to be wrong, but they'll probably be similarly wrong. And a lot of the time they're off.
by opposites. So like a previous step, like how many people per double clinic is going to be off by a factor of three, let's say. But then, you know, the next step, which is how many people actually have a problem might be off by a factor of three in the other direction.
So all these uncertainties actually do average out and you end up with something quite reasonable. And then what's really cool is you can actually have AI do a lot of this stuff for you too. AI is wicked at Fermi approximates, and I use it often as a first principle of a reasoner.
I will like give it questions like, hey, you know, let's through first principles and Fermi approximates, let's run through X, Y, Z problem and like, let me see what you think. And a lot of the time it'll do just as good as I was capable of doing. And then I can take that estimate and I can use it in the next chain of my process.
Okay, and that takes me to an action question, which is what is a potential project, business or idea you're excited about right now? Fermi it in 60 seconds. Remember, that's decompose, estimate and multiply.
All right, let's now talk the hourly rate and the offload rule. Most people never calculate their hourly rate. They will just implicitly set it at zero.
And that is even if you're the sort of person who thinks that they know their hourly rate. But now that you understand opportunity costs, you should actually sit down and calculate this. Because the moment that you do, it'll prevent you from spending three hours on a task that an assistant or an AI could do for you for $20 or $30.
An example of that might be buying stuff, which until recently I didn't realize was a very poor use of my time. How about doing spreadsheet work or stuff like that? So why don't we start with what's your hourly rate?
Now, first, you don't just need one hourly rate. You actually need two hourly rates. The first is your average hourly rate.
The second is your marginal hourly rate. In case you didn't know, average hourly rate is just your income over the last year divided by the total number of hours that you worked in order to make that income. So here you're going to want to use actual hours.
If you work a side hustle or Uber on the weekends or whatever, you have to include those quick jobs that you do as well. If not, you will probably be misrepresenting what that is. So that's your average hourly rate.
Your marginal hourly rate, on the other hand, is the value of your best work. Now, when I say best work. What I typically refer to is work that when you allocate it towards the number one most important thing in your day, whatever the big moneymaker is, sales calls, writing a script, in my instance, I don't know, doing business partnerships or whatever, the amount of money that you would make doing that.
And this is the number that matters a lot for people like us who presumably want to improve our stations in life, because when we offload some task that is preventing us from spending another hour on our marginal rate, we get that entire hour back. Also, usually your marginal rate is way higher than your average rate. So for instance, my marginal rate managing Maker School, which is my 90 -day accountability community, if I spend a single hour on Maker School, I will make somewhere between $6 ,000 to $8 ,000.
I know that sounds crazy, but such is how Maker School works. If I spend a single hour on Maker School answering questions, recording videos, or responding to people, I make $6 ,000 to $8 ,000. but that's not my average hourly rate.
Meaning I can't just work eight hours a day and make 6 ,000 to 8 ,000 times eight. So I don't know, what's that? $56 ,000 a day.
If I made $56 ,000 a day, I'd be making $2 million a month and I don't make $2 million a month. I make about a quarter of that. So the difference there is the average hourly rate takes into account all of the time that I'm spending suboptimally.
Okay. Just the average hourly rate, the average usage of my time while working. Whereas the marginal hourly rate is basically like just that power law chunk that is the highest ROI.
And so ideally, your marginal hourly rate would be equal to your average hourly rate, by the way, because you just spend every waking moment of your time on the most effective thing. But let's be real, most of the time it is not, and it's important to understand that. Okay, so logically, you should offload a task if the number of hours to complete the task yourself multiplied by your hourly rate, your marginal hourly rate, is greater than the direct price of offloading plus the number of hours you need to supervise the task that somebody else does times your marginal hourly rate.
If that doesn't make sense to you, we can look at it with numbers. Now, on the left -hand side here is cost of your hours and dollars. And so over here, for instance, if you were to do this hypothetical task, whatever it is, it's a five -hour task, and your rate is $100 per hour, Well, $5 per hour times, sorry, five hours, I should have said, times $100 per hour is equal to $500, right?
That's just logical. Now, that's if you do it yourself. If you offload it, okay, what you'll do is let's say you could pay somebody $150 to do it.
You'll pay $150, which will be the price, but you'll still have some additional costs and opportunity that you don't typically take into account. For instance, it may take you an additional hour to scope, send, and then supervise the completion of the task. And maybe you'll do that in 20 minute batches, you know, 20 minutes to scope it, 20 minutes to send it off, 20 minutes to supervise it throughout the course of the task.
If you do the math there, then the total amount of time that you just spent multiplied by the obviously price and your hour of supervision is now $250. Meaning that what you've done is you've offloaded a task that was $500 of your own time for a total of $250. Or $150 plus one hour of your time.
Okay, so naturally, if you were to do it yourself, that would cost $100 or $500 of your time. Five hours times $100 an hour. The only real, I'd say, error rate here, the issue, is just that supervision portion.
A lot of the time, people assume the supervision portion is basically zero. They'll say, well, you know, I hire good people. I only have good people on my team, so they'll be able to do it.
But the question is not, are they able to do it? It's, are they able to do it as well as you are? And in most cases, they will not be able to do it as well as you are.
And so what you will have to do is you will have to poke and prod and move things around just enough, okay, just enough for that to take a little bit more of your time, reducing the total profitability of that calculation. Nobody talks about that, but I figure I might as well because, yeah, I mean, obviously everything looks like sunshine and rainbows if you could spend $150 to do something that might have taken you $500.
But you do have to take into account you are always now going to have to scope the task, send the task, and then delegate. and supervise the task as well. So hypothetically, let's say you give yourself a little bit of error bars and you say, I might screw something up in the task.
So now they have to send it back for revisions and then I have to do it. Maybe this is some design task or something. So it'll actually take twice as long for me to supervise.
What do we do? All we do is we add an additional one hour of supervision cost to this to make this two hours of supervision or $200 with a price of $150, putting the total cost at 350. Now that is still well within profitability, right?
You're spending $350 to make $500. Obviously, that's fine. Ideally, I'm happy to offload tasks that perform at least a 2 .5x multiple.
So that $150 to $500 is actually basically the precise example of tasks that I would delegate. Anything around there gives me enough of an error bar that I'm okay having to do a little bit of supervision just in case. And it also allows me to as we know through the planning fallacy, it allows me to kind of massage the rate.
I don't think we actually talked about the planning fallacy yet. It allows me to massage the rate a little bit in case my math is a little bit off. So consider the prior probability that you know your exact hourly rate and that your marginal hourly rate is equivalent to the true marginal hourly rate, not just your calculated marginal hourly rate.
If the prior probability is quite high, then there's a high correlation between what you think your hourly rate is and what it actually is, but we still have a little bit of a gap and that controls for that gap. So supervision is the main issue in practice.
In delegation, you know, we'll always require your time to write a brief check -in, answer questions, review outputs, and so on and so forth. So you just have to be honest with yourself about those supervision costs or the task will boomerang back to you. It's just now it'll be half done and half done is almost worse than not done.
Also, just because a task is offloadable, that doesn't necessarily mean you should always offload that task. Some things are how you make your money and what your customers are paying you for and should just not be offloaded. For instance, I respond to every post in Maker School myself.
I also film all of my own videos. The reason why is because I make an insane amount of money doing this, close to about $10 ,000 per hour right now, which my estimate was nearly correct on. What marginal rate would be $10 ,000 if I have to make at least $11 ,000?
And not only would I have to make 11, I'd have to be sure that like, you know, the work that I'm doing and offloading is actually genuinely as good, if not better. I see virtually no way I could do that unless the marginal rate is now 40 or $50 ,000 an hour. So I just can't see any reason in which I'd ever do it.
If I was making $50 ,000 an hour, you know, I could be one of the richest people on earth, right? Well, maybe not richest people on earth, but you know what I mean? $50 ,000 an hour, let's do some Fermi approximates, eight hours a day, okay?
That's $400 ,000 a day or $80 million a month. $80 million a month, 700, well, $960 million a year, that's a billion dollars a year.
I'd be very, very hard pressed to find a way for me to make a multiple on that. Final thing, your time is only worth the hourly rate if you actually use that time productively. Like a lot of people offload $150 task and then they'll just spend their newly freed up time on TikTok or whatever.
If you're doing that, you are basically paying somebody $150 to scroll TikTok. And I have been guilty of this, which is why I do not have TikTok on my phone. So some software here can change the math that we just did.
especially today with AI agents, they can do a lot of things for less money than you'd usually pay a person. I would say that the main thing with AI agents these days is oversight, which is basically your ability to supervise and then maintain high levels of quality throughout the process. So if you can't verify an AI agent's work in less time than you'd have spent doing it yourself, you should obviously not automate the process.
But if you can, then you should automate it if it's high volume. Another thing to consider is a scientific study that was done that showed that buying time actually improves your happiness. Now, I don't usually think about this while I'm doing offloading math, but I probably should.
Basically, there's a fair amount of research out there that suggests that when you offload something, you're not just paying $150 to not do the task, you're actually buying yourself a whole hour of your time back. And if you think about it, that hour has additional value that I haven't talked about yet. The marginal value equal to your hourly rate.
plus however much peace of mind or happiness that you will get from having a less stressful week and that might be considered its true value okay so some caveats with all that said if you guys are solopreneurs or freelancers or very small agencies despite what i just said you will probably find that most things are not worth offloading the reason why is because offloading is linked to chain law as we saw from the chain law if you add more steps to a process you get worse performance because of these joint probabilities.
Because 0 .9 times 0 .75 times 0 .3 times 0 .8 means the entire process is now subservient to that 0 .3. So if you do a task at 100 % in one step, the end result is 100%, right? And you know it's going to be 100 % because you're good at it and you've done it before.
But if you introduce a new person into the task, you'd have error bars around 70%, meaning your end result is now 70 at minimum. And that's not even taking into account logical other confounders like supervision and delegation and management time. feedback loops and so on and so forth.
