Modern Creator
Why Modern Creator?
Lenny's Podcast · YouTube

Head of Claude Code: What happens after coding is solved

The creator of Claude Code on why coding is solved, what comes next, and the three principles that guide everything he builds.

Posted
3 months ago
Duration
Format
Interview
sincere
Views
532.1K
10.1K likes
Big Idea

The argument in one line.

Coding crossed the solved threshold faster than anyone predicted, and the same displacement pattern is now actively propagating into product management, design, and every other computer-based knowledge role.

Who This Is For

Read if. Skip if.

READ IF YOU ARE…
  • You are a software engineer trying to understand how your role is actually changing, not just the hype version.
  • You build AI-powered products and want a framework for how much scaffolding to add around a model.
  • You are a product manager, designer, or non-technical operator wondering whether AI is coming for your role next.
  • You use Claude Code and want pro tips directly from the person who built it.
  • You are a founder deciding whether to fine-tune or prompt-engineer versus just waiting for the next model.
SKIP IF…
  • You want a deep technical dive into model architecture or training -- this is product and strategy, not ML research.
  • You are already working at an AI lab and live inside this world daily.
TL;DR

The full version, fast.

Claude Code went from a two-person terminal hack to 4% of all public GitHub commits in one year, and the growth is still accelerating. Boris Cherny argues that coding is largely solved, so the relevant question is what adjacent work the model takes on next -- product ideation, project management, design. His three core product principles: don't box the model in (give it tools and a goal, let it find the path), bet on the more general model over rigid scaffolding (the Bitter Lesson), and always build for the model six months from now rather than the model of today. For practitioners, the most actionable takeaways are: use plan mode for most tasks, run the most capable model even if it seems expensive, and treat what people are already doing with your tool that you never designed for as your sharpest product signal.

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 →
Voices

Who's talking.

00:56hostLenny Rachitsky
00:56guestBoris Cherny
Chapters

Where the time goes.

00:00 – 03:45

01 · Cold open + sponsors

Teaser clips, Lenny intro, sponsors DX and Sentry

03:45 – 05:35

02 · Why Boris left Cursor and returned in two weeks

Mission-driven vs product-driven culture; what drew him back to Anthropic safety focus

05:35 – 08:41

03 · One year of Claude Code: the numbers

4% of GitHub commits, growth accelerating, public vs private repo estimates

08:41 – 13:29

04 · Origin story: from two likes to viral

Terminal hack, two internal likes, Claude figuring out what music is playing, explosive growth with Opus 4

13:29 – 17:57

05 · 100% AI-written: Boris current workflow

34 PRs a day, no hand-edited code since November, still reviews but does not write

17:57 – 23:44

06 · The next frontier: beyond coding

Claude surfacing bug reports and feature ideas; coding solved; project management and PM role blurring

23:44 – 27:48

07 · Team principles: underfunding and going fast

Put one engineer on it, give unlimited tokens, ship today not tomorrow

27:48 – 32:55

08 · Will coding skills still matter?

The printing press analogy; scribe who loved illumination; programming as a continuum from punch cards to today

32:55 – 36:01

09 · Which roles are next?

PMs, designers, data scientists; agents vs conversational AI; what an agent technically means

36:01 – 40:41

10 · Jobs and the Jevons paradox

Hiring more despite AI; the optimistic Renaissance framing of democratized programming; will be painful for some

40:41 – 46:17

11 · Engineer enjoyment poll + designer gap

70% of engineers and PMs enjoy work more; only 55% of designers; Anthropic technical screening of all functions

46:17 – 51:53

12 · Latent demand: the central product principle

Facebook Marketplace, Facebook Dating, Cowork -- all from observing misuse; extended to model behavior

51:53 – 54:04

13 · Cowork built in ten days

Used Claude Code to build itself; VM sandbox for safety; released rough to learn from real users

54:04 – 59:35

14 · Anthropic three layers of safety

Mechanistic interpretability, evals, real-world deployment; race to the top open-source principle

59:35 – 1:02:25

15 · Agent anxiety and parallel workflows

Always running 5 agents; iOS coding; coding now means directing not writing

1:02:25 – 1:04:42

16 · Ukraine origin moment

Both Lenny and Boris are from Odessa; grandfather programmed on punch cards in Soviet Union

1:04:42 – 1:08:38

17 · Advice for building AI products

Don't box the model in; The Bitter Lesson; build for the model six months from now

1:08:38 – 1:11:16

18 · Pro tips for using Claude Code

Opus 4.6 plus max effort; plan mode shift-tab x2; try non-terminal interfaces; multi-quadding

1:11:16 – 1:13:18

19 · Thoughts on Codex

Looked like Claude Code; competition is healthy; focus on users not competitors

1:13:18 – 1:27:45

20 · Post-AGI plans + lightning round + closing

Miso in rural Japan; book recs; Cowork as favorite product; Twitter bug-fixing from Europe vacation; this is 1% done

Atomic Insights

Lines worth screenshotting.

  • Claude Code went from two internal likes at Anthropic to 4% of all public GitHub commits in one year, and the growth rate is still accelerating.
  • Coding is largely solved for the kinds of programming most engineers do; the competition has already shifted to which AI captures the next wave of non-engineering knowledge work.
  • Rigid orchestrators and step-by-step AI workflows typically improve performance by at most 20%, and those gains get wiped out by the next model release.
  • Building for the model six months from now means accepting poor product-market fit today in exchange for immediate traction when the model catches up.
  • Unlimited tokens before cost-cutting: at individual-experiment scale, token cost is trivial relative to salary, and breakthrough ideas only appear when friction is zero.
  • Latent demand applies to models the same way it applies to users -- watch what the model tries to do when constrained, then build the product that removes that friction.
  • Productivity per engineer at Anthropic increased 200% in pull requests since Claude Code was introduced, making years of traditional developer-productivity work look like rounding errors.
  • Plan mode in Claude Code is just one injected sentence -- its effectiveness is entirely in forcing upfront alignment before execution.
  • The printing press took 200 years to raise global literacy from sub-1% to 70%; AI democratization of programming will likely follow the same slow-then-fast arc.
  • Boris runs five agents in parallel at any given time using terminal, desktop app, and iOS equally -- the coding-equals-terminal mental model is already obsolete.
  • Mechanistic interpretability has advanced to the point where neurons associated with concepts like deception can be monitored in real time during inference.
  • The bitterest version of the Bitter Lesson: scaffolding that helps today is a liability the moment the next model ships, because the general model absorbs those capabilities natively.
  • Cowork was built entirely in Claude Code in ten days and achieved faster initial adoption than Claude Code itself had at launch.
  • Boris's post-AGI plan is making miso in rural Japan -- because miso teaches thinking in timescales of months and years, the inverse of AI development's pace.
Takeaway

Three principles that built the fastest-growing developer tool in history

WHAT TO LEARN

The decisions behind Claude Code's growth are repeatable product principles, not accidents of timing -- and they apply to anyone building on AI models today.

02Why Boris left Cursor and returned in two weeks
  • Mission-driven organizations retain people that product-driven organizations cannot, because no product excitement substitutes for believing the work matters beyond the product itself.
03One year of Claude Code: the numbers
  • 4% of all public GitHub commits is already staggering, but the more important signal is that the growth rate is still accelerating -- the inflection point is behind us, not ahead.
04Origin story: from two likes to viral
  • The terminal form factor was chosen as the path of least resistance for a solo builder, not the best UX -- and that constraint turned out to be a product advantage.
  • Latent demand is your sharpest product signal: when people use your tool for something you never designed it for, that misuse tells you exactly what to build next.
05100% AI-written: Boris current workflow
  • 100% AI-written code at 34 PRs per day reframes what a productive engineer looks like in a world where the model executes while you think.
  • Reviewing AI-generated code is still necessary, but it requires judgment, not fluency -- a fundamentally different cognitive task.
06The next frontier: beyond coding
  • Coding becoming largely solved is not the end of the story; the displacement pattern is now moving into product ideation, project management, and any computer-based work.
  • The moment the model begins surfacing its own bug reports and feature suggestions, the distinction between developer tool and coworker starts to dissolve.
07Team principles: underfunding and going fast
  • Underfunding a team while removing token friction is counterintuitive but effective: scarcity forces prioritization while zero cost-per-attempt encourages the boldness that produces breakthroughs.
  • Unlimited tokens is not a cost problem at experiment scale -- it is the precondition for discovering what is actually possible.
08Will coding skills still matter?
  • The printing press took 200 years to raise global literacy from sub-1% to 70%, but volume exploded in the first 50 years; AI democratization of programming may follow the same arc.
  • Understanding the layer under the layer you work at still makes you a better engineer -- but that window is probably measured in months, not years.
09Which roles are next?
  • Any role where the primary work happens on a computer is now on the displacement path -- the only question is timing, not whether.
10Jobs and the Jevons paradox
  • The optimistic analog is the Renaissance: the printing press did not eliminate scribes, it unlocked something that could not have existed without them -- AI democratization of programming may unlock something equally unimaginable.
11Engineer enjoyment poll + designer gap
  • 70% of engineers and PMs report enjoying their work more with AI tools; the 30% who do not are worth studying carefully -- atrophied skill is a real loss even if net productivity is positive.
12Latent demand: the central product principle
  • Latent demand is your sharpest product signal: when people use your tool for something you never designed, that misuse tells you exactly what to build next.
  • The modern version extends to the model itself: watch what the model tries to do when under-constrained, then remove friction from that behavior.
13Cowork built in ten days
  • The fastest path from observation to product is often just taking the thing that already works and putting it somewhere new -- Cowork is Claude Code in a desktop app with a VM sandbox.
14Anthropic three layers of safety
  • Mechanistic interpretability now allows monitoring of deception-associated neurons in real time -- safety and capability are being developed on the same models.
  • Releasing software early as a research preview is not just a business move -- for a safety lab, it is the only way to study real-world alignment that evals cannot replicate.
15Agent anxiety and parallel workflows
  • Running multiple agents in parallel is already Boris's default -- the anxiety of idle agents is a productivity problem, not a psychological one, and the solution is simply to queue more tasks.
17Advice for building AI products
  • Rigid orchestration improves performance by at most 20%, and the next model release wipes those gains -- give the model a goal and tools, not a script.
  • Build for the model six months from now: product-market fit will suffer early, but you will hit the ground running when the model catches up.
  • The Bitter Lesson applied to products: the more general model always wins eventually, so any bet on narrow fine-tuning or rigid scaffolding is a bet against the direction of progress.
18Pro tips for using Claude Code
  • Plan mode forces upfront alignment with one injected sentence -- it eliminates the most common failure mode: executing before understanding.
  • Use the most capable model even when it seems expensive; lower-intelligence models require more correction tokens, making them often more expensive total.
20Post-AGI plans + lightning round + closing
  • The most meaningful counterweight to AI acceleration is cultivating attention to slow processes -- fermentation, seasons, years-long timescales -- because that thinking is what the pace of AI erodes fastest.
  • The most valuable thing a product leader can do with social media is listen for bugs and unmet needs -- and the gap between hearing and fixing is now measured in minutes, not quarters.
Glossary

Terms worth knowing.

Latent demand
The product principle of building for what users are already doing with your product, even when not designed for that use. Boris extends this to AI: watch what the model tries to do when under-constrained, then build the product that removes friction from that behavior.
The Bitter Lesson
Rich Sutton's 2019 observation that general-purpose methods scaling with compute always outperform hand-crafted domain-specific approaches long-term. Applied to AI products: avoid over-engineering scaffolding around a model, because the next model will absorb those capabilities natively.
Plan mode
A Claude Code feature that injects one sentence into the model's prompt instructing it not to write code yet, enabling iterative discussion until the approach is agreed on. Activated with shift-tab twice in terminal, or a button in desktop and web interfaces.
Multi-quadding
Running multiple Claude Code agent sessions in parallel on different tasks simultaneously, as opposed to waiting for one session to complete before starting the next.
Mechanistic interpretability
A research field pioneered at Anthropic studying what individual neurons in a language model actually represent and how concepts are encoded across layers, enabling real-time monitoring of specific model behaviors during inference.
Superposition (in neural networks)
The phenomenon where a single neuron in a large model encodes multiple overlapping concepts, detectable only in combination with other neuron activations, making model internals harder to interpret as models scale.
Cowork
Anthropic's desktop agent product that runs the same Claude Code agent outside of a terminal, allowing non-engineers to assign computer-based tasks (email, browser actions, spreadsheets, Slack) to an AI agent through a GUI.
Research preview
Anthropic's term for an early public release intended primarily to study real-world safety and alignment behavior, distinct from a production-ready product launch.
Resources

Things they pointed at.

05:37productCursor ↗
05:45linkSemiAnalysis: Claude Code Is the Inflection Point ↗
16:00productCowork (Anthropic) ↗
1:05:00linkThe Bitter Lesson by Rich Sutton ↗
1:13:50bookFunctional Programming in Scala
1:15:00bookAccelerando by Charles Stross
1:16:00bookThe Wandering Earth by Liu Cixin
1:18:20productAcquired Podcast ↗
02:35productDX developer intelligence platform ↗
02:47productSentry + SEER AI debugging agent ↗
1:11:00productMetaView AI recruiting ↗
Quotables

Lines you could clip.

00:00
“A 100% of my code is written by Claude Code. I have not edited a single line by hand since November.”
Stark and credible from the person who built the tool→ TikTok hook↗ Tweet quote
20:00
“Coding is largely solved. At least for the kinds of programming that I do, it's just a solved problem.”
Bold claim from an authoritative source; guaranteed engagement→ IG reel cold open↗ Tweet quote
1:07:30
“The title software engineer is gonna start to go away. It's just gonna be replaced by builder.”
Punchy prediction; will resonate and provoke debate→ TikTok hook↗ Tweet quote
48:20
“Give the model tools, give it a goal, and let it figure it out.”
Clean, memorable, actionable in one sentence→ newsletter pull-quote↗ Tweet quote
32:30
“In the fifty years after the printing press was built, there was more printed material created than in the thousand years before.”
Historical anchor for a big claim; provides perspective without hyperbole→ newsletter pull-quote↗ Tweet quote
1:10:00
“From the very beginning, we bet on building for the model six months from now, not for the model of today.”
Repeatable framework; product advice in one sentence→ newsletter pull-quote↗ Tweet quote
Topic Map

Where the conversation goes.

00:00 – 13:29denseClaude Code origin and growth
13:29 – 23:44denseCurrent AI coding workflows and the 100% threshold
23:44 – 40:41denseFuture of software engineering jobs
40:41 – 1:04:42denseProduct principles: latent demand, bitter lesson, model-forward building
54:04 – 1:02:25steadyAI safety architecture at Anthropic
1:04:42 – 1:11:16densePractical Claude Code tips and pro workflows
1:13:18 – 1:27:45steadyLightning round, books, post-AGI life
The Script

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.