So in general, I would remember that . Okay, so time for some action questions. Why don't we do some Fermi approximations to practice and then you guys can grab a calculator.
First, work out both of your rates just off the top of your head. How much money did you make last year roughly? Is that 100k?
Okay, cool, fix that. Then divide it by the number of hours worked for the average hourly rate. The average number of hours that people work per year, by the way, is about 2000.
So if you made about $100 ,000 last year and you worked about full -time jobs, about 2000 hours, probably made about 50 bucks an hour average. Once we're done with that, this takes a little bit more time and energy to think about your job and think about the work that you do or your business or whatever. Of the hours available in your day, what tasks make you more than $50 an hour or whatever your average hourly rate is that you just figured out?
How much more? Add a multiple to that. If that task is two or three times as important, well now you know it's $100 or $150 an hour worth.
Now you know your marginal hourly rate, and you also have done some reflecting on what it is that is power law distributed in your day -to -day work. Once you're done, pick one of the tasks that you are currently doing and price the offload cost. Will the price plus your supervision hours at your rate be higher than that, essentially?
Once you've done this, what is stopping you from actually offloading? Let's chat about the planning fallacy multiplier. Now, tasks in general tend to take longer than you think they will.
That's not exactly rocket science. Even after years of missing deadlines and after learning about this planning fallacy multiplier, you will also continue to almost certainly make estimates that are off in the same direction every time. So what is the planning fallacy?
Well, when you plan a task, anything that you're going to do, you almost always go through all the steps in your head as if it is a smooth plan from start to finish. Because human beings don't work very well on hypotheticals, It is virtually impossible to consider all of the cases where something could go wrong, simply because there are far more ways for something to go wrong than there are ways for something to go right.
What is seen is seen, and what is not seen is not seen. Well, it turns out that there are more things that are not seen than are seen. The vast majority of all of the possible routes that you could take or paths that you could take to screw something up are far larger combinatorially than all of the ways that you could make something right.
It's like how many ways are there to build a building? Well, there's a few. But how many ways are there to destroy a building?
There are a lot more. Because of this, the planning fallacy means that your plan is almost never an actual estimate of how long something will take you. Instead, your plan is almost always a best -case scenario, and because of that, you should treat it like such.
So how much worse are estimates in reality? Well, a good example is a study on 258 transport projects that cost $90 billion in total. Of these 258 transport projects, 9 out of 10 were underestimated, with an average cost overrun of 28%.
If you're counting rail projects specifically, the average cost overrun was 45%. Now these were not amateurs. These were 258 high -quality large infrastructural projects managed by experienced engineers, each of which who had decades of data on similar projects.
My own experience is quite similar. As much as I hate to admit it, I routinely go off of my own estimates by anywhere from 30 to 50%. So what that means is, if planning is always a best case scenario, how can we make this more realistic?
One approach to do this is to use what's called a planning fallacy multiplier. This is where you take your inside estimate, which is the one that you would have come up with using the simulation in your head of all the ways that things are going to go right. And then just multiply that amount of time that you came up with by some constant number.
Believe it or not, that has to be one of the simplest and easiest ways to readjust for your internal bias the wrong way. In practice, many people multiply their planning fallacy estimates by three. Others do it by two.
And in my case, I do it by about 1 .5. And I do it by about 1 .5 because I typically run small projects with very few people involved. So I know that I am usually the majority of the variability in the outcome.
Now, it's important to note that your inside view estimate will almost always be way off. So this is necessary for any stakeholders in a project not to hit your guts. You see this a lot when you pitch a client on timetables and timelines for a project.
I do a lot of AI automation. And so a lot of the time I say something along the lines of this will take two to three weeks. How long do I actually think it'll take?
I think it'll take about a week. But I add on. an additional 1 .5 to 2x amount of time, just because I know that my internal estimates are not to be trusted.
Another strategy is to stop considering a project in isolation and instead begin looking at other projects that are similar to that project. Whereas before we talked about the inside view, which is where you think about how long it would take yourself, this is the outside view. Instead of how long is my course going to take to write, it is how long did the last three courses I took to write.
take. And the way that all of that works realistically is, you know, what you're doing is you're averaging out the actual real times and then comparing them against your estimates. And then what you can do is you can just work on a simple multiple.
As you see in this graph, which represents a study that was done on how often people's estimates of when they were sure something would be finished actually got done. You can see here that in the case of, I give it a 50 % sure chance that we'll probably finish by that date. Despite people's confidence at being at 50%, the actual percentage of projects that were finished were only 13%, which is almost a 4x multiplier.
When you ask people and tabulated the number of people that said that they would be 75 % sure something would finish by a date, you found that only 19 % of projects were actually done by that date. Same thing with 99 % sure, in which case only 45 % of projects were done. Imagine saying, I'm 99 .9 % sure this thing's going to finish today, and you're only right 45 % of the time.
You probably consider yourself a pretty shitty planner, right? Well, that is just how human beings are. On net, we are very, very shitty planners.
The good news is you can kind of reliably just multiply this by two or three, in some cases four, in order to get yourself back to an actual fair number. So that's reference classes. If you look at the actual outcome or the outside view and then compare that to the income or the inside view, rather, you'll find that the reference class will always feel really pessimistic.
But the feeling is the fallacy. We just suck at this stuff. So from here on out, all you really do is you just rinse and repeat.
You continue making whatever your usual estimate is, call that the inside view, then apply a multiplier to the estimate, and then use the scaled number to quote clients, write contracts, and ultimately build your life around. And if that number seems crazy, statistically, it is probably closer to reality than your estimate is.
It is your estimate that is crazy. Do not treat this as a license to slack off, obviously. I recommend you scope every individual step very tightly and aggressively internally.
So add very tight timelines to individual steps within an overall plan, just to make sure you have enough leeway with your stakeholders. This allows you to set both strong internal deadlines and then reasonable external deadlines as well. And this is at the core of thinking, clearly, because thinking is usually planning and planning is usually long -term strategizing.
So some action questions for you. Think back to the last project you estimated a time on. Could be an essay or something like that that you thought you'd be done by X date.
Do you remember the length that you proposed versus how long it took? Now this can be anything, a personal thing, a relationship thing, a business thing, whatever. Whatever the ratio is versus how long you thought it would take to do and how long it actually took, that is a good multiplier to start using.
And just use that. Explore exploit trade -off. Now imagine you're in a casino and you're presented with 10 awesome looking fancy slot machines.
Every machine has a different payout rate, but you don't know which machine is which. And what you have is 100 pulls to distribute amongst each of the possible slot machines. Every pull on a machine that you know pays out well, aka an exploit of a machine that you know is going to pay you, costs you a pull that you technically could have used to discover an even better machine elsewhere.
This technically is called exploring. On the flip side, every pull on a machine that you don't know the payout of, aka ones that you are exploring, will cost you pulls on a machine that you do know pays out well. This is a classic problem in research and development and also in planning and acting throughout your life.
How many of these machines should I pull before I should feel satisfied with what I have and just pull the highest performer? This is called the exploit -explorer trade -off. Should I do what works?
Or should I try something new? Exploiting is always about milking a known good. Exploring is about spending time and accepting suboptimal outcomes.
If you think about it from a local maximum perspective, this is where you descend your local maximum. To learn more about other options that might be better performing than the ones that you're currently exploiting. Now, super annoyingly, I find both of these are seen as virtues in the real world, which prevents a frank and honest discussion about how to proceed.
For instance, somebody who never explores is usually called focused or specialized. And then when their thing, whatever it is that they're doing, dries up because they've been exploiting the hell out of it and you can only exploit anything for a certain amount of time, then they die because they've never innovated or grew past that.
On the other hand, somebody who never exploits could be seen as growth -minded or always learning for a very long time. But when you peel back the curtain, you realize that they're poor as hell. They're broke.
They haven't actually achieved anything. Why? Because they've never stopped to spend enough time milking any one thing that they've discovered to actually reap the rewards of having found it.
So the way you can think of this as the exact same local and global maxima problem that we discussed earlier, it's just now we're adding a time dimension to it. So when should you explore more in life? Well, there's actually some science here, and there are some neural circuits that have been studied that imply that human beings actually have this built in.
So in this study, we found that human subjects tended to explore more when there were many trials remaining. But as the number of trials decreased, or when the rewards began to decline, subjects shifted their energy primarily towards exploitation. Now, I find this sort of comforting, because if you think about it, unlike most of these other biases and programs, in this case, it's like our brains actually have built an algorithm to do this for us.
You know, we've been installing a lot of new ones, but this is literally something hardwired, probably for evolutionary reasons, because exploring and exploitation are core to survival. So how do you actually use this built -in algorithm to approach your decisions rationally? Well, to make a long story short, if you have a lot of time on your hands, let's say you're very young, you should explore a lot.
If you find a better option now, that means that you will get to exploit it for all future time steps. If you are 20 years old and you can expect to live to, say, 100, that means that you have 80 time steps left to exploit a good thing that you find now. But you should probably err on the side of finding new things as opposed to just sitting down and exploiting.
You can think about this in the case of a job. If you found a $30 an hour job right now, that might be pretty good compared to where you were before. But if you settle there for the rest of your life, the next 80 years at a $30 an hour job, that $30 an hour job will only last you so long before you can no longer exploit it.
And is that really the highest you think you can climb? I think about this like investing in risky portfolios. Since you guys have more time for a risky portfolio to pay off in investing, and on average, riskier portfolios do pay out more.
If you're not investing or if you're not expecting to withdraw money from your investment soon, the number one piece of advice is usually, well, put it in a riskier investment. Put it in something a little more aggressive.
Why? Well, because over the course of 60 years or something like that, 50, 40 years, however long you're actually keeping your money invested for, you know, the market has gone up a lot. It's only in the short term that it tends to dip.