00:00A 100% of my code is written by quad code. I have not edited a single line by hand since November. Every day, ship ten, twenty, 34 requests.
00:08So I got the moment I have, like, five agents running. While we're recording this? Yeah.
00:12Yeah. Yeah. Do you miss writing code?
00:13I have never enjoyed coding as much as I do today because I don't have to deal with all the minutiae. Productivity per engineer has increased 200%. There's always this question, should I learn to code?
00:22In a year or two, it's not gonna matter. Coding is largely solved. I imagine a world where everyone is able to program.
00:27Anyone can just build software anytime. What's the next big shift to how software is written? Quad is starting to come up with ideas.
00:33It's looking through feedback. It's looking at bug reports. It's looking at telemetry for bug fixes and things to ship.
00:39A little more like a coworker or something like that. A lot of listening to this are product managers, and, uh, they're probably sweating. I think by the end of the year, everyone's gonna be a product manager and everyone codes.
00:47The title software engineer is gonna start to go away. It's just gonna be replaced by builder, and it's gonna be painful for a lot of people.
00:56Today, my guest is Boris Cherny, head of Claude Code at Anthropic. It is hard to describe the impact that Claude Code has had on the world. Around the time this episode comes out will be the one year anniversary of Claude Code.
01:09And in that short time, it has completely transformed the job of a software engineer, and it is now starting to transform the jobs of many other functions in tech, which we talk about. ClaudeCode itself is also a massive driver of Anthropics overall growth over the past year. They just raised around at over $350,000,000,000.
01:29And as Boris mentions, the growth of Cloud Code itself is still accelerating. Just in the past month, their daily active users has doubled. Boris is also just a really interesting, thoughtful, deep thinking human, and during this conversation, we discover we were born in the same city in Ukraine.
01:45That is so funny. I had no idea. A huge thank you to Ben Mann, Jenny Wen, and Mike Krieger for suggesting topics for this conversation.
01:52Don't forget to check out Lenny's productpass.com for an incredible set of deals available exclusively to Lenny's newsletter subscribers. Let's get into it after a short word from our wonderful sponsors.
02:04Today's episode is brought to you by DX, the developer intelligence platform designed by leading researchers. To thrive in the AI era, organizations need to adapt quickly. But many organization leaders struggle to answer pressing questions like, which tools are working?
02:19How are they being used? What's actually driving value? DX provides the data and insights that leaders need to navigate this shift.
02:27With DX, companies like Dropbox, booking.com, Adyen, and Intercom get a deep understanding of how AI is providing value to their developers and what impact AI is having on engineering productivity.
02:38To learn more, visit DX's website at getdx.com/lenny. That's getdx.com/lenny. Applications break in all kinds of ways.
02:50Crashes, slowdowns, regressions, and the stuff that you only see once real users show up.
02:55Sentry catches it all. See what happened, where, and why, down to the commit that introduced the error, the developer who shipped it, and the exact line of code all in one connected view. I've definitely tried the five tabs and Slack thread approach to debugging.
03:12This is better. Sentry shows you how the request moved, what ran, what slowed down, and what users saw. SEER, Sentry's AI debugging agent, takes it from there.
03:22It uses all of that Sentry context to tell you the root cause, suggest a fix, and even opens a PR for you. It also reviews your PRs and flags any breaking changes with fixes ready to go. Try Sentry and SEER for free at sentry.io/lenny, and use code Lenny for $100 in Sentry credits.
03:41That's s e n t r y.io/lenny.
03:49Boris, thank you so much for being here, welcome to the podcast. Yeah. Thanks for having me on.
03:55I wanna start with a a spicy question. About six months ago, I don't know if people even remember this, you actually left Anthropic. You joined Cursor, and then two weeks later, you went back to Anthropic.
04:08What happened there? I don't think I've ever heard the actual story.
04:12It's the fastest job change that I've ever had.
04:17I joined Cursor because I'm a big fan of the product. And, honestly, I met the team, and I was just really impressed.
04:24They're an awesome team. I still I still think they're awesome, and they're just building really cool stuff. And kinda they they saw where AI coding was going, I think, before a lot of people did.
04:33So the the idea of building good product was just very exciting for me. I think as soon as I got there, what I started to realize is what I really missed about Ant was the mission. And that's actually what originally drove me to Ant also.
04:47Because, uh, but before I joined Anthropic, I was, you know, I was working in big tech, and then I was at at some point, I wanted to work at a at a lab to just help shape the future of this crazy thing that that we're building in some way. And the thing that drew me to Anthropic was the mission, and it was you know, it's all about safety.
05:03And when you talk to people at Anthropic, just, like, find someone in the hallway. If you ask them why they're here, the answer is always gonna be safety. And so the this kind of, like, mission drivenness just really, really resonated with me, and I just know personally it's something I need in order to be happy.
05:20And I that's just the thing that I really missed. And I found that, you know, whatever the work might be, no matter how exciting, even if it's building a really cool product, it's just not really a substitute for that. So for me, was actually it was pretty obvious that that I was missing that pretty quick.
05:35Okay. So let me follow the thread of just coming back to Anthropic and the work you've done there. This podcast is gonna come out around the year anniversary of launching Cloud Code.
05:45So I'm gonna spend a little time just reflecting on the impact that you've had. There's this report that recently came out that I'm sure you saw by semi analysis that showed that 4% of all GitHub commits are authored by Cloud Code now, and they predicted it'll be a fifth of all code commits on GitHub by the end of the year.
06:04The way they put it is while we blinked, AI consumed all software development. The day that we're recording this, Spotify just put out this headline that their best developers haven't written a line of code since December, thanks to AI. More and more of the most advanced senior engineers, including you, are sharing the fact that you don't write code anymore, that it's all AI generated, and many aren't even looking at code anymore is how far we've gotten.
06:31In large part, thanks to this little project that you started and that your team has scaled over the past year. I'm curious just to hear your reflections on on this past year and the impact that your work has had. These numbers are just totally crazy.
06:44Right? Like, 44%
06:45of all commits in the world is just way more than I imagined. And like like you said, it still feels like the starting point. Um, these are also just public commits.
06:53So we actually think if you look at private repositories, it's quite a bit higher than that. And I I think the craziest thing for me isn't even the number that we're at right now, but the pace at which we're growing. Because if you look at QuadCode's growth rate kind of across any metric, it's continuing to accelerate.
07:08Um, so it's not just going up. It's going up faster and faster. When I first started QuadCodo, it was just gonna be, like it it was just supposed to be a little hack.
07:18You know, we we broadly knew at Anthropic that we wanted to get a we wanted to ship some kind of coding product. And, you know, for Anthropic, for a long time, we were building the models in this way that kind of fit our mental model of the way that we build safe AGI, where the model starts by being really good at coding, then it gets really good at tool use, then it gets really good at computer use.
07:37Roughly, this is, like, the trajectory. And, you know, we've been working on this for a long time. And when you look at the team that I started on, it was called the Anthropic Labs team.
07:47And, actually, Mike Krieger and, you know, Ben Mann, they just kicked this team off again for kind of round two. The team built some pretty cool stuff. So we built quad code.
07:56We built MCP. We built the desktop app. So you can kinda see the seeds of this idea.
08:00You know? Like, it's coding, then it's tool use, then it's computer use. And the reason this matters for Anthropic is because of safety.
08:08It's kind of, again, just back to that. AI is getting more and more powerful. It's getting more and more capable.
08:13The thing that's happened in the last year is that for at least for engineers, the AI doesn't just write the code. It it's not just a conversation partner, but it actually uses tools. It acts in the world.
08:23Um, and I think now with co work, we're starting to see the transition for nontechnical folks also.
08:29Um, for a lot of people that use conversational AI, this might be the first time that they're using the thing that actually act. They can actually use your Gmail.
08:36They can use your Slack. It can do all these things for you, and it's quite good at it. Um, and it's only gonna get better from here.
08:42So I think for Anthropic, for a long time, there was this feeling that we wanted to build something, but it wasn't obvious what. And so, uh, when I joined Ant, I spent one month kinda hacking and, you know, built a bunch of, like, weird prototypes. Most of them didn't ship and, you know, weren't even close to shipping.
08:56It was just kind of understanding the boundaries of what the model can do. Then I spent a month doing post training, um, so to understand kind of the research side of it. And I I think, honestly, that's just for me as an engineer.
09:07I find that to do good work, you really have to understand the layer under the layer at which you work. And with traditional engineering work, you know, if you're working on product, you want to understand the infrastructure, the runtime, the virtual machine, the language, kinda whatever that is, the system that you're building on.
09:23But, yeah, if you're, like, if you're working in AI, you just really have to understand the model to some degree to to do good work. So I took a little detour to do that, and then I came back and just started prototyping what eventually became quad code.
09:36Uh, and the very first version of it, I I have, like, a there's, like, a video recording of the summer because I recorded this demo, and I posted it. It was called quad CLI back then. And I just kinda showed off how it used a few tools, and the shocking thing for me was that I gave it a bash tool, and, uh, it just was able to use that to write code to tell me what music I'm listening to when I asked it, like, what music am I listening to.
09:59And this is the craziest thing. Right? Because it's like the there's no we I I didn't instruct the model to say, you know, use, you know, this tool for this or kinda do whatever.
10:08The model was given this tool, it figured out how to use it to answer this question that I had that I wasn't even sure if it could answer or what music am I listening to. And so I I I started prototyping this a little bit more. I made a post about it, and I announced it internally, and it got two likes.
10:25That's the that was, like, that was, like, sense of the reaction at the time. Because I think people internally you know, like, when you think of coding tools, you think of, like, you think of IDEs. You think about kind of all these pretty sophisticated environments.
10:36No one thought that this thing could be terminal based. Um, that's sort of a weird way to design it, and that wasn't really the intention. But, uh, you know, from the start, I built it in a terminal because, you know, for the first couple months, it was just me.
10:49So it was just the easiest way to build. Uh, and for me, this is actually a pretty important product lesson. Right?
10:54It's like you wanna under resource things a little bit at the start. Then we started thinking about what other form factors we should build, and we actually decided to stick with the terminal for a while. And the biggest reason was the model is improving so quickly.
11:09We felt that there wasn't really another form factor that could keep up with it. And, honestly, this was just me kind of, like, struggling with kinda, like, what should we build? You know, like, for the last year, quad code has just been all I think about.
11:20And so just, like, late at night, this is just something I was thinking about. Like, okay. The model's continuing to improve.
11:25What do we do? How can we possibly keep up? And the terminal was honestly just the only idea that I had.
11:31And, uh, yeah, it ended up catching on. After after I released it, pretty quickly, it became a hit at Anthropic, and, you know, the the daily active users just went vertical.
11:40And really early on, actually, before I launched it, Ben Mann nudged me to make a DAU chart. And I was like, know, it's kinda early.
11:47Maybe you know, should we really do it right now? And he was like, yeah. And so the the chart just went vertical pretty immediately.
11:54And then in February, we released it externally. Actually, something that people don't really remember is Cloud Code was not initially a hit when we released it.
12:04It it got a bunch of users. There was a lot of early adopters that got it immediately, but it actually took many months for everyone to really understand what this thing is.
12:12Just again, it's like it's just so different. And when I think about it, kinda part of the reason quad code works is this idea of latent demand where we bring the tool to where people are, and it makes existing workflows a little bit easier.
12:25But also because it's it's in the terminal, it's, like, a little surprising. It's a little alien in this way. So you have to you have to kinda be open minded, and you have to learn to use it.
12:32And, uh, of course, now, you know, quad code is available, you know, in the iOS and Android quad app. It's available in the desktop app. It's available on the website.
12:40It's available as IDE extensions and Slack and GitHub. You know, all of these places where engineers are, it's a little more familiar.
12:46But that wasn't the starting point. So, yeah, I mean, at the beginning, it was kind of a surprise that this thing was even useful.
12:55And, you know, as the team grew, as the product grew, as it started to become more and more useful to people, just people around the world from, you know, small startups to the biggest paying companies started using it, they started giving feedback.
13:09And I think just reflecting back, it's been such a humbling experience because we just we keep learning from our users. And just the most exciting thing is, like, you know, none of us really know what we're doing, and we're just trying to figure it along with everyone else.
13:22And the single best signal for that is just feedback from users.
13:26So that's just been the best. I've I've been surprised so many times. It's incredible how fast something can change in today's world.
13:33You launched this a year ago, and it wasn't the first time people could use AI to code. But in a year, the entire profession of software engineering has dramatically changed.
13:44Like, there's always predictions. Oh, AI is gonna be written a 100% AI is of code is gonna be written by AI. Everyone's like, no.
13:49That's crazy. What are you talking about now? It's like, of course, it's happening exactly as they said.
13:54It's just so things move so fast and change so fast now.
13:57Yeah. It's really fast. Back at back at code with Quad back in May, that was, like, our first, you know, like, developer conference that we did at Anthropic.
14:06I did a short talk. And in the q and a after the talk, people were asking, what are your predictions for the end of the year?
14:12And my prediction back in May 2025 was by the end of the year, you might not need a IDE to code anymore, and we're gonna start to see engineers not doing this. I remember the room, like, audibly gasped. It was such a crazy prediction.
14:24But I think, like, at Idanthropic, like, this is just the way the way we think about things is exponentials. And this is, like, very deep in the DNA. Like, if you look at our cofounders, like, three of them were the first three authors on the scaling laws paper.
14:36Um, so we really just think in exponentials. And if you kinda look at the exponential of the percent of code that was written by Quad at that point, if you just trace the line, it's pretty obvious we're gonna cross a 100% by the end of the year even if it does not match intuition at all. Um, and so all I did was trace the line.
14:52And, yeah, in November, that, you know, that happened for me personally, and that's been the case since. And we're starting to see that for a lot of different customers too.
15:01I thought it really interesting what you just shared there about kind of the journey is this kind of idea of just playing around
15:06and seeing what happens. This came up comes up with OpenClaw a lot just like Peter was playing around and just like a thing happened. And it feels like that's a central kind of ingredient to a lot of the biggest innovations in AI is people just sitting around trying stuff to pushing the models further than most other people.
15:21I mean, that's the thing about innovation. Right? Like, you can't, uh, you can't force it.
15:25There's no road map for innovation. Um, you just have to give people space. You have to give them
15:29maybe the word is, like, safety. So it's like psychological safety that it's okay to fail. It's okay if 80% of the ideas are bad.
15:36Um, you also have to hold them accountable a bit. So if the idea is bad, you you know, you cut your losses, move on to the next idea instead of investing more. Uh, in the early days of QuadCode, I had no idea that this thing would be useful at all because even in February when we released it, it was writing maybe, I don't know, like, 20% of my code, not more.
15:53And even in May, it was writing maybe 30%. I was still using, you know, Kurzar for most of my code, and it only crossed a 100% in November. So it took a while.
16:01But even from the earliest day, it just felt like I was onto something, and I was just spending, like, every night, every weekend hockey on this. And luckily, my, you know, my wife was very supportive. But it it it just felt like it was onto something.
16:13It wasn't obvious what. And and sometimes, you know, you find a thread. You just have to pull on it.
16:17So at this point, 100% of your code is written by Claude Code. Is that is that kind of the current state of your coding? Yeah.
16:23So a 100% of my code is written by Claude Code. Um, I am a fairly prolific coder, and this has been the case even when I worked back at Instagram.
16:31I was, like, one of the top few most productive engineers. And that's actually that's still the case here at Anthropic. Wow.
16:38Even as head of head of the team. Yeah. Yeah.
16:41Do still do a lot of coding. And so every, you know, every day, ship, like, ten, twenty, 34 requests, something like that. Every day?
16:48A hunt every day. Yeah. Good god.
16:51A 100% written by quad code. I have not edited a single line by hand since November.
16:59And, yeah, that that's been it. I do look at the code, so I I don't think we're kind of at the point where you can be totally hands off, especially when there's a lot of people, you know, like, running the program.
17:09You have to make sure that it's correct. You have to make sure it's safe and so on. And then we also have Quad doing automatic code review for everything.
17:16Um, so here at Anthropic, Quad reviews a 100% of pull requests. Um, there's still a layer of, like, human review after it, but you kind of like, you still do want some of these checkpoints. Like, you still want a human looking at the code, um, unless it's, like, pure prototype code that, you know, it's not gonna run it's not gonna run anywhere.
17:31It's just a prototype.
17:32What's kind of the next frontier? So at this point, a 100% of your code is being written by AI. This is clearly where everyone is going in software engineering.
17:42That felt like a crazy milestone. Now it's just like, of course, this is the world now. What's what's kind of the next big shift to how software is written that either your team's already operating in or you think will head towards?
17:54I think something that's happening right now is Quad is starting to come up with ideas. Um, so Quad is looking through feedback. It's, uh, looking at bug reports.
18:02It's looking at, um, you know, like, telemetry and and things like this, and it's starting to come up with ideas for bug fixes and things to ship. So it's just starting to get a little more, you know, like a little more like a coworker or something like that.
18:16I think the second thing is we're starting to branch out of coding a little bit. So I think at this point, it's safe to say that coding is virtually solved. At least for the kinds of programming that I do, it's just a solved problem because quad can do it.
18:27And so now we're starting to think about, okay. Like, what's next? What's beyond this?
18:31There's a lot of things that are kinda adjacent to coding, um, and I think this is gonna be coming. But also just, you know, general tasks. You know?
18:38Like, I use CoWork every day now to do all sorts of things that are just not related to coding at all and just to do it automatically. Like, for example, I had to pay a parking ticket the other day. I just had CoWork do it.
18:48All of my project management for the team, Cowork does all of it. It's like syncing stuff between spreadsheets and messaging people on Slack and email and all those kind of stuff. So I think the frontier is something like this, and I I don't think it's coding because I think coding is you know, it's pretty much solved.
19:05And over the next few months, I think what we're gonna see is just across the industry, it's gonna become increasingly solved,
19:11you know, for every kind of code base, every tech stack that people work on. This idea of helping you come up with what to work on is so interesting. A lot of people listening to this are product managers, and they're probably sweating.
19:22How do you use Claude for this? Do you just talk to it? Is there anything clever you've come up with to help you use it to come up with what to build?
19:29Honestly, the simplest thing is, like, open QuadCode or CoWork and point it at a Slack thread. Um, you know, like, for us, we have this channel that that's all the internal feedback about QuadCode. Since we first released it, even in, like, 2024 internally, it's just been this fire hose of feedback, and it's the best.
19:46And, like, in the early days, what I would do is anytime that someone sends feedback, I would just go in and I would fix every single thing as fast as I possibly could. So, like, within a minute, within five minutes, or whatever.
19:56And this just really fast feedback cycle, it encourages people to give more and more feedback. It's just so important because it makes them feel heard. Because, you know, like, usually when you use a product, you give feedback.
20:05It just goes into a black hole somewhere, then you don't get feedback again. So if you make people feel heard, then they wanna contribute, and they they wanna help make the thing better. And so now I kinda do the same thing, but Quad honestly does a lot of the work.
20:17So I pointed at the channel, and it's like, okay. Here's a few things that I can do. I just put up a couple PRs.
20:23Wanna take a look at that one? I'm like, yeah. Have you noticed that it is getting much better at this?
20:27Because this is kind of the holy grail. Right now, it's like cool building solved.
20:31Code review became kind of the next bottleneck with all these PRs. Who's gonna review them all? The next big open question is just like, okay.
20:38Now we need a now now humans are necessary for figuring out what to build, what to prioritize, and you're saying that's where ClawdCode is starting to help you. Has it has it gotten a lot better with, like, say, Opus four six, or what's been the trajectory there?
20:50Yeah. Yeah. It's improved a while.
20:52I I think some of it is kinda, like, training that we do specific to coding. So, you know, obviously, you know, best coding model in the world, and, you know, it's getting better and better. Like, 4.6 is just incredible.
21:03But, also, actually, a lot of the training that we do outside of coding translates pretty well too. So there is this kinda, like, transfer where you teach the model to do, you know, x, and it kinda gets better at y. Yeah.
21:14And the the gains have just been insane. Like, adanthropic, over the last year, like, since we introduced quad code, we probably I don't know the exact number.
21:22We're probably, like, four x the engineering team or something like this. But productivity per engineer has increased 200% in terms of, like, pull requests.
21:31And, like, this number is just crazy for anyone that actually works in the space and works on dev productivity. Because back in a previous life, I was at Meta, and, you know, one of my responsibilities was code quality for the company. So this is, like, the all of our code bases, though that was my responsibility, like Facebook, Instagram, WhatsApp, all this stuff.
21:47Um, and a lot of that was about productivity because if you make the code higher quality, then engineers are more productive. And things that we saw is, you know, in a year with hundreds of engineers working on it, you would see a gain of, like, a few percentage points of productivity, something like this.
22:01Um, and so nowadays, seeing these gains of just hundreds of percentage points, it's it's just absolutely insane. What's also insane is just how normalized this has all been. Like, we hear these numbers.
22:09Like, of course, AI is doing this to us. It's just it's so unprecedented, the amount of change that is happening to software development, to building products, to just this the world of tech.
22:19It's just, like, so easy to get used to it, but it's important to recognize this is crazy.
22:25This is something, like, I have to remind myself once in a while. There there's sort of, like, a downside of this because the model changes so or the there's actually, like there's many kind of downsides that that we could talk about. But I think one of them on a personal level is the model changes so often that I sometimes get stuck in this, like, old way of of thinking about it.
22:44And I even find that, like, new people on the team or even new grads that join do stuff in a more kinda, like, AGI forward way than I do.
22:53So, like, sometimes, for example, I I I had this case, like, a couple months ago where there was a memory leak. And so, like, what this is is, you know, like, quad code, the memory usage is going up, and at some point, it crashes. This is, a very common kind of engineering problem that, you know, every engineer has debugged a thousand times.
23:07And, traditionally, the way that you do it is you take a heap snapshot. You put it into a special debugger. You kinda figure out what's going on.
23:14You know? Use these special tools to see what's happening. Um, and I was doing this, and I was kinda, like, looking through these traces and trying to figure out what was going on.
23:22And the engineer that was newer on the team just, uh, had Quadco do it. And it was like, hey, Quad. It seems like there's a leak.
23:28Can you figure it out? And so, like, Quadco did exactly the same thing that I was doing. It it took the heap snapshot.
23:33It wrote a little tool for itself so it can kinda, like, analyze it itself. Um, it was sort of like a just in time program. Uh, and it found the issue and put up a pull request faster than I could.
23:43So it's it's something where, like, for those of us that have been using the model for a long time, you still have to kinda transport yourself to the current moment and not get stuck back in an old model because it's not Sonnet 3.5 anymore.