Now, you can contrast this exploration kind of guidepost with the exploitation guidepost. which is that if you have a very short time horizon, you should exploit instead of explore. The reason why is because you don't have the time to make back what you spent exploring.
Here's some other rules. If your current reward is declining, you should explore more. The benefit of exploiting is going down, meaning the value of information and potentially value of other candidate options is going up.
And if your current reward is increasing, you should explore more because you just found a winner. So here's a quick visual representation of what that actually looks like. In gray, you have returns from the proven thing that you are exploiting.
What you will notice is as you exploit the thing more and more and more and more, eventually you will start plummeting because there's only so much time and energy that you can exploit something before the opportunity disappears. You can think of it like a business model, per se. At the same time, the red is the share of your effort spent exploring.
So ideally what you do is you'd actually explore very, very hard early. And then after you make it to maybe halfway through, what you would do is you would milk the winner over and over and over and over and over again. And then when your winner starts decaying, like they do in gray over here, you spike up and start exploring again.
And so this red here, to be clear, is just how much you explore. At the beginning, you explore a lot. Then though, eventually you explore very little because you're milking a known winner.
But eventually the winner stops producing a valuable return on investment, which is when it goes down, and then you start exploring again. And so always explore hard very early. You can think of my own life very similar to this chart.
The door -to -door period of my life at around 22, where I was knocking at 50, 80 doors a day, that was extreme exploration until something stuck. Once I figured that out, I exploited this script that worked, which was just rinsing and repeating the exact same client acquisition strategy book that I was running when I was going door to door in cold calling, cold emailing, cold DMing, and a variety of other outreach -based methods.
And I just continued doing that until we got to about $20 ,000 a month. However, eventually, that client acquisition approach stopped working. It literally just stopped.
Why? Because opportunities only last a certain amount of time. And then I went back to exploration again.
It's just this time, instead of doing client acquisition the regular way, I started doing AI. I started doing content marketing. Now, I didn't make money off of that right away because I was in explore mode.
But eventually, exploration paid off, and then I found a winning lever, something that I could pull, like that slot machine that delivered me a higher return on investment. Once I found that, I shifted to exploit mode, and I repeated that until about $90 ,000 a month. That was my ceiling.
Now, after that, I was on that local maxima. Couldn't get any higher. So I took a step back, and I started exploring more.
That took a while, but eventually I found another lever like content, personal branding, and so on and so forth that I started pulling to scale to $540 ,000 a month, which is my largest recent monthly revenue. So there are seasons of life in all things. You'll probably go through multiple seasons, including ones where you want to explore a lot and ones where you want to exploit a lot.
And that is totally fine. The good news is now you have a rigorous and procedural name for it. So you don't need to worry.
that what you are doing is wasting time. Now you can frame it as exploring or you can frame it as exploiting. And you also have a framework by which you can function.
You'll know, for instance, you should not just be picking up every shiny object and exploring that all day. There will come a time when you need to select the shiniest of the objects that you picked up in, say, the last 90 days and then commit to it. Only then will you actually be able to exploit and then maximize your return on investment.
some action questions. And I find it funny how all of these say two questions to answer because I just copy and pasted the same thing. Two answers to question before you continue.
No, one question to answer before you continue. Map your own explore to exploit pathway. At what point in your life were you exploring?
And at what point in your life were you exploiting? Okay, now it's time for my favorite part of the course. All the software that we talked about so far is a bunch of very general ways to clarify your thinking.
And once you see them, you cannot unsee them, which is why it's valuable to learn this, because once you learn them, they will always be with you. But what I will discuss next is a little different. Rather than a general purpose software application, these are short and very punchy programs.
They're immediately actionable techniques that you can use and abuse to chain together thought with action. This will round out your arsenal and make you not only a prolific thinker, but also a prolific doer. And I use stuff like this all the time in my own life.
In fact, TAPS, which I will cover first because I think they're very high ROI, probably my number one most exploited tool that I've never talked about before this day. So if you've been paying attention, you guys will have a pretty good idea of how to solve a problem and make high -level decisions, right? But there is still a common failure mode.
Many of you will understand and agree with each of these techniques, and then when it is time to apply them, you will not do anything. That is called Acratia. Remember?
The whole idea about TAPS is solving Acratia. So TAP stands for Trigger Action Plan. And they're a very simple way to bridge the gap between knowing something and doing.
They take the form as follows. When X trigger or Q occurs, I will do Y action or behavior. This is like a line of code.
You know, if code isn't deprecated because of AI. What it is is an if or then statement. You can also think of it as an SOP in business, like what to do when a client sends you an email.
Or you can also think of it as a rule of life. The idea of taps are to eliminate the question that everybody asks themselves when they are faced with a decision. Should I do this thing?
And this should question often acts as a blocker to real growth. Here are some examples of simple trigger action plans that I have made. If my phone illuminates while it is in my workspace, or if it vibrates while it is in my workspace, I will get up and move it to another room.
And it's very funny that I just said that because I had my phone with me because I just had to take a call. And because my phone did just illuminate or vibrate, I'm now going to go and move it into another room.
That single decision that I just made has been programmed in my mind for the better part of the last year. And in me making that decision right now, the resulting maybe 10 or 20 % of this course will likely be 10 or 20 % better.
It takes you only a few moments when you start hardwiring these rules like circuits or switches in your mind. A next rule I have is when I sit down at my desk, I will open the document I need to work on before doing anything else. In my case, I just came back to sit down at my desk.
The document's already open, so I don't need to do anything. If I ever catch myself saying a word or phrase like, I'll figure this out later, I actually set a five -minute timer and then I write a list of actionable steps to get it done right then and there. This is a simple plan called a resolution cycle, which I will run you through later.
Now, while this if then statement idea sounds very trivial, it is actually one of the most strongly evidenced techniques in this entire course. In academic literature, this is referred to as an implementation intention. The term was coined by a man called Peter Goldwitzer, who studies behavior and how to optimize it.
Many other meta -analyses have also shown that people that make these if then plans are always significantly more likely to achieve their goals than those who merely set vague goals. Specifically, a 2006 meta -analysis showed an effect size around 0 .65 standard deviations, which to be clear, is quite large, for a simple 30 second intervention.
For those of you guys that don't know standard deviations, this is basically taking you from a 50th percentile to maybe the 75th percentile, which means just using taps or trigger action plans, can legitimately help you leapfrog a quarter of the entire known world in your ability to do things. There are very few things in life that allow you to leapfrog a quarter of the known world that you can figure out in less than five minutes.
That is what you are doing right now. So this is a quick visualization of this. 94 studies on implementation intentions, the fancy word for trigger action plan, over almost...
8 ,500 people, they saw a standard deviation improvement in their ability to achieve a goal by 0 .65. There is opportunity and gold everywhere for those with eyes to see it. And this is a very simple example.
This is something that I've been doing for a very long time. And I guarantee you immediately after starting to do it, I leapfrogged 25 % of the entire population of planet Earth without even knowing in terms of my economic effectiveness. When you do that with action and thought, It is only a matter of time before other areas of your life follow.
So in my case, my income, in my case, my life contentment, and so on and so forth. So why does something that's this small actually work? Well, if you guys think back to the willpower section for a second, the people that have high self -control always set their lives up such that they do not win the fights in the moment.
What they do is they move the fight or the decision out of that moment entirely, and they either defer to their former selves, like via ulysses contracts or they remove the opportunity to make a mistake in the first place like i just did with my phone a tap does something very similar on purpose the idea is you will make the choice once now while smart calm and relaxed and then you will hand the reins to the if or then statement to work for you now at the end of the day deliberation is expensive and difficult Every time you have to decide, should I do this?
It is equivalent to having to go to the gym, lie down on a bench press and rep out 225. If you have to do that before you do anything, why the hell? Like you're not going to want to do half of what you have to do, right?
It's a very energetically involved task. Well, taps just allow you to skip the bench press entirely. They allow you to skip most of the hard work.
So how do you actually make a trigger in the trigger action plan work? Because to me, that's by far the most important thing. As long as you can immediately and viscerally notice a trigger, you are usually more than half of the way to doing the task.
To be successful, the literature on implementation intentions suggests that your trigger must have the three following attributes. Number one, it must be obvious. Number two, it must be visceral.
And number three, it must be reliable. So obvious means you can't miss it. When I feel like I'm going to procrastinate is not an obvious trigger, and that would not work.
And I've tried stuff like this, and I used to use triggers like this way back in the day, and they did not do anything for me. The reason why is because your brain doesn't actually signal it in a way that you can detect. You don't even know when you're procrastinating 80 % of the time.
You just are. A much better way to do this is when I open a new tab in my browser. Because that is something that is immediately clear.
It is very, very obvious. And there is a Boolean yes or no. yes I opened it, or no I didn't open it.
Visceral means that it is tied to a physical or sensory cue. So when you feel a phone in your hand, or when you hear the door click, or when you sit down in your chair, or when you take your first sip of coffee, all this stuff is self -explanatory. Human beings recognize triggers very viscerally when they're physically stimulating us in some way.
And so when I open a new tab, for instance, I open it by holding the command key and pressing T. This is both like satisfying from a haptic perspective, but it is also immediately stimulating to me visually because I see it open up on my window. And so both of these things sort of hit my mind simultaneously.
And now I have a trigger that I can reliably depend on. Speaking of reliably depending on things, reliable means that it triggers every time that you find yourself in the target situation. And ideally, it never triggers when you are not in the target situation.
I'll give you an example of a very bad trigger. 3pm. That is not reliable because human beings, like we don't have the ability to tell when we're procrastinating, also do not have the ability to tell what the time is internally.
Some days, if you set your trigger to be 3pm, 3pm will pass you and you won't miss it. It is just part of being human. But instead of tying a trigger action plan to 3pm, why not tie it to something like an alarm that sets in advance?
That way you are far more likely to trigger on said alarm. My recommendation is use alarms. They're awesome.