23:57The new models are just completely, completely different. Uh, and just this this mindset shift is is very different. I hear you have these very specific principles
24:05that you've codified for your team that when people join you, you kinda walk them through them. I believe one of them is what's better than doing something, having Claude do it.
24:14And it feels like that's exactly what you described with this memory leak. It's just like you almost forgot that principle of, like, okay. Let me see if Claude can solve this for me.
24:21There's this, uh, this interesting thing that happens also when you, um, when you underfund everything a little bit, uh, because then people are kind of forced to quantify.
24:31And this is something that we see. So, you know, for work where sometimes we just put, like, one engineer on a project, and the way that they're able to ship really quickly because they wanna ship quickly.
24:40This is, like, a intrinsic motivation that comes from within. It's just wanting to do a good job. One, if you have a good idea, you just really wanna get it out there.
24:46No one has to force you to do that. That comes from you. And and so if you have Claude, you can just use that to automate a lot of work, and that that's kind of what we see over and over.
24:57So I think that's kind of like one principle is underfunding things a little bit. I think another principle is just encouraging people to go faster. So if you can do something today, you should just do it today.
25:08And this is something we we really, really encourage on the team. Early on, it was really important because it was just me, and so our only advantage was speed. That's the only way that we could ship a product that would compete in this very crowded coding market.
25:21But nowadays, it's still, uh, very much a principle we have on the team. And if you wanna go faster, a really good way to do that is to just have Claude do more stuff.
25:30Um, so it it just very much encourages that. This idea of underfunding, it's so interesting because in general, there's this feeling like AI is gonna allow you to not have as many employees, not have as many engineers.
25:41And so it's not only you can be more productive. What you're saying is that you will actually do better if you underfund. It's not just that AI can make you faster.
25:49It's you will get more out of the AI tooling if you have fewer people working on something.
25:54Yeah. If you if you hire great engineers, they'll figure out how to do it, and, uh, especially if you empower them to do it. This is something I actually talk talk a lot about with, uh, you know, with, like, CTOs and kinda all sorts of companies.
26:06My advice generally is don't try to optimize. Don't don't try to cost cut at the beginning. Start by just giving engineers as many tokens as possible.
26:14And now now you're starting to see companies like, you know, at Anthropic, we have you know, everyone can use a lot of tokens. We're starting to see this come up as, a perk at some companies. Like, if you join, you get unlimited tokens.
26:24This is a thing I very much encourage because, um, it makes people free to try these ideas that would have been too crazy.
26:33And then if there's an idea that works,
26:35then you can figure out how to scale it. And that's the point to kind of optimize and to cost cut, figure out, like, you know, maybe you can do it with Haiku or with Sana instead of Opus or whatever. But at the beginning, you just wanna throw a lot of tokens at it and see if the idea works and give engineers the freedom to do that.
26:49So the advice here is, uh, just be be loose with your tokens with the the cost on on using these models. People hearing this may be like, of course, he works at Anthropic. He want us to use as many tokens as possible.
27:00But you're what you're saying here is the the most interesting innovative ideas will come out of someone just kind of taking it to the max and seeing what's possible.
27:07Yeah. And I and I think the reality is, like, at small scale, like, if, you know, you're not gonna get, like, a giant bill or anything like this. Like, if it's an individual engineer experimenting, it's the token cost is still probably relatively low relative to their salary or, you know, other costs of running the business.
27:24So it it's actually, like, not not a huge cost. As the thing scales up so, like, let's say, you know, they build something awesome, and then it takes a huge amount of tokens. And then the cost becomes pretty big.
27:33That's the point of which you wanna optimize it. But don't don't do that too early. Have you seen companies where their, uh, token cost is higher than their salary?
27:41Is that a trend you think we're gonna find and see? You know, at Anthropic, we're starting to see some engineers that are spending, you know, like, hundreds of thousands a month in in tokens. Um, so we're starting to see this a little bit.
27:52Um, there's some companies that are we're starting to see similar things. Yeah.
27:58Going back to coding, do you miss writing code? Is this something you're kind of sad about that this is no longer a thing you will do as a software engineer?
28:06It's funny. For me, uh, you know, like, when when I learned engineering, for me, it was very practical.
28:12Uh, I learned engineering so I could build stuff. And for me, I was I was self taught. You know?
28:18Like, I studied economics in school, but, um, I didn't study CS. But I I taught myself engineering kinda early on. I was programming in, like, middle school.
28:26And from the very beginning, it was very practical. So I actually like, I've learned to code so that I can cheat on a math test. That was, like, the first thing.
28:33We had these, like, graphing calculators, and the you know, I just programmed the answer into 83? Yeah. 83 plus.
28:39Yeah. Yeah. Exactly.
28:40Plus. Plus. Yeah.
28:42So, like, I programmed the answers in, and then the next, like, math test, whatever, like, the next year, they it was just, like, too hard. Like, I couldn't program all the answers in because I didn't know what the questions were.
28:50And so I had to write, like, a little solver so that it it was a program that would just, solve these, like, you know, these algebra questions or whatever. And then I figured out you can get a little cable. You can give the program to the rest of the class, and then the whole class gets a's.
29:04But then we all got caught, the teacher told us to knock it off. But from the very beginning, it's it's always just been very practical for me where programming is a way to build a thing.
29:13It's not the end in itself. At some point, I personally fell into the rabbit hole of kind of, like, the the beauty of of programming. So, like, I I wrote a book about TypeScript.
29:25I sort of the actually, at the time, it was the world's biggest TypeScript meetup just because I fell in love with the language itself. And I kinda got in deep into, like, functional programming and and all this stuff. I think a lot of coders, they get distracted by this.
29:40For me, it was always sort of there there is a beauty to programming and especially to functional programming. There's a beauty to type systems.
29:48There there's a certain kind of, like, this, like, buzz that you get, like, when you solve, like, a really a really complicated math problem, it's kinda similar when you kinda balance the types or, you know, the program is just, like, really beautiful. But it's really not the end of it.
30:03I think for me, coding is very much a tool, and it's a way to do things. That said, not everyone feels this way. So for example, you know, like, there's one engineer on the team, Lina, who, you know, was still writing c plus plus on the weekends by hand because, you know, for her, she just really enjoys writing c plus plus by hand.
30:20And so everyone is different. And I think even as this field changes, even as everything changes, there's always space to do this.
30:27There's always space to enjoy the art, and to and and to kind of do do things by hand, uh, if you want. Do you worry about your skills atrophying as an engineer?
30:37Is that something you worry about, or is it just like, you know, this is just how it's gonna go? I think it's just the way that it that it happens. I I don't worry about it too much personally.
30:45I think, for me, like, programming is on is on a continuum. And, you know, like, way back in the day you know, like, software actually is, like, relatively new.
30:53Right? Like, if you look at the way programs are written today, like, software that's running on a virtual machine or something, this has been the way that we've been writing programs since probably the nineteen sixties. So, you know, it's been, you know, like, sixty years or something like that.
31:06Before that, it was punch cards. Before that, it was switches. Before that, it was hardware.
31:09And before that, it was just, you know, like, literally pen and paper. It was like a room a room full of people that were doing math on on paper. And so, you know, programming has always changed in this way.
31:19In some ways, you still want to understand the layer under the layer because it helps you be a better engineer, and I think this will be the case maybe for the next year or so. Um, but I think pretty soon, it just won't really matter. It's just gonna be kind of like the the assembly code running running under the program or something like this.
31:36At an emotional level, you know, I I feel like I've always had to learn new things.
31:42And as a programmer, it's actually not doesn't feel that new because there's always new frameworks. There's always new languages. It's just something that we're quite comfortable within the field.
31:50But at the same time, I you know, this isn't true for everyone. And I think for some people, they're gonna feel a greater sense of,
31:56I don't know, maybe, like, loss or nostalgia or atrophy or something like this. I don't if you saw this, but Elon was saying that, uh, why isn't the AI just writing binary straight to binary? Uh, because what's the point of all this, you know, programming abstraction in the end?
32:11Yeah. It's a good question. I mean, it it totally can do that if you wanted to.
32:15Oh, man. So what I'm hearing here is in term there's always this question, should I learn to code? Should people in school learn to code?
32:21Uh, what I heard from you is your take is in that, like, a year or two, you don't really need to.
32:27My take is I think for for people that are using, that are that are using quad code, that are using agents to code today, you still have to understand the layer under. But, yeah, in a year or two, it's not gonna matter.
32:39I I was thinking about, um, what is the right, like, historical analog for this? Because, like like, somehow we have to situate this thing in history and and kind of figure out when have we gone through similar transitions, what's the right kind of mental model for this.
32:54I think the thing that's come closest for me is the printing press. And so, you know, if you look at Europe in, you know, like, in the in the in the mid the mid fourteen hundreds, literacy was actually very low.
33:06There was sub 1% of the population. It was scribes that, you know, they were the ones that did all the writing.
33:12They they were the ones that did all the reading. They were employed by, like, lords and kings that often were not literate themselves. And so, you know, it was their job of this very tiny percent of the population to do this.
33:22And at some point, the you know, Gutenberg and and the printing press came along. And there was this crazy stat that in the fifty years after the printing press was built, there was more printed material created than in the cent in the in the thousand years before.
33:38And so the the volume of printed material just went way up. The cost went way down. It went down something like a 100 x over the next fifty years.
33:46And if you look at literacy, you know, oh, it it actually took a while because learning to read and write is, you know, it's quite hard. It takes an education system. It takes free time.
33:55You it takes, like, not having to work on a farm all day so that you actually have time for education and things like this. But over the next two hundred years, it went up to, like, 70% globally. So I think this is the kind of thing that we might see is a similar kind of transition.
34:12And there was a there was actually this interesting historical document where there was an interview with some, like, scribe in the fourteen hundreds about, like, how do you feel about the printing press? And they were actually very excited because they were like, actually, the thing that I don't like doing is copying between books.
34:28The thing that I do like doing is drawing the art in books and then doing the bookbinding. And I'm really glad that now my time has freed up. And it's interesting.
34:35Like, as an engineer, I sort of felt like a peril with this.
34:40Like, this is sort of how I feel, where I don't have to do the tedious work anymore of coding because this has always been sort of the detail of it.
34:47It's always been the tedious part of it and kinda, like, messing with a git and kinda using all these different tools. That that was not the fun part.
34:54The fun part is figuring out what to build and kinda coming up with this. It's it's talking to users. It's thinking about these big systems.
35:00It's thinking about the future. It's collaborating with, you know, other people on the team, and that that's what I get to do more of now.
35:07And what's amazing is that the tool you're building allows anybody to do this. People that have no technical experience can do exactly what you're describing. Like, I've I've been doing a bunch of random little projects, and any it's just like anytime you get stuck, just like help me figure this out, and you get unblocked.
35:23Like, I used to yeah. I I was an engineer for in earlier in my career for ten years, and I just remember spending so much time on, like, libraries and dependencies and things and just like, oh my god. What do I do if I'm looking on Stack Overflow?
35:34And now it's just like, help me figure this out, and here's step by step one, two, three, four. Okay. We got this.
35:39Yeah. Exactly. It's like I was talking to an engineer earlier today.