Alternatively, tie taps to actions that you do and you have to do every single day. Think things like waking up, going to bed. A lot of people here are hopelessly addicted to caffeine, myself included.
Why not tie something to something that you know you will do every single day? A lot of people here are hopelessly addicted to other things like turning the television on. for instance, or watching your favorite show or whatever.
You can actually use these anchors, these things that are going to occur as immediate triggers to begin some action or some sequence of actions that improve your life. Finally, the action must be small and physical. Do not do hard things like spend two hours writing.
That is not how you do a tap. Instead, do something small, concrete. and ultimately an action that gets you started.
And then what you do is you just trust that the inertia of the small action will guide you through the rest of what you need to do. For instance, open a Google Doc and write one sentence is an excellent action. If the trigger to that is every time I sit back down at my desk, I open a Google Doc and write one sentence, that is a fantastic trigger action plan.
The probability of you actually completing this is quite high, assuming you do the steps that I will talk about in a moment. And the value there is it's very hard to have just one sentence worth of an idea. Usually after a human being has written one, their brain just starts filling in the blanks and continuing.
So you'll naturally want to chain that together into something else and so on and so forth. The whole point of a tab is to get you moving and to solve akrasia, which is your biggest problem by far. Now, how do you actually install one and maximize the probability that this tab works?
Well to start, it's not sufficient to simply brainstorm and write down a bunch of taps. You have to rehearse the tap over and over and over again before you can eventually start doing it on an automatic level. Once it's on the automatic basis, then you will do it yourself and it will self -reinforce.
But that initial on -ramp to the beginning of the positive flywheel, that is difficult. Luckily there's a very powerful method for doing this. After coming up with your tap, I want you to close your eyes and vividly imagine the trigger.
So in my case of sitting down and then opening up a Google Doc, I want you to visibly imagine and viscerally imagine sitting down in your chair. So right now I'm actually imagining the process of me sitting down in my chair. What does the trigger look like?
What does me sitting down in my chair look like? Well, when I sit down in my chair, what do I see? I see this.
So I see my laptop, my computer monitor, my laptop, and then I see, you know, a camera. And then I typically have a microphone. I see my desk.
I see the sun on my left side. You know, I see a big white wall on my right side. What does it feel like?
Well, my chair is really uncomfortable by the way. So I'm going to have to replace it with a steel case leap in a moment. Um, just need to unbox it.
I have little arm rests where I typically put my arms. Um, what does it feel like? Hmm.
It's kind of cold in the room right now, which is good, which I like. okay what does it smell like luckily for me it doesn't smell like anything then after you've really just tried to visualize and rehearse you feeling the trigger imagine yourself performing the actual action itself so in my case opening up a google doc sitting down and then holding command t and typing docs .new i will do that 10 times in a row this whole process here believe it or not takes less than two minutes And it might seem silly, but what you're doing is you're training your brain's cue detection system.
You're taking these cues and your intention from working memory, and then you're actually encoding it into long -term memory. Another thing you can do is you can actually practice this multiple times. So for instance, when I was practicing this sit down and then open a Google Doc thing, what I would have done is I would have actually closed my door, went outside, opened it, came down, sit down, and then open a Google Doc and write a sentence.
And I would have repeated four or five times. I would have done this over and over and over again until the next time that I sit down, some little part of my brain would have been activated, you know, like a monkey. My neurons would have gotten activated and I would have gone, all right, wasn't I supposed to do something?
The second that you can get yourself to that point, everything else is done for you. You are an automatic machine capable of doing whatever the hell it is that you want. Another way to do this is putting the trigger directly in your path, which is the friction rule.
Now, if your tap is, when I see my running shoes at the door, I will put them on. Just make sure your shoes are actually at the door. The more tangible the trigger and the more in your control, the better.
In the case of our sitting down on a Google Doc, for instance, in order for me to do anything with my computer, I need to sit down at this chair. which is why this chair is such an extremely powerful trigger for anything that I want to do. Same thing with opening my computer or turning on OBS or whatever it is that your particular trigger is.
Now, every now and then, once you've installed it, just ask yourself, did I actually trigger that tap? If not, the problem is not you, because we are not the problem here. It is the trigger that you chose.
It's not a willpower thing. It's not like, oh, next time I'll do it because I'll remember. No, no, no.
It's not that at all. It is the trigger. Your willpower has nothing to do with this.
At that point, you can choose something more noticeable and then rehearse it again. Okay, so this is the flow. Pick a trigger, write a sentence, rehearse it 10 times, put the trigger in your path and just check and go over and over and over and over again.
The debug loop here is at the trigger stage, nothing else. Most of my taps, to be clear, are at the first hour of my day, and they're also very boring, but they're also extremely effective. What I do is I will chain my taps together such that the action of one of my taps is the cue or trigger of another.
This is sort of advanced tapping, but hear me out. I'm going to give you guys some pretty grisly detail here. After I wake up, which is my cue, I will immediately walk over to my coffee machine and then I will brew.
That is my action. After I click brew, cue, I then use bathroom and then I wash my face, which is an action. After I wash my face, cue, I will grab my coffee from the kitchen and I will put a bit of protein milk in it, which is my action.
After I put a bit of protein milk in it, which is my cue, I will go to desk and I will open my laptop, which is my action. After I open my laptop, which is my cue, I will click a bookmark that opens my school tabs, which is my action. Right, that's right over here.
After I open my school tabs, which is a cue, I will read the first question and then voice record my first answer, which is my action. From waking up, to voice recording my first answer, this process takes me one minute and 30 seconds.
But by the end of that one minute, 30 seconds, all of the brain circuitry that was typically involved in, you know, doing in my case, school management is ready to go. It's like, it's like I'm a world class athlete or something like a swimmer, like Michael Phelps listening to his favorite track and doing his favorite stretches right before he goes on.
It is the exact same thing every day. And what he has done and what all professional athletes and high performers in any field have done is they've just standardized their inputs. So do you guys see how easy it is?
Eventually, all this becomes completely automatic. And then I have repeated taps until maybe 2 .5 hours in my day, after which I am usually done most of the economically valuable chunks of my work. The key here, again, is to use the clarity of thinking you get when you are in a pristine condition, aka motivated and ambitious, aka right now, to ensure that even when you are not in a pristine condition, you are still moving the needle of your life.
An action question here. Take a technique from this course that you most want to use. Write it as a tap, which is a cue.
and then an action. Then rehearse it 10 times right now using the visualization practice that I gave you guys earlier. Now I want to talk a little bit about noticing.
What I'm about to tell you can be thought of as a background function. Once you install this background function, it will immediately improve your growth rate across your entire life. The way I like to think about this is as a savings account.
and taking funds from the savings account and moving it into, say, an investing account, which typically delivers higher returns. If you don't have this software platform or program installed in your brain, you will grow at 0%. But the second that you install this concept in your brain, and this concept has many forms, so some of you will already have this in your mind.
But the second that you install this concept in your brain, it's akin to making 7 % to 10 % on an investment year over year. You will become clearer. and more effective simply by virtue of time passing.
So simply by virtue of having learned about this idea, noticing, you will naturally and organically be growing. And it is not hyperbole. It's not woo -woo spirituality stuff.
It is like 100 % concrete and based in science. So what does noticing refer to? Put simply, noticing really stands for noticing that you're confused.
In real life, Confusion typically lasts just a few seconds before your subconscious mind will start squashing it away out of fear. So noticing that you're confused is the ability to recognize in your mind that something is wrong and then actively label that thing that's wrong by saying or explicitly verbalizing, I notice I am confused before that feeling disappears and before you immediately shove it away to a blind spot or there's never to be looked at again.
This helps make your map line up with your territory. which improves your internal model of the world. And if you do this several times per day, because we typically confuse several times per day, you can think of every single time of it occurring as a 1 % improvement to your map.
Eventually, your map gets closer and closer and closer to the territory that it represents, and you become a lot more effective as a result. So let's say hypothetically that you're listening to an audiobook, and something does not add up. Maybe you weren't paying attention.
the wording all of a sudden seems off, or there's a big gap and you don't remember how you got there because you were looking at a bird. Well, you can use noticing to fix this. And there's a great essay all about noticing that I'll link you guys here just in case, but I'm going to give you a shorthand.
When you have the sensation of noticing that you're confused, it means that your internal model of reality is off. Rather than hiding away from it, you can just verbalize it to yourself. I noticed I'm confused.
Why? Then, because you are now forced to verbalize your hypotheses, you also put actionable steps in place to prevent or minimize it from happening again. I know that this sounds stupid, and I know it sounds extremely, extremely simple.
But you'd be surprised at the small fraction of the confusion events that occur in a 24 -hour period that you are actively addressing. You probably attend to less than 5 % of all of the events throughout your day in which you are confused. You are almost certainly just allowing these events to occur and then moving on and not even questioning why you feel confused.
Instead, you just feel confused. This is a dumb move because being confused is about as clear an opportunity to rectify your map and make it as close to the territory as you will ever get. Another example, let's say you're talking to a client for whatever service and they tell you that your campaign is going great, but then the dashboard on your screen says something different.
In this case, one of you guys must be wrong, right? Rather than just having that idea, something's wrong here. I'll mention it later.
Just notice it. I'm confused. Your data does not line up with my experience, which is that the campaign is not going very well.
Could we chat for a second about why that might be? A simple statement like that doesn't just help yourself out, but it helps multiple people plus an entire organization out. And now you and your client can grow together.
And I don't mean this in like, again, feel good sense. I mean, people will both respect you more for pointing out that you are confused, and it is also far more effective and efficient because now you and the person you are talking with are both operating on a shared model of reality. You are not just hiding away your confusion because you want to seem smart.
It is the people that are open to seeming dumb that are typically the smartest ones amongst us. So do not shy away from confusion. That window is where so much value in real life lies.
Another example of a similar concept is the flinch. The flinch is when you see something that you need to do, or maybe you're about to go look at a number, or you see an email in your inbox, and all that just makes you go, oh, God. Maybe this is a text from your ex.
Maybe it is a letter from your boss or your client. Or maybe it is an email from a supplier. Whatever the case is, when you go, ah, aka when you flinch away, your inherent desire is to look anywhere but there and think about something else, right?
The issue is every time that a thought hurts your head, right? Every time you go, ah, and then you don't act on it. You are training your brain to avoid that.
And it's not just an opinion, it is actual neuroscience. Neurons that correspond to actions that you take regularly get stronger and more defined. The ones that correspond to actions that you avoid always grow weaker and shrink.
So, when you avoid the thing, you are literally training the circuitry of your brain that goes, you should avoid stuff that is painful. The issue with that is most of the good things in life occur behind doors that are painful. Now, interestingly, I also find this happens with careers.
How many people, myself included, spend half a decade or more preparing for their dream career, only to maybe get the job or think about the job, and then they go ugh every morning when they wake up? Well, this is a newsflash, but you don't have to live that way. Rather than flinch away from something and then immediately try and bury the emotion, try sitting with it for at least 10 seconds.
Like, actually sit down and consider, why do I feel that way? Is it the job itself? Is it a person at my workplace?
Is it the process of commuting to my workplace? What is it about this thing that makes me go, ugh? If you go, ugh, 10 times a day, and then every single time that occurs, you ask yourself, let me clarify exactly why I feel bad.
Quite simply and quite quickly, you will soon never ugh again, because you will have solved all of those problems. Every time you ugh or flinch away is legitimately a massive opportunity that has just appeared on your plate to unfuck some character of your being, some part of who you are. So your mind and body are very, very valuable and sensitive instruments.
And I found as I grow older, you can often think far more clearly just by listening to them. Don't avoid their signals. You should sit with them for a moment and actually listen.
The last thing I want to talk about is the should. So how many times have you said some variant of this today? I should really post more.
Oh man, I should really read that book. I've been putting it off. Or man, I really got to call that person back.
How many shoulds have you tried guilt tripping yourself with in the last 24 hours? Have you ever actually stopped for a second? Have you ever considered why you should do these things?
I mean, you are a donut -shaped sack of meat around a hole that metabolizes and poops out food. Why should you, a meat sack on a random rock somewhere floating through the universe, call somebody else? Why should you put up a microphone device to a listening hole, slam your fingers around, and then reverberate the talking hole for 30 minutes to communicate with another meat sack?
Is it written anywhere that meat sack A needs to communicate with meat sack B? No, obviously not. So where is all of this stuff coming from if you ask yourself?
Well, here's another newsflash. Most of your beliefs are what you should be doing come from somewhere else. They come from outside of your body and your mind.
So to be clear, you can and should do whatever the hell you want to do. My issue is most people just never actively examine what it is that they want to do, and they never actually think for themselves. We would rather mollify or placate ourselves with the notions of, oh, I should do this because X person said so.
I should do that because my family used to do this. I should do blah, blah, blah, blah. We mollify ourselves from shoulds borrowed from friends, spouses, social media, institutions, and so on and so forth.
And all of this at the end of the day means that most of us spend our lives not for ourselves, but for others. And I don't mean that in a noble sort of honorable sense. I mean, literally just a, I've never really considered any other way of living sense.
That sucks, right? I want to live for other people, but I want that to be my choice. I don't want to live in a structure that somebody else creates for me unless I have consciously sat down and said, you know what?
I respect the values of the structure and I like the structure and I want to abide by it. The good news is you can actually solve this in just a few seconds. The next time you find yourself saying, man, I should really do X thing.
Just ask yourself, why should I do X? Do I actually want X? man, I should really call my friend.
Ask yourself, why should you call your friend? Do you really want to call your friend? In many cases, you will want to call your friend, which is awesome.
But in some cases, to be honest, you will not want to call your friend. In some cases, you will say, why am I calling my friend? You know, we don't really get along super well anymore.
I don't feel really good after I call them. I always feel guilty for not calling them. They're kind of an asshole.
Maybe I just shouldn't call them to begin with. Huh. If that is you, great.
Because now you don't have to should yourself the next time your friend's like, dude, you got really got to call me. Alternatively, if you do want to call them, if you go, man, I should call them. Yeah.
And the reason why I should is because I like them. They add value to my life. They genuinely improve my state of mind and my being.
And I want to help them just like they've helped me. Well, then great. Now all you need to do is ask yourself, what is the next physical step to actually take that should and actualize it in the real world?
Do I set a calendar reminder? Maybe do I let them know, hey, you know, are you game to talk on a weekly basis? Because I keep on running out of time because I don't have this written in a calendar somewhere and I keep on juggling a lot of stuff.
Are you game to just chat every Thursday at 4 p .m.? If that's the case, great. They'll probably do it more often.
At the end of the day, you don't actually have to do anything. But at least working out why you feel the way that you do and what the next logical step is means that you are now thinking productively and proactively and clearly as opposed to just flinching and constantly feeling confused. So at the end of the day, these are three signals.
And for each of these signals, there are three five second moves. The good news is you can actually turn all of these into taps. You know, flinching, for instance, that is a clear concept.
Should is a clear cue. Confusion, I notice I am confused. That is a clear model.
And so when you're confused, that means your model, aka your map, is wrong. All you should do is check. Why is it wrong?
Is it really wrong? Why do I feel this way? If you flinch away from something, That's like an ugh field.
It's a field of ugh that you just don't even want to like look at or address. What you should do is just look at it. Look at it for just 10 seconds.
Just like stare at the thing that you're avoiding. If it's that email or text message, stare at it. Look at it.
Understand it. If you are going to do something about it, you can do it after you do that. Finally, with shoulds, whose goal are you actually living?
If you actually want it, then you should take actions. And this is going to help you do so. So how do you actually train this?
When you catch yourself noticing one of these thinking traps, I would recommend verbalizing which one it is immediately. So for instance, I will often think to myself, I'm confused, or I just flinched, or I just said should, etc. And then I would connect it to a tap.
Like when I notice I'm confused, I must state that I notice I'm confused out loud and then ask why. So for instance, if I'm confused, right? If I don't know, I mean, I'm looking right now at my camera, for instance, and there's some sensor there that I think means temperature, but I don't really know.
So I notice I'm confused. Why? Because there's a sensor readout and it says temperature.
And I'm not really sure if that means that the camera is about to go dark because it's too hot in that corner or something else. Something is odd here. The very act that I just said that means later on, I'm going to address it and fix it.
It's extraordinarily easy for you to do. As long as you notice and then label and then have a tap of some kind, you can usually eliminate most problems in your life within a few weeks. And you don't even have to worry about taking action on the noticing just yet anyway.
The real value is literally just being aware that all of this is occurring. When you start being aware of all this stuff, your brain naturally starts trying to improve it. So as mentioned, this is the passive investing strategy of clear thinking.
This is a quick example of how much better things get. If every time you feel, ugh, you avoid it, your discomfort on the left -hand side, every future time you think about the thing, will just skyrocket exponentially. But if every time you go, ugh, when you get a piece of stimulus that you don't like, you actually look at it and sit with it, well, the discomfort will then go down and basically ground down to zero.
Okay, so two action questions before we move on. First is, can you think of, off the top of your head, something that has made you confused or flinch recently? Next, what are three things that you commonly tell yourself you should be doing and why?
I want to talk a little bit about premortems. Premortems are a simple and very effective way to improve your ability to plan projects. Now, I use them all the time, and they're a very vital part of my personal life and relationships, my business life, aka my deals and brand strategies, and even my fitness, like my workout plans.
If you guys have ever been part of a committee or an organization or a corporation, usually the way that project management occurs is somebody will deliver a plan. And then what you do is you look at the plan and then stress test it, usually by roundtabling answers via some sort of brainstorm way. You guys will ask yourselves, what might screw up with our plan?
After that, you'll spend a couple of hours on calls, brainstorming additional ideas. And then at the end of it, you'll usually get three to five vague risks out of the process. After the project, when it inevitably fails, it will not have failed.
for one of those risks. It will almost always have failed for some other reason that no one predicted, simply because there are many more ways for something to go wrong than there are ways for something to go right, like we talked about. Post -mortems are fairly standard project management techniques.
In corporate, what you'll do is you will round up all the people involved in a project that did not succeed, and then you will say, so what went wrong here? What can we learn from this in order to make next try better? The idea here is to figure out the failure mode so you can apply that to future projects and then improve your total success rate.
Now, obviously this is a good idea and postmortems are awesome and I'd recommend them, but there's an even better strategy you can employ called the premortem. So what is a premortem? Very simple trick.
Just change the tense. Instead of post, it's pre. Instead of asking, why did something fail?
You ask, what did fail if I were to have done this and it failed? So for instance, let's say you have an idea to start a project and the project's going to take you a week. Rather than asking, so what could go wrong here?
Or rather than waiting until the project fails and then asking, so what went wrong here? You could just say, so let's pretend it's one week from today and the project failed. Why did it fail?
This is an extremely clear, empirically proven way of increasing the quality of predictions that people make in a project and the probability that they are correct with their predictions as well. Why? There are many explanations for why.
The first is when you ask the question, what might fail? It's a very abstract question and your brain treats it like an abstract question and produces a very abstract answer. The reality is that human beings do not dwell in the abstract world.
We don't dwell on hypotheticals or conditional probabilities. That's why we have to learn what Bayes' theorem is. We have to learn about expected value and we sort of have to add that onto our biology.
The place that human beings dwell is in stories. It's in narratives and it's in explanations. So when you ask, why did the project fail instead of why might the project fail, your brain will start treating it like a story and it'll actually start filling in the details.
It'll create characters, events, like what happened on Tuesday, specific failure cases, like, you know, somebody forgot to send an email, et cetera. What you do is you do this. So you, everybody asks how, how and why did the project fail?