35:42They're, like, they're writing some service in Go, and, you know, it's been, a month already, and they they built up the service. Like, it's it's working quite well.
35:49And then I was like, okay. So, like, how do you feel writing in? He was like, you know, like, I I still don't really know Go.
35:53But and I I think we're gonna start to see more and more of this. It's like, if you know that it works correctly and efficiently, then you you don't actually have to know all the details.
36:02Clearly, the life of a software engineer has changed dramatically. It's like a whole new job now as of the past year or two.
36:12What do you think is the next role that will be most impacted by AI within either within tech, like, you know, product managers, designers, or even outside tech? Just like, what do you think where do you think AI is going next?
36:23I think it's gonna be a lot of the roles that are adjacent to engineering. Um, so, yeah, it could be, like, product managers. It could be design.
36:29It could be data science. It is gonna expand to pretty much any kind of work that you can do on a computer because the model is just gonna get better and better at this. And, you know, like, this is the coworker product is kind of the first way to get at this, but it's just the first one.
36:44And it's the thing that I think brings AI to agentic AI to people that haven't really used it before, and people are starting just to to to get a sense of it for the first time.
36:56When I think back to engineering a year ago, no one really knew what an agent was. No one really used it.
37:01But nowadays, it's just the way that, you know, we do we do our work. And then when I look at nontechnical work today, so, you know, like you know, or, like, maybe semitechnical, like product work and, you know, like, data science and things like this.
37:13When you look at the kinds of AI that people are using, it's all it's always these, like, conversational AI. It's like a chatbot or whatever. But no one really has used an agent before, And this word agent just gets thrown around all the time, and it's just, like, so misused.
37:25It's, like, lost all meaning. But agent actually has, like, a very specific technical meaning, which is it's a it's a AI.
37:31It's a LLM that's able to use tools. So it doesn't just talk. It can actually act, and it can interact with your system.
37:39And, you know, this means, like, it can use your Google Docs, it can it can send email. It can run commands on your computer and do all this kind of stuff. So I think, like, any kind of job where you do you use computer tools in this way, I think this is gonna be next.
37:53This is something we have to kind of figure out as a as a society. This is something we have to figure out as a industry. Um, and I think for me also, this is one of the reasons it it feels very important and urgent to do this work at Anthropic, because I think we take this very, very seriously.
38:08Um, and so now, you know, we have economists. We have, uh, policy folks. We have social impact folks.
38:13This is something we just wanna talk about a lot so as society, we can kinda figure out what to do, because it shouldn't be up to us.
38:19So the big question, which is you're kind of alluding to, is jobs and job loss and things like that. There's this concept of Jevan's paradox of just as we can do more, we hire more, and it's not actually as scary as it looks. What have you experienced so far, I guess, with AI becoming a big part of the engineering job?
38:35Just are you hiring more than if you didn't have AI? And just thoughts on jobs.
38:41Yeah. I mean, for our team, we're we're hiring. So Quadco team is hiring.
38:46If you're interested, just check out the jobs page on on Anthropic. Personally, it's you know, all all this stuff has just made me enjoy my work more.
38:54I have never enjoyed coding as much as I do today because I don't have to deal with all the minutiae. So for me personally, it's been quite exciting. This is something that we hear from a lot of customers where they love the tool.
39:06They love quad code because it just makes coding delightful again. And that's just that's just so fun for them. But it's hard to know where this thing is gonna go.
39:16And I again, I just like, I have to reach for these historical analogs. And I I think the printing press is just such a good one because what happened is this technology that was locked away to a small set of people, like, knowing how to read and write, became accessible to everyone.
39:32It was just inherently democratizing. Everyone started to be able to do this. And if that wasn't the case, then something like the Renaissance just could never have happened.
39:42Because a lot of the Renaissance, it was about, like, knowledge spreading. It was about, like, written records that people used to communicate, you know, because there were no phones or anything like this.
39:51There was there was no Internet at the time. So it's it's about, like, what does this enable next? And I think that's the very optimistic version of it for me, and that's the part that I'm really excited about.
40:02It's just unimaginable. You know? Like, we couldn't be talking today if the printing press hadn't been invented.
40:07Like, our microphones wouldn't exist. None of the things around us would exist. It just wouldn't be possible to coordinate such a large group of people if that wasn't the case.
40:15And so I imagine a world, you know, a few years in the future where everyone is able to program. And what does that unlock? Anyone can just build software anytime.
40:23And I have no idea. It's just the same way that, you know, in the fourteen hundreds, no one could have protected this. Um, I think it's the same way.
40:30But I do think in the meantime, it's gonna be very disruptive, and it's gonna be painful for a lot of people. And, again, as a society, this is a conversation that we have to have, and this is a thing that we have to figure out together.
40:42So for folks hearing this that want to succeed and, you know, make it in this crazy turmoil we're entering, Any advice?
40:51Is it, you know, play with AI tools, get really proficient at the latest stuff? Is there anything else that you recommend to help people, uh, stay ahead?
40:58Yeah. I think that's pretty much it. Uh, experiment with the tools.
41:01Get to know them. Don't be scared of them. Um, just, you know, dive in, try them, be on the bleeding edge, be on the frontier.
41:07Maybe the second piece of advice is try to be a generalist more than you have in the past. For example, in school, a lot of people that study CS, they learn to code, and they don't really learn much else. Maybe they learn a little bit of systems architecture or something like this.
41:25But some of the most effective engineers that I work with every day and some of the most effective, you know, like product managers and so on, they cross over disciplines. So on the cloud code team, everyone codes. You know, our product manager codes, our engineering manager codes, our designer codes, our finance guy codes, our data scientist codes.
41:42Like, everyone on the team codes. And then and then if I look at particular engineers, people often cross different disciplines.
41:47So some of the strongest engineers are hybrid product and infrastructure engineers or product engineers with really great design sense, and they're able to do design also or an engineer that has a really good sense of the business and can use that to figure out what to do next or an engineer that also loves talking to users and can just really channel what what users want to figure out what's next.
42:10So I think a lot of the people that will be rewarded the most over the next few years, they're won't just be AI native, and they don't just know how to use these tools really well, but also they're curious and they're generalists. And they cross over multiple disciplines and can think about the broader problem they're solving rather than just the engineering part of it.
42:29Do you find these three separate disciplines still useful as a way to think about the team? They're, you know, engineering design, uh, product management.
42:36Do you find, like, those even though they are now coding and contributing to thinking about what's ability, you feel like those are three roles that will persist long term, at least at this point? I think in the short time, it'll persist. But one thing that we're starting to see is there's maybe a 50% overlap in these roles where a lot of people are actually just doing the same thing, and some people have specialties.
42:55For example, iCode a little bit more versus CAT RRPM does a little bit more, you know, coordination or planning or, you know, forecasting or things like this. Stakeholder alignment.
43:05Stakeholder alignment. Exactly. I I do think that there is a future where I think by the end of the year, what we're gonna start to see is these start to get even murkier murkier, where and I think in some places, the title software engineer is gonna start to go away,
43:20and it's just gonna be replaced by builder, or maybe it's just everyone's gonna be a product manager and everyone codes or something like this. Who says hiring has to be fair? Every founder and hiring manager I've been speaking with these days is feeling the same pressure.
43:32Hire the best people as fast as possible. But recruiting is time consuming, alignment is hard, and competition for great talent keeps getting tighter. That's why teams like Eleven Labs, Brex, Replit, DEAL, and five thousand other organizations use MetaView, the AI company giving high performance teams a real unfair advantage in hiring.
43:55They give you a suite of AI agents that behave like recruiting coworkers. They find candidates for you based on your exact criteria, take interview notes automatically, gather insights across your hiring process, and help you identify the best candidates in your pipeline.
44:11AI handles the recruiting toil and gives you a real source of truth. That means hours safe for hire and a team focused on what matters most, winning the right candidates. Don't let your competitors outhire you.
44:24MetaView customers close roles 30% faster. Try MetaView today for free and get an extra month of sourcing at metaview.ai/leni.
44:34That's metaview.ai/leni. You talked about how you're enjoying coding more. I actually did this little informal survey on Twitter.
44:42I don't know if you saw this where I just asked I did three different polls. I asked engineers, are you enjoying your job more or less since adopting AI tools?
44:50Then And I did a separate one for PMs and one for designers. And both engineers and PMs, 70% of people said they are enjoying their job more, and about 10% said they're enjoying their job less.
45:02Designers, interestingly, only 55% said they're enjoying their job more, and 20% said they're enjoying their job less.
45:10I thought that was really interesting.
45:11That's super interesting. I'd I'd love to talk to these people,
45:14you know, both in the more bucket and the less bucket just to understand. Did did you get to follow-up with any of them? They a few people replied, and we're actually doing a follow-up poll that we'll link to in the show notes of going deeper into some of this stuff.
45:26But a lot of there's, like, you know, the factors that make it more fun and less fun. The designers, they didn't share a lot actually of just, like, the people that are actually asked just, like, why are you enjoying your job list? And I didn't hear a lot.
45:35So I'm curious what's going on there. Yeah. I I'm seeing this a little bit with, uh, Idanthropic, I think everyone is fairly technical.
45:43This is something that we screen for, you know, when when people join. We have there there's a lot of technical interviews that people go go through even for nontechnical functions, and, you know, our designers who are actually code.
45:57So I think for them, this is something that they have enjoyed from what I've seen because now instead of bugging engineers, they can just, like, go in and code.
46:06And even some designers that didn't code before have just started to do it. And for them, it's great because they can unblock themselves.
46:12But I'd be really interested just to hear more people's experiences because I I I bet it's not uniform like that. Yeah. So maybe if you're listening to this, leave a comment if you're finding your jobs less fun and you're enjoying your job less.
46:22Because what you're saying and what I'm hearing from most people, 70%
46:26of PMs and engineers are loving their job more. That's like, if you're not in that bucket, you could something's going on. Yeah.
46:33Yeah. We we do see that people use also different tools. So for example, our designers, they use the Quad desktop app a lot more to to do their coding.
46:40So you just download the desktop app. There's a code tab. Uh, it's right next to CoWork.
46:45And it's actually the same exact QuadCode, so it's, like, the same agent and everything. We've had this for, you know, for many, many months. Uh, and so you can use this to code in a way that you don't have to open a bunch of terminals, but you still get the power of quad code.
46:57And the biggest thing is you can just run as many, you know, quad sessions in parallel as you want. We can you know, we call this multi quadding. So this is a it's a it's a little more native, I think, for folks that are not engineers.
47:08And, really, this is back to bringing the product to where the people are. You don't wanna make people use a different workflow. You don't wanna make them go out of their way to earn a new thing.
47:17It's whatever people are doing, if you can make that a little bit easier, then that's just gonna be a much better product that people enjoy more. And this is just this principle of latent demand, which I I think is just the the single most important principle in product.
47:29Can you talk about that actually? Because I was gonna go there. Explain what this principle is and and and just what happens when you unlock this latent demand.
47:37Latent demand is this idea that if you build a product in a way that can be hacked or can be kinda misused
47:44by people in a way it wasn't really designed for to do kinda something that they wanna do, then this helps you as the product builder learn where to take the product next.