It did fail a week from now. So why, why did it fail? And then you just list all of the reasons why people think that the project failed from yourself included.
And then you bin them into the actual failure case. Like if a bunch of stories have to do with people not sending emails when they should have. Now you know that that is actually probably a likely failure mode.
The end result is you end up with a much richer tapestry of all possible failures alongside clear solutions that ultimately improve project outcomes. Okay, so this is called the pre -mortem loop. You write the plan.
You pretend it's a week later and it failed. Then you write the story in the past tense. Everybody brings their stories together and you just fix all the causes.
You do this over and over and over and over again until your project is airtight. So let me give you an example. I want to record and publish a video every single day for 40 days.
I would do a pre -mortem message from the perspective of someone 40 days in the future saying, hey, it's 40 days in the future and I'm looking back on my goal and I only published 11 out of the 40 videos I wanted to. Why? Well, if this were the case, I would reflect on that and then I would say something like, I think the issue was I got caught up in client projects about a weekend.
And then I broke the chain of my daily uploads, which is historically always the most important thing. I also made it really hard to keep up after that, probably because I tried being a perfectionist about it and aiming for higher quality than I probably needed. Then instead of once per day, I fizzled out and produced maybe one or two a week.
Now, because I put myself in future Nick's shoes, the failure modes are far more tangible and far clearer. And I can now put systems in place to avoid all of these from occurring in the first place. In this situation, for instance, it sounds like client projects were my main problem.
And breaking the chain was another main problem. So I would make sure my recording time was the very first thing I do every morning, meaning my client projects can't push me off the habit, and then I would use something like a Ulysses contract to publicly state to my audience that I want to do this, or have somebody else hold me accountable, let's say, by donating to my competitor or something of that nature.
Two more things. First, the premortem is the single best use of AI that I know of in planning. Because models can generate hundreds of thousands of narratives in just a few moments, you can use AI to generate a lot of these and then read the narratives and failure cases and then go, yeah, I could have screwed that up big time.
You will be able to label these instinctively and then it's super quick to do so. The next is the premortem is very cheap. Even if you don't use AI and you just use your old meat brain, it is like 15 minutes on one page.
The thing that it replaces, which is actual failure, will often take you weeks or months. So spending 15 minutes to reduce the probability of a failure by like 3x is very high leverage.
And also, if it doesn't do anything, very little skin off your back. It's been 15 minutes. Okay, action questions.
First, pick the plan that you're most excited about right now. Pretend that it is one month later and it failed. Write the story in five past tense sentences.
Name the week that you failed, the mechanism behind why, and so on. Next, what's the cheapest one fix for the most likely cause in the story? And can you do it before the plan starts rather than after?
Let's talk resolve cycles next. The following few paragraphs are from an old post of mine that I wrote back in 2019. Few things destroy a good idea as effectively as the phrase, I'll get back to it.
To solve this, set a five minute timer and actually try to solve the problem before it runs out. I mean, don't just plan how you're going to solve it or sketch out a possible approach. Actually attempt to solve the problem in five minutes.
The next time you have a problem that seems too hard, time yourself. How long have you actually spent thinking about it? I guarantee you haven't even given it a full 30 seconds for giving up and saying, this is too hard for me, or I'll deal with this later.
If you spend five minutes on it, that is 10 times the work, which means it's no wonder that you get a different result. In practice, you would not believe how many actions the average person procrastinates on for days, weeks, or even years sometimes that are legitimately doable in five minutes or less. So I'll give you an example from my own life.
Right now, I'm trying to get my Bulgarian passport. I'm technically a Bulgarian citizen, because my family members were born in the country, and I know Bulgarian, and I go back there reasonably often. And I just want to get a passport to prove it.
It'll also help me expedite EU travel. However, proving Bulgarian citizenship takes a fair amount of time and paperwork. And a lot of this includes, quite frankly, annoying steps, like I have to go to my local BC guys here and notarize stuff.
I have to episteal my documents, and so on and so forth. One step that until recently held up my citizenship proof process by 45 days, which is a month and a half, was simply calling a notary down my street. You know, I could have found their number in a minute.
I could have figured it out what I wanted to say in maybe another two minutes. And the actual call itself would have taken me maybe a minute or two as well. So five minutes in total.
But it still took me over 45 freaking days to do the thing. So why did I avoid it? I mean, all of this ties back to the concept of akrasia that I talked about earlier.
The problem is never knowing what to do, really. It is merely starting to do the thing. And usually that involves a hurdle at the first step.
In my case, getting the phone number of the notary, which is honestly a trivially easy thing to do. Well, resolution or resolve cycles or five minute timers solves the problem. And it is the simplest thing on earth.
You literally just set a timer and you see, can I finish the task in five minutes or less? The reality is you can always fit five minutes in somewhere. There's no man or woman on earth that is too busy for five minutes.
Also, this solves multiple problems, like the should I be doing this now problem. You ultimately have no idea whether you should or shouldn't be doing something until you've at least looked at it. So you can actually use these five minute timers to answer the should I be doing this now problem.
In many cases, I will look at a task that I've been avoiding, set a five minute timer, then while starting it, realize, oh, you know, I don't actually need to do that anymore. And now I'm already done. And it only took me 30 seconds.
The five minute timers assist me in doing this very, very easily. They assist me in resolving the problem. It'll also take advantage or take care of the this -is -going -to -take -all -afternoon problem, which is a separate but related problem.
Because the reality is you can't know how long something will actually take until you start, right? You can guess, obviously, but as mentioned with the planning fallacy, many of your guys' guesses will be very wrong. Well, with the five -minute timer, you have capped that.
You will not spend more than five minutes on this task, so you can get back to whatever the hell you were doing before the five minutes. This is the simplest version of an old trick called the Try Harder Luke, with a funny post on LessWrong all about this, if you want to take a look. So let's say you set a five -minute timer and you actually try and do something on your to -do list.
From here, there are three possible outcomes. The first is that you will realize that it is not worth doing, in which case you will win. Now it is off your plate.
Excellent. The second is you will do the task within five minutes, which is also a win. Nice.
It's also off your plate. The third is in attempting to do the task, you realize, shit, this is going to take more than five minutes. Well, great.
Now you know what you need to do and you can actually get into it. Now, the first two outcomes are incredible, but the last one is still quite manageable. How?
You just set another five minute task. It's just instead of this time, instead of doing the task, your goal is now planning your next five steps. Now, when I say planning your next five steps, I mean, in those five minutes, just write down the order of operations in bullet points that you need to take in order to complete the task.
So the majority of Acrasia isn't just not having a clearly defined next action. You know, most people can ultimately do what they put their minds to as long as it is in front of them. This will cut to the root of the issue by giving you a very simple and action -oriented plan.
And after that, you can spend another five -minute resolve timer actually doing the thing. With the task clearly defined, I think you'd be surprised at how far you could go to completing something in just five minutes. So let me give you an example.
Let's say your task is send tax information to accountant. It's been sitting there for the last, you know, I meant to say five weeks, but five minutes. Cycle one, what you'll do is you'll set your resolve timer and go.
So you'll try sending the tax information to your accountant. The second that you start, you'll realize, damn, that's a vague statement. What the hell is tax information?
What does send actually mean? There's a lot of kind of uncertainties here. On cycle two, you'll write out your next five actions.
And you'll just do this by pulling actions straight off the dome. For instance, step one in this case would be determine the scope of what tax information revolves around. Do I need to send like 2025?
Is it 2026? Does he want to see all my books? What exactly does he want to see?
So think back to my conversation or run through my email thread, figure that out. Number two, decide on a way to send it. Should I do email?
Should I do Google Drive? Next, compile the docs. So actually get all the files.
Fourth, write the email. And five, send. Once you have it all laid out, okay, that easily, after thinking about it for literally five minutes, you can just set another timer and then get through it.
And you can literally probably get through most of this. I know it sounds really scary at heart, but like determining the scope of tax information, that could take you like a minute. Deciding on a way to send it, that could take another two or three minutes.
You'd probably be halfway through compiling the damn docs before you actually finish that timer. And at that point, you're almost already done. So you might as well just spend a few minutes and wrap up the task, right?
That is in practice how you do it. You try and solve the task in batch number one. Then if you can't, in batch number two, you list five minute next actions.
And then in task number three, you actually just do as many of those as you can. So I will do this all the time, especially when my to -do list gets super bloated and unwieldy, which unfortunately happens maybe once a week. Today is one of those days.
Here are some practical tips that I've learned over the years to make the five -minute resolve timer work. The first is a physical timer or a desk clock is ideal just because that takes you away from your phone. If you're on your phone and you're setting their default alarm, what you'll do is you will constantly be looking at this thing and then maybe a notification will pop up or you'll feel tempted to swipe down and then boom, now you're out of flow.
The second is to write while you think. What I mean by this is if I was solving the accounting problem from before, I would actually write stuff down. Like for instance, scope of tax information, bullet points, year end, total revenue, total savings, whatever.
When you write, it immediately makes you like 20 IQ points smarter. And this is not facetious. I mean, people have done a lot of analysis on this stuff.
The quality of thought that you get when you write everything down and can reflect on it later is way higher than if it's just all in your head. The main reason for that is the four working memory chunk limit. your thoughts on paper, you've expanded your working memory limit from four things in your head to four things in your head, plus however many things you can look at simultaneously.
So what you've done is you've realistically grown that chunk limit from maybe four to six or seven. The benefit of that as well is you'll have a log or asset you can look back on later, and you can also use that to delegate a task to somebody if you do end up needing it, like say if it's a business task. The third is to define your end state beforehand.
Do you want to make a decision? Do you want to write a draft? Do you want to take a concrete step?
You cannot rely on, I have thought about it. That is not good enough. What you need is some sort of concrete end state.
And so in my case of sending the stuff to my tax account, my end state there is literally the email is sent. The guy has my information. Now, to be clear, resolve cycles don't work every single time.
Sometimes the task is too hard or it is just way too long for you to meaningfully chip at it away. But even then, at the end of a couple of timers, you can usually build a concrete list of what you need to find out or even who you need to find it out from. And that is a very far cry from, I will figure it out later.
Okay, and an action question here. What's a problem or task you guys have been thinking about for a long time? Set a five minute timer right now and try and do it.
And if you can't, set another five minute timer and write out your five next steps. Finally, I want to talk about AI. Now I teach people, hundreds of thousands of people all over the world, how to use AI.
both for their personal gain and for the gain of their business or some sort of economic activity. So on net, it is an incredible invention, okay? I have made many, many millions of dollars with AI, and I've helped people make many hundreds of millions of dollars with AI at this point.
I would consider AI to have already helped humanity accomplish some of our biggest challenges, and I firmly believe that technology will eventually bring abundance to billions, maybe even trillions of humans, if you zoom out and think about how long humanity will be around for. But like any disruptive technology, it takes a lot of time to learn how to wield a technology to maximize its benefits.
It took us a while to learn how to take fire from out in the environment and put it inside of, say, a steam engine, right? Same idea. Right now, most people are wielding fire, or AI, in ways that make them worse, not better.
And as a result, we are frying our attention spans, frying our brains, and really impacting our abilities to think clearly and deeply. What I want to do is cover a few of these problems now and then show you guys how to avoid them afterwards. The first major problem with AI is sycophancy.
Basically, modern models are post -trained to maximize the score that they receive from humans. If human raters routinely give higher scores to answers that they like, the model will naturally learn to give answers that humans like. The thing is, humans usually like answers that agree with them, obviously.
What that means is models end up being trained to agree with everything a human says in one way, shape, or form. Anthropic recently tested their five AI assistants and found lots of sycophancy across all the tasks that they measured. This is the Claude family of models.
And the results were not surprising. Human raters were significantly more likely to prefer a persuasive or flattering response over one that was just accurate or true, which is why all this happened in the first place. Now, I think most people here intuitively understand this.
When you tell a model something like, wait, isn't XYZ thing true? The model will not go, no, you are wrong. The truth is, why?
It will go, well, you raise an interesting point there, and it's the interesting half that's half true. And then it'll bury the message in a really nicely massaged length of words that'll make you feel like you were kind of right. Why is this such a big issue?
Do you guys remember maps and territories? The issue here is that AI is by definition a map. The way that neural networks work is they literally compress information about our world into a highly performant information structure that allows the efficient retrieval of a bunch of concepts and patterns.
So in this way, these models are maps of the world around them. We're increasingly using these maps to make sense of the world upon which they're trained. If you guys are heading in the wrong direction on a hike somewhere and you check your map and rather than your map making it clear that you're going in exactly the opposite way you should be, it says, hey, you're heading in the right direction.
What good is that map? It has lost all of its utility. And to be clear, this is not because of some hidden agenda or conspiracy to lie.
It is just the model doing what it was trained to do, which is maximize human ratings. You can see this in practice with a simple experiment. Next time you chat with an AI, try feeding in one of your business plans and asking, is this good?
Because you're already implying the answer that you want, the AI will almost always give you a confident, yes, here are three reasons why, or absolutely, you're getting on something here, or so on and so forth. And this may seem like an opinion, but really it's just your own thoughts dressed up and mirrored back at you. It's almost like .
So what is the solution to this? Think about it like the following. If you were to go to your mom and say, hey mom, do you think my business idea is good?
You know that your mom wants to make you happy, right? So what do you end up doing? You don't actually take her 100 % seriously.
You listen and you're like, maybe there's some nuggets in there. But for the most part, you're like, okay, my mom just loves me. So she's probably going to tell me what I want to hear.
The solution would be the same as asking your mom. Instead of asking variations of, is this a good business idea? Start asking variations of, why is this a bad business idea?
When you do this, you are no longer leaking in your hope. What you are doing is you are forcing the model to explicitly find the worst parts of your plan or your business model and then say them back to you. For instance, if you were to ask, is my plan good?
Your hope leaks in. The model will guess at the answer that you like. and then you will rate agreement above accuracy, which then loops around in this virtuous, or somewhat vicious, I should say, cycle.
But if you ask, hey, what is wrong with this idea? The model will not have any of that hope leaked in, and then it will actually be able to apply its intelligence to enumerating all of the reasons why your plan sucks, many of which will probably be true. For instance, here are four prompts that I've used many times.
The first is, hey, here's my plan, or here's my idea. Please assume this plan is a mistake. Give me the three strongest reasons that a smart, experienced person would not do this.
The next one is, hey, here's my plan. Out of 100 similar attempts at this plan, how many would get the result that I want? I also want you to go and find the reference class that you're using.
Now this is Bayes and natural frequencies plus the planning fallacy multiplier in one prompt. The other is the premortem. So here's my plan.
Unfortunately, it's six months later and the whole project failed. I want you to write the postmortem. Then you take the postmortem and then you run it multiple times and then you look at all the reasons why it might have failed.
And the last is sort of like the Fermi approximation. Hey, here's my plan. I want you to break this opportunity or idea into a bunch of pieces.
Then I want you to estimate each piece's range and then Fermi approximation my results, which is basically straight from the Fermi approximates section. So in the literature, everything I just talked about actually has a name. And that name is red teaming.
Red teaming is usually applied in a security context because you're probably going to hear a lot more about it soon. But what red teaming means is adversarially attacking something. So what you're doing when you ask an AI to pick apart your super great idea is you're asking it to adversarially attack you, your ideas, and your plans, and tell you reasons why it sucks and how it could be better.
All of these are what you actually want. You don't just want somebody to pat you on the head and say, good job. You want something to point at all the ways that you suck and actually look at that, assess whether or not the map and the territory are actually misaligned, and if they are, take steps to fix it.
That's what noticing, confusion, Ugg fields, and shoulds are all about. The second major point I want to get across is do not let AI make final decisions for you. I would never ever let an AI make a final decision for me.
So in the final sentence of any planning or stretch of thinking where you state your decision and your reasoning, you must write or verbalize all of your reasoning in your own words. The reason why is because you are the only one that will bear the consequences of that decision. You cannot blame an AI for something because it is essentially right now a search engine.
Logically, if you are the one that is going to be facing the consequences of a decision, you should be the one to articulate it. And if you start handing this off, you will find that your skills in doing all of this will degrade. Like with, uh, and flinching.
If you get in the habit of seeing a heart problem and immediately start flinching away and giving it to an AI assistant instead, you will downgrade the neurocircuitry that underlies your strong decision making. Another issue is that humans are prone to deep rationalization biases. That means that rather than reason to make a decision, we often will make a decision and then we will reason about why we made that decision and pretend that we thought about it beforehand.
Unfortunately, we do not. If the AI writes your decision for you, many of you guys will probably do the thing and then rationalize why you did the thing using the AI's writing. And like we mentioned in the expected value section, you cannot work by resulting.
If you want to improve, you cannot only work. By looking at decisions that you've made that have achieved positive outcomes, you actually need to rigorously analyze why you made the decision in the first place, because statistical chance means sometimes even poor decisions yield good outcomes. So you must be able to evaluate your own process, see why it is incorrect or correct.
And even if the outcome was poor, if your decision was right, you should keep doing it. Okay, so two main ways to use the same model. Most people will do the wrong one.
You'll ask a question and we'll answer and then you will act. The way that I'd recommend it is you draft the decision. AI will attack the decision, tell you all the reasons why it's wrong.
You will then revise and then finally actually write the decision line yourself based off the revised reasoning. Two action questions here for you. First, try red teaming an idea of yours right now.
So find an idea you have and then feed it to an agent with a prop like, hey, here's my plan. This is a bad idea, but I'd like you to explain all the reasons why. The next, I'd like you to think back.
Have you already let an AI make a decision for you at some point in the last year? If so, why and what ended up happening? Alright guys, in closing, that is a wrap.
Thank you for being with me the whole way through. What I'm going to do next is summarize everything that we've learned so far to help you encode that in your long -term memory and hopefully carry it with you for many days, weeks, months, and years. So, a summary of everything we've learned.
At the beginning, we learned how the brain works, which was the biological basis of cognition. If you guys remember, this was three components. Attention, working memory, and executive function.
We then talked about the human brain's upper limit to working memory, which was... four chunks, before discussing ways to store more information in that limited window. Do you guys remember the date that we discussed?
That was quite a while ago. But I bet you, if you sit back and think, you might be able to make it happen. Feel free to pause the video if you'd like to give it a try.
From there, we went to mice and rats. We covered enriched environments, which was the idea that mice, given enriched stimulating environments, grew 15 % more neurons in the centers of their memory without having to change anything about their effort. as well as a couple of cool studies on people that show things like how air quality immediately impact clarity of thinking, sometimes by 50 % or more.
Afterwards, we defined willpower and what it means to have strong willpower. In general, we found that people with high self -control simply do not let themselves be tempted in the first place. Then we covered choice architecture, which is the idea that the more options and choices that you have seen in your life, the more times you have probably been manipulated because every option or choice has always been arranged by someone else and for their benefit, not yours.
We saw the difference between a 12 % and a 99 .98 % organ donation number simply by changing the wording and the default version of whether or not you should opt in or opt out to having your organs donated. Afterwards, we talked friction. Now, if you move a bowl of peas 10 inches further away in a cafeteria and make it uncomfortable to grab, consumption of those peas drops 13%.
What areas of your life can you do the same thing to? In that way, if you want to do more of a thing, make it easier, and if you want to do less of a thing, make it harder. Afterwards, we spent a long time on hardware, sleep disorders, exercise, nutrition, etc.
The end result of this, if you stack them all, was around 150 -200 % improvement to your ability to think clearly and make good decisions. Once we were done with hardware, we moved to software, where I ran through expected value, which was simply the expected value equaling the sum of all of a choice's outcomes times probabilities.