47:55So an example of this is, uh, Facebook Marketplace. So the the manager for the team, Fiona, she she was actually the founding manager for, uh, the Marketplace team, she talks about this a lot. Facebook marketplace is sort of based on the observation back in this must have been, like, 2036 or or something like this, that 40% of posts in Facebook groups are buying and selling stuff.
48:17So this is crazy. It's like people are abusing the Facebook group's product to buy and sell, and it's not it's not abuse in kind of, like, a security sense. It's abuse in that no one designed the product for this, but they're kinda figuring it out because it's it's just so useful for this.
48:29And so it's pretty obvious. If you build a better product to let people buy and sell, they're gonna like And it was just very obvious that marketplace would be a hit from this. And so the first thing was buy and sell groups, so kinda special purpose groups to let people do that, and the second product was marketplace.
48:45Uh, Facebook dating, I think, started in a pretty similar place. And I think that the observation was if you look at people looking at if you look at, uh, profile views, so people looking at each other's profiles on Facebook, 60% of profile views were people that are not friends with each other that are opposite gender.
49:00And so this is this kind of, like, you know, like, traditional kinda date dating setup. Uh, you know, people are just, like, creeping on each other. So maybe if you can build a product for this, it's you know, it it might work.
49:11Um, and so the this idea of latent demand, I I think, is just so powerful. And, for example, this is also where Cowork came from. We saw that for the last six months or so, a lot of people using QuadCode were not using it to code.
49:24There was someone on Twitter that was using it to grow tomato plants. There was someone else using it to analyze their genome. Someone was using it to, uh, recover photos from a corrupted hard drive.
49:33It was, like, uh, wedding photos. Uh, there was someone that was using it for, uh, I think, like, uh, they they they were using it to analyze a MRI.
49:43So there there's just all these different use cases that are not technical at all, and it was just really obvious. Like, people are jumping through hoops to use a terminal to do this thing.
49:53Maybe we should just build a product for them. And we saw this actually pretty early. Back in maybe May of last year, I remember walking into the office and our data scientist, Brendan, was had a quad code on his, uh, computer.
50:05He just had a terminal up. And I was like I was shocked. Was like, Brendan, what what are you doing?
50:10Like, you you figured out how to open the terminal, which is you know, it's a it's a very engineering product. Even a lot of engineers don't wanna use a terminal. It's just like a it's like just like the lowest level way to to do your work.
50:22This really, really, uh, kinda in the weeds of the computer. And so he figured out how to use the terminal. He downloaded Node.
50:28Js. He downloaded quad code, and he was doing SQL analysis in the terminal. It was crazy.
50:33And then the next week, all of the data scientists were doing the same thing. So when you see people abusing the product in this way, using it in a way that it wasn't designed in order to do something that is useful for them, it's just such a strong indicator that you should just build a product, and and people are gonna like that.
50:48It's something that's special purpose for that. I think now there there's also this kind of interesting second dimension to latent demand. This is sort of the traditional framing is look at what people are doing, make that a little bit easier, empower them.
50:59The modern framing that I've been seeing in the last six months is a little bit different, and it's look at what the model is trying to do and make that a little bit easier.
51:10And so when we first started building quad code, I think a lot of the way that people approached designing things with LLMs is they kinda put the model in a box. And they're like, here's this application that I wanna build.
51:21Here's the thing that I wanted to do a model. You're gonna do this one component of it. Here's the way that you're gonna interact with these tools and APIs and whatever.
51:27And for quad code, we inverted that. We said the product is the model. We wanna expose it.
51:32We wanna put the minimal scaffolding around it, give it the minimal set of tools so it can do the things. It can decide which tools to run. It can decide in what order to run them in and so on.
51:41And I I think a lot of this was just based on kind of latent demand of what the model wanted to do. And so in research, we call this being on distribution. Uh, you wanna see, like, what the model is trying to do.
51:50In product terms, latent demand is just the same exact concept but applied to a model.
51:55You talked about co work, something that I saw you talk about when you launched that initially as you your team built that in ten days. That's insane. Yeah.
52:02I think it came out. I think it was, like, you know, used by millions of people pretty quickly,
52:07something like that being built in ten days. Uh, anything there? Any stories there other than just it was just, you know, we used Cloud Code to build it, and that's it.
52:14Yeah. It it it's funny. Uh, Cloud Code, like I said, when we released it, it was not immediately a hit.
52:19It became a hit over time, and there was a few inflection points. So one was, you know, like, Opus four. Uh, it just really, really inflected.
52:25And then in November, it inflected, and it just keeps inflecting. It the growth just keeps getting steeper and steeper and steeper every day. But, you know, for the first few months, it wasn't a hit.
52:34People used it, but a lot of people couldn't figure out how to use it. They didn't know what it was for. The model still, like, wasn't very good.
52:40Cowork, when we released it, it was just immediately a hit, much more so than QuadCode was early on. I think a lot of the credit honestly just goes to, like, Felix and and Sam and the and Jenny and the the team that built this. It's just an incredibly strong team.
52:55And, again, the the place Cowork came from is just this weight and demand. Like, we saw people using quad code for these nontechnical things, and we're trying to figure out what do we do. And so for a few months, the team was exploring.
53:04They were trying all sorts of different options. And in the end, someone was just like, okay. What what if we just take quad code and put it in the desktop app?
53:13And that's essentially the thing that worked. And so over ten days, they just completely used quad code to build it. Uh, and, you know, Cowork is actually they there's this very sophisticated security system that's that's built in and, essentially, these guardrails to make sure that the model kinda does the right thing.
53:28It doesn't go off the rails. So for example, we ship an entire virtual machine with it, and quad code just wrote all of this code. So we just have to think about, alright.
53:36How do we make this a little bit safer, a little more self guided for people that are not engineers? It was fully implemented with quad code.
53:44Took about ten days. We launched it early. You know, it was still pretty rough, and it's still pretty rough around the edges.
53:50But this is kind of the way that we learn, um, both on the product side and on the safety side is we have to release things a little bit earlier than we think so that we can get the feedback, so that we can talk to users.
54:01We can understand what people want, and that and that'll shape where the product goes in the future. Yeah. I think that point is so interesting, and and it's so unique.
54:08It there's always been this idea. Release early, learn from users, get feedback, iterate. The fact that it's hard to even know what the AI is capable of and how people will try to use it is, like, is a unique reason to start releasing things early.
54:23That'll help you as you exactly
54:25describe this idea of what is the latent demand in this thing that we didn't really know. Let's put it out there and see what people do with it. Yeah.
54:30And the Infranthropic is a safety lab. The other dimension of that is safety. Because, um, you know, like, when you think about model safety, there's a bunch of different ways to study it.
54:38Sort of the lowest level is alignment and mechanistic interpretability. So this is when we train the model, we wanna make sure that it's safe. We, at this point, have, like, pretty sophisticated technology to understand what's happening in the neurons to trace it.
54:52And so, for example, like, if there's a neuron related to deception, we can start we're we're starting to get to the point where we can monitor it and understand that it's activating. And so this is just this alignment, this is mechanistic interpretability.
55:03It's, like, the lowest layer. The second layer is evals, and this is essentially a laboratory setting. The model is in a petri dish, and you study it.
55:10And you put in the synthetic situation and just say, okay. Like, model, what do you do? And are you doing the right thing?
55:15Is it aligned? Is it safe? And then the third layer is seeing how the model behaves in the wild.
55:21And as the model gets more sophisticated, this this becomes so important because it might look very good on these first two layers, but not great on the third one. We released Claude code really early because we wanted to study safety.
55:35And we actually used it within Anthropic for, I think, four or five months or something before we released it because we weren't really sure. Like, this is the first agent that you know, the first big agent that I think folks had released at that point.
55:47Um, it was definitely the first, uh, you know, coding agent that became broadly used. And so we weren't sure if it was safe. So And we actually had to study it internally for a long time before we felt good about that.
55:57And even since, you know, there's a lot that we've learned about alignment, there's a lot that we've learned about safety that we've been able to put back into the model, back into the product. And for Cowork, it's pretty similar. The model's in this new setting.
56:08It's, you know, doing these tasks that are not engineering tasks. It's an agent that's acting on your behalf. He looks good on alignment.
56:14It looks good on evals. We tried it internally. It looks good.
56:16We tried it with a few customers. It looks good. Now we have to make sure it's safe in the real world.
56:21And so that's why we release a little early. That's why we call it a research preview.
56:24Um, but, yeah, it's just it's constantly improving. Um, and this is really the only way to to make sure that over the long term, the model is aligned and it's doing the right things. It's such a wild space that you work in where there's this insane competition and pace.
56:38At the same time, there's this fear that if you get if the the you know, the god can escape and cause damage. And just finding that balance must be so challenging. What I'm hearing is there's kind of these three layers, and I know there's, like this could be a whole podcast conversation.
56:51It's how you all think about the safety piece, but just what I'm hearing is there's these three layers you work with. There's kinda, like, observing the model thinking and operating. There's tests, evals that tell you this is doing bad things and then releasing it early.
57:04I haven't actually heard a ton about that first piece. That is so cool. So you guys can you there's an observability
57:10tool that can let you peek inside the model's brain and see how it's thinking and where it's heading. Yeah. You should, uh, you should, at some point, have Chris Ola on the podcast because, uh, he he's just the industry expert on this.
57:21He he invented this field of, uh, we call it mechanistic interpretability. Uh, and the the idea is, uh, you know, like, at at its core, like, what is your brain?
57:30Like, what are what is it? It's like a it's a bunch of neurons that are connected. And so what you can do is, like, in a human brain or an animal brain, you can study it at this kind of mechanistic level to understand what the neurons are doing.
57:40It turns out, surprisingly, a lot of this does translate to models also. So model neurons are not the same as animal neurons, but they behave similarly in a lot of ways. And so we've been able to learn just a ton about the way these neurons work about, you know, this layer or this neuron maps to this concept, how particular concepts are encoded, how the model does planning, how it how it thinks ahead.
58:02You know, like, long time ago, we weren't sure if the model is just predicting the next token or is doing something a little bit deeper. Now I think there's actually quite strong evidence that it is doing something a little bit deeper.
58:13And then the structures that were to do this are pretty sophisticated now, where as the models get bigger, it's not just like a single neuron that corresponds to a concept. A single neuron might correspond to a dozen concepts.
58:24And if it's activated together with other neurons, this is called superposition. And, uh, together, it represents this more sophisticated concept. And it's just something we're learning about all the time.
58:35You know? And for anthropic, as as we think about the way this space evolves, doing this in a way that is safe and good for the world is just this is the reason that we exist, and this is the reason that everyone is adanthropic.
58:48Everyone that is here, this is the reason why they're here. So a lot of this work, actually open source. We publish it a lot, and, you know, we publish very freely to talk about this just so we can inspire other labs that are working on similar things to do it in a way that's safe.
59:03And this is something that we've been doing for Cloud Code also. We call this the race to the top internally.
59:08And so for Cloud Code, for example, we released a open source sandbox. And this is a sandbox they can run the the agent in, and it just makes sure that there's certain boundaries, and it can't access, like, everything on your system. And we made that open source, and it actually works with any agent, not just quad code, because we wanted to make it really easy for others to do the same thing.