We then covered power laws, which were equivalent to x equals y raised to the n. That's kind of a hard way to remember it, so put another way, most of your outputs come from one or two inputs. I'd like you guys to remember that exponential graph, where the beginning of the graph was very tall, and then it smoothly gradated down.
After that, we learned about the chain law, which was the inverse of the power law, where processes are only as strong as their weakest link. Then we covered local and global maxima. If you want to reach higher peaks, you usually have to cross a lower valley.
Then we covered attentional residue, which was a way to use notes, resume notes, to minimize how much residue sticks after you context switch from task A to task B. Then we covered akrasia, which was knowing while not doing the thing that you know that you should. And everything after this was essentially software upgrades that help you minimize the degree of akrasia you feel.
So my favorite idea in the whole course is map versus territory. Now maps are useful compressed versions of the wider world, and that's exactly why they're useful. They're compressed.
But a map is not the territory, and it is valuable for you to remember that. Then we covered Ulysses or Odysseus contracts, which were making decisions while still in a clear state of mind, such that later on when you're not in a clear state of mind, you defer to the earlier version of you. This also allows you to tie consequences to poor actions like donating money to competitors or charities you dislike.
Once we finished Ulysses contracts, we moved on to opportunity cost, which is where you learned that price and cost are actually two different things. The total cost of something isn't just the price you pay. It is also the time multiplied by what you could be doing with that time.
Then we learned about Goodhart's law, where when a measure becomes the target, it stops being a measure. We also learned how to reverse it by tying an additional measure to it and tracking if it goes up or down. In this case, think speed versus accuracy.
Think tickets closed per unit time versus tickets reopened over unit time, and so on. We then covered Chesterton's fence, which is the idea where you must understand a thing before you tear it down. Then Bayes' theorem, which is a personal favorite, where you stop asking, is this true, yes or no?
And you instead ask, how much more likely is this evidence if X is true versus if X is false? After that, we covered Fermi approximates, the three tenets of which are decompose, estimate, and multiply. Then your hourly rate and the offload rule.
We talked about marginal, which was the best thing you could be doing with your time, and then average, which is the usual thing you spend your time on, as well as supervision costs, which are the hidden opportunity cost in most delegation. We covered the planning fallacy multiplier, minus 1 .5x. Finally, we covered explore versus exploit and the optimal way to structure your explore versus exploit behavior according to the seasons of your life.
If you guys remember, if you have a long time horizon, you should explore first aggressively and then later start exploiting aggressively until your exploitation results in the initial resource being depleted. Finally, we're talking immediately actionable techniques, taps or trigger action principles, where when you ahead of time turn a decision into an if then and then wait for the cue.
We talked noticing where you catch your own confusion, flinch or should statements and use that to level up your life. Very similar to passive income. We also covered premortems rather than postmortems, which is where you use stories to find failure cases.
Then we chatted resolve cycles, which are simple five minute timer approaches that minimize your akrasia. And at the end, we covered how to use AI without letting it flatter you or make your decisions for you. So it is now time to think.
I hope you guys have enjoyed this journey as much as I have. You'll find that thinking is a skill that you can indeed learn and is a skill that will improve your life in many ways that you probably can't imagine. In my view, thinking is not about how neat or tidy your thoughts look when written down on paper.
It is not about how you can cajole an AI into doing something for you. Thinking is about the outcomes that your thoughts produce. And ultimately, your gains here are exponential.
Every quarter of your life, you will face hundreds of decisions that shape your future. Like what projects should you take on? What should you say no to?
How big of a bet should you place? When should you quit? every time that you make a slightly better decision because of the clarity of your thought, it will actually improve your ability to make the next decision.
And so the gap between you and then a less defined version of yourself does not grow linearly, but exponentially. I'm very glad you guys made it this far. Keep on thinking.
Have a lovely rest of the day.
The Hook
The bait, then the rug-pull.
Five hours is the pitch and the filter at the same time. The opening line frames the whole thing as a recovery operation rather than a productivity course: the world profits from your fog, and this is the manual for taking the fog back. What follows is one man reading a document he wrote, for three hundred straight minutes, with no cuts and no music.
Named ideas worth stealing.
06:40list
The three components of cognition
- Attention (gatekeeper)
- Working memory (workbench)
- Executive function (foreman)
Every thought is a product of what gets let in, what fits on the bench, and who decides what stays. Hardware and software upgrades are allocated to these three stats like a character sheet.
Steal fordiagnosing whether a focus problem is an input problem, a capacity problem, or a prioritization problem
1:52:17model
Hardware versus software
Hardware is biology and environment and can be improved maybe 150 to 200 percent. Software is the models you reason with and can be improved by orders of magnitude, because a better algorithm changes the class of problem you can solve at all.
Steal fordeciding whether to fix your body or your method when output stalls
54:07concept
The do-nothing diagnostic
At every transition in your day, ask what happens if you do nothing. If the answer benefits a company, that default was set for you and needs changing. If it benefits you, you already own it.
Steal forauditing a day for defaults you never chose
1:55:00model
Expected value
- List each outcome
- Assign a value
- Assign a probability
- Multiply and sum
Sum each outcome's value times its probability. The arithmetic matters less than being forced to write a number down, which makes the belief falsifiable and revisable later.
Steal forany decision costing more than a few days or ten thousand dollars
2:08:10model
The resulting quadrant
- Good decision, bad outcome: bad luck, keep going
- Good decision, good outcome: deserved
- Bad decision, bad outcome: deserved
- Bad decision, good outcome: lucky, do not repeat
Separates decision quality from outcome quality so a correct high-variance bet does not get abandoned after one loss.
Steal forreviewing a quarter's calls without letting variance rewrite the lesson
2:19:00model
Power law versus chain law
Power laws apply when choosing between systems: sort outputs, double down on the head, cut the tail. Chain laws apply inside one system: reliability is the product of the steps, so fix the weakest link and delete links wherever possible.
Steal forknowing whether a given problem calls for concentration or for hardening
2:27:24model
Local versus global maxima
- Improvements are shrinking
- More input produces no more output
- Someone less skilled on another hill is beating you
Three signs you are on a small peak, plus two ways to cross the valley: quantify the loss you will accept in advance, and allocate only 10 to 20 percent of hours to the new hill instead of jumping.
Steal forrecognizing a plateau that looks like success
2:58:20concept
Ulysses contract
- Physical distance (phone in another room)
- Accountability to others (public schedule)
- Financial stakes (pre-pay someone to donate against you)
A decision made while clear-headed that removes the tempted version of you from the loop entirely. The rope is not the point; the timing is.
Steal forany habit where willpower has already failed more than twice
3:41:40model
The offload rule
Offload when hours to do it yourself times your marginal rate exceeds the price plus supervision hours times your marginal rate. Supervision is the term people zero out and it is the term that kills the math.
Steal fordeciding what to delegate to a person or an agent
3:25:00model
Bayes in natural frequencies
Skip the notation. Ask how likely this evidence is if the claim is true versus if it is false, convert to odds, and multiply through one piece of evidence at a time.
Steal forgrading pipeline signals instead of guessing at close probability
3:32:30list
Fermi approximation
- Decompose into parts
- Estimate each part within about 3x
- Multiply
Round aggressively, accept an order-of-magnitude answer, and trust that errors in opposite directions cancel. Used as a go or no-go filter before real research.
Steal forsizing a market or a campaign in sixty seconds
3:51:40concept
Planning fallacy multiplier
Take your inside-view estimate and multiply by a fixed constant, commonly 1.5 to 3. Scope individual steps tightly internally while quoting the multiplied number externally.
Steal forclient timelines and any deadline you have to commit to publicly
3:58:20model
Explore-exploit seasons
Explore hard early when the time horizon is long, exploit hard once a winner is found, then explore again the moment returns start declining. Framed as seasons rather than personality traits.
Steal fordeciding whether this quarter is for testing or for milking
4:06:20acronym
TAP (trigger action plan)
- Trigger must be obvious
- Trigger must be visceral
- Trigger must be reliable
- Action must be small and physical
An if-then rule installed by rehearsal: write it, visualize the trigger in detail, physically run it ten times, then debug only the trigger when it misfires.
Steal forconverting any intention into behavior without relying on motivation
4:21:40list
Three signals, three five-second moves
- Confusion: say 'I notice I am confused' and ask why
- Flinch: stare at the thing for ten seconds
- Should: ask whose goal this is
A background process that compounds like an index fund. Each catch nudges your map closer to the territory without any project attached.
Steal forturning small daily discomfort into a diagnostic loop
4:35:00list
The premortem loop
- Write the plan
- Pretend a month has passed and it failed
- Write the story in past tense
- Bin the causes
- Fix the cheapest one first
Past tense makes the brain produce narratives with specific characters and dates instead of abstract risks. Fifteen minutes on one page against weeks of real failure.
Steal forstress-testing any plan before the first day of work
4:40:00list
Resolve cycles
- Cycle one: try to finish the task in five minutes
- Cycle two: write the next five concrete steps
- Cycle three: run as many steps as fit
Three outcomes are all wins: it turns out not to be worth doing, it gets done, or you finally know what it involves. Use a physical timer and write while you think.
Steal foranything that has been sitting on a list for weeks
4:50:00list
The AI red-team loop
- Draft the decision yourself
- Have the model attack it
- Revise
- Write the final decision line in your own words
Four prompt patterns: assume this is a mistake, give me a reference class out of 100 attempts, write the postmortem six months out, and Fermi it into pieces.
Steal forany plan you were about to run past a model for approval
How they asked for the click.
VERBAL ASK
05:54link
“This entire document is available just in the link in the description, so you guys can download it and then keep it as a companion guide.”
The only real pitch lands six minutes in and it is the free document, not the paid community. Maker School gets mentioned repeatedly as the source of his data (45,500 threads, 14,000 members) rather than as an ask, which makes the single description link do all the selling.
MENTIONED ON CAMERA
Visual structure at a glance.