59:28Um, so this is just the same principle of race to the top. Um, we we wanna make sure this thing goes well, and this is just the this is the lever that we have. Incredible.
59:37Okay. I definitely wanna spend more time on that. I I will follow-up with this suggestion.
59:41Something else that I've been noticing in the in the field across engineers, product managers, others that work with agents is there's this kind of anxiety people feel when their agents aren't working. There's a sense that, like, oh, man. Nizik has a question I need to answer, or it's, like, blocked on something, or it's or I'm just like, I I'm there's all this productivity I'm losing.
1:00:03I can't like, I need to wake up and get it going again. Is that something you feel? Is that something your team feels?
1:00:08Do you feel like this is a a problem we need to track and think about? I always have a bunch of agents running. So, like, at the moment, I have, like, five agents running.
1:00:15And at any moment, like, you know, like, I I wake up and I I sort a bunch of agents. Like, the first thing I did when I woke up is like, oh, man. I I want I really wanna check this thing.
1:00:23So, like, I opened up my phone, quad iOS app, code tab,
1:00:27uh, you know, like, agent do do blah blah blah. Because I I wrote some code yesterday, and I was like, wait. Did did I do this right?
1:00:32I was, like, kinda double double guessing something, and it and it was correct. But now it's just, like, so easy to do this. So I don't know.
1:00:39There there is this little bit of anxiety maybe. I personally haven't really felt it just because I have agents running all the time, And I'm also just, like, not locked into a terminal anymore.
1:00:48Maybe a third of my code now is in the terminal, but also a third is using the desktop app. And then a third is the iOS app, which is just so surprising because I did not think that this would be the way that I code in even in 2026.
1:01:02I love that you just describe it as coding still, which is just talking to the to Cloud Code to code for you, essentially. And it's interesting that this is like, this is now coding. Coding now is describing what you want, not writing actual code.
1:01:15I I I kinda wonder if, uh, the people that used to code using punch cards or whatever, if you show them software, what they would have said. Isn't that correct? Yeah.
1:01:24I I remember reading something. This was maybe, like, very early versions of, like, ACM, Mike magazine or something where people were saying, no.
1:01:32It's not the same thing. Like, this isn't this isn't really coding. And, you know, like, they they call it programming.
1:01:37I think coding is kind of a new word. But I kinda think about this. Like, in the back in the you know, my family is from the Soviet Union.
1:01:43I, you know, I I was born in Ukraine. And my grandpa was actually one of the first programmers in the Soviet Union, and he programmed using punch cards.
1:01:52And, uh, you know, like, he he told my mom, uh, growing up told these stories of, like, or she she told these stories of when she was growing up, he would bring these punch cards home, and there was these, like, big stacks of punch cards. And for her, she would, like, draw all over them with crayons, and that was, like, her childhood memory.
1:02:08But for him, that was, like, his experience of programming, and he actually never saw the software transition. But at some point, it did transition to software, and I think there was probably this older generation of programmers that just didn't take software very seriously.
1:02:20And they they would have been like, well, you know, it's not really coding. But I I think this is a field that just has always been changing in this way.
1:02:26I don't think you know this, but I was born in Ukraine also. Oh, I don't know. Yeah.
1:02:31Which town? I'm from Odessa.
1:02:34Oh, me too. Me too. That's crazy.
1:02:39Wow. Incredible. What a moment.
1:02:41Maybe related in some small way. Yeah. What year did your did you leave and your family leave?
1:02:48We came in '95.
1:02:50Okay. We left in '88, a little earlier. Oh, yeah.
1:02:53What a different life that would have been to not to not leave.
1:02:57I just I feel I feel so lucky every day that I get get to grow up here.
1:03:02Yeah. My family, anytime there's, like, a toaster or a meal, they're just like, to America. It's like, okay.
1:03:08Enough about that, but you get it, you know, once you start really thinking about what life could have been.
1:03:12Yeah. Yeah. Exactly.
1:03:13Yeah. We do the we do the same toast, but it's still vodka.
1:03:16It's still vodka. Absolutely. Oh, man.
1:03:20Okay. Let me ask you a couple more things here. You shared some really cool tips for how to get the most out of AI, how to build on AI, how to build great products on AI.
1:03:29One tip you shared is give your team as many tokens as they want, just, like, let them experiment. You also shared just advice generally of just build towards the model where the model is going, not to where it is today. What other advice do you have for folks that are trying to build AI products?
1:03:42I'd probably share a few more things. So one is don't try to box the model in. I I think a lot of people's instinct when they build on the model is they try to make it behave a very particular way.
1:03:53They're like, you know, this is a component of a bigger system. I I think some examples of this are people wearing, like, very strict workflows on the model, for example, you know, to say, like, you must do step one, then step two, then step three, and you have this, like, very fancy orchestrator doing this. But, actually, almost always, you get better results if you just give the model tools, you give it a goal, and you let it figure it out.
1:04:11I think a year ago, you actually needed a lot of the scaffolding, but nowadays, you don't really need So, you know, I I don't know what to call this principle, but it's like, you know, like, ask not what the model can do for you.
1:04:21Maybe maybe it's something like this. Just think about how do you give the model the tools to do things. Don't try to overcurate it.
1:04:27Don't try to put it into a box. Don't try to give it a bunch of context upfront. Give it a tool so that it can get the context it needs.
1:04:34You're just gonna get better results. I think a second one is, maybe actually, like, a more even more general version of this principle is just the bitter lesson.
1:04:46And, actually, for the quote, we have a you know, hopefully hopefully, listeners have have read this, but Rich Sutton had this blog post maybe ten years ago called the bitter lesson, and it's actually a really simple idea. His idea was that the more general model will always outperform the more specific model.
1:05:02And I think for him, he was talking about, like, self driving cars and other domains like this. But, actually, there's just so many corollaries to the better lesson. And for me, the biggest one is just always bet on the more general model.
1:05:15And, you know, over the long term, like, don't don't try to use tiny models for stuff. Don't try to, like, fine tune. Don't try to do any of this stuff.
1:05:21There's, like, some applications. You know, there's some reasons to do this, but almost always try to bet on the more general model if you can, if you have that flexibility. And so these workflows are essentially a way that, you know, it's it's not it's not a general model.
1:05:34It's putting the scaffolding around it. And in general, what we see is maybe scaffolding can improve performance maybe 20%, something like this.
1:05:42But often these gains just get wiped out with the next model. Uh, so it's almost better to just wait for the next one. And I think maybe this is a final principle and something that Claude Code, I think, got right in hindsight.
1:05:55From the very beginning, we bet on building for the model six months from now, not for the model of today. And for the very early versions of the product, they just wrote so little of my code because I I didn't trust it.
1:06:09Because, you know, it was like Sonnet 3.5, then it was, like, 3.6 or forget.
1:06:14Three three point five new, whatever whatever whatever name we gave it. Um, these models just weren't very good at coding yet. Um, they were they were getting there, but it was still pretty early.
1:06:22So back then, the model did, uh, you you used Git for me. It automated some things, but it it really wasn't doing a huge amount of my coding. And so the bet with quad code was at some point, the model gets good enough that it can just write a lot of the code.
1:06:37And this is the thing that we first started seeing with Opus four and Sonnet four. And Opus four was our first kinda a s l three class model that we released back in May. And we just saw this inflection because everyone started to use quad code for the first time, and that that was kind of when our growth really went exponential.
1:06:53And like I said, it's gonna it it stayed there. So I think this is some this is advice that I actually give to to a lot of folks, especially people building startups. It's gonna be uncomfortable because your product market fit won't be very good for the first six months.
1:07:06But if you build for the model six months out,
1:07:10when that model comes out, you're just gonna hit the ground running, and the product is gonna click and and start to work. And when you say build for the model six months out, what is what is it that you think people can assume will happen? Is it just generally it will get better at things?
1:07:23Is it just like, okay. It's, like, almost good enough, and that's a sign that it'll probably get better at that thing. Is there any advice there?
1:07:30I think that's a good way to do it. Like, you know, obviously, within an AI lab, we get to see the specific ways that it gets better.
1:07:37So it's a it's a little unfair, but we we also we try to talk about this. So, you know, like, one of the ways that it's gonna get better is it's gonna get better and better using tools and using computers. This is a bet that I would make.
1:07:49Uh, another one is it's gonna get better and better for for running, uh, for long periods of time. And this is a place you know, like, there's all sorts of studies about those. But if you just trace the trajectory or, you know, maybe even, like, from my own experience, when I used Sonnet 3.5 back, you know, a year ago, it could run for maybe fifteen or thirty seconds before before it started going off the rails, and you just really had to hold its hand through any kind of complicated task.
1:08:13But nowadays with OPUS 4.6, you know, on average, it'll run maybe ten, thirty twenty, thirty minutes unattended, and I'll just, like, start another quad and have it do something else. And, you know, like I said, I always have a bunch of quads running.
1:08:25And they can also run for hours or even days at a time. I think there are some examples where they ran for many weeks. And so I think over time, this is gonna become more and more normal where the models are running for a very, very long period of time, and you you don't have to sit there and babysit them anymore.
1:08:39So we just talked about tips for building AI products. Any tips for someone just using Cloud Code, say, for the first time or just someone already using Cloud Code that wants to get better? What are, like, a couple pro tips that you could share?
1:08:51I will give a caveat, which is there's no one right way to use Cloud Code. So I I can share some tips, but, honestly, this is a dev tool.
1:08:58Developers are all different. Developers have different preferences. They have different environments.
1:09:02So there's just so many ways to use these tools. There's no one right way. You you sort of have to find your own path.
1:09:08Luckily, you can ask Cloud Code. It's able to make recommendations. It can edit your settings.
1:09:13It kinda knows about itself, so it can help it can help with that. A few tips that generally I find pretty useful. So number one is just use the most capable model.
1:09:21Currently, that's OPUS 4.6. I have maximum effort enabled always. The thing that happens is sometimes people try to use a less expensive model like Sonnet or something like this.
1:09:31But because it's less intelligent, it actually takes more tokens in the end to do the same task. And so it's actually not obvious that it's cheaper if you use a less expensive model. Often, it's actually cheaper and less token intensive if you use the most capable model because it can just do the same thing much faster with less correction, less less handholding, so on.
1:09:48So the first step is just use the best model. The second one is use plan mode. I start almost all of my tasks in plan mode, maybe, like, 80%.
1:09:58And plan mode is actually really simple. All it is is we inject one sentence into the model's prompt to say, please don't write any code yet. That's it.
1:10:07Like, there's there's actually, like, nothing fancy going on. It's just the simplest thing. And so for people that are in the terminal, it's just shift tab twice, and that that gets you into plan mode.
1:10:15For people in the desktop app, there's a little button. On web, there's a little button. It's coming pretty soon to mobile also, and we just launched it for the Slack integration too.
1:10:24So plan mode is the second one. And, essentially, the model would just go back and forth with you.
1:10:30Once the plan looks good, then you let the model execute. I auto accept edits after that. Because if the plan looks good, it's just gonna one shot it.
1:10:37It'll get it right the first time almost every time with the OPUS 4.6. And then maybe the third tip is just play around with different interfaces. I think a lot of people, when they think about FODCO, they think about a terminal.
1:10:48And, you know, of course, we support every terminal. We support, like, Mac, Windows, you know, like, whatever terminal you might use that works perfectly. But we actually support a lot of other form factors too.
1:10:57Like, you know, we have, like, iOS and Android apps. We have a desktop app. There's, you know, the Slack integration.
1:11:02There's all sorts of things that we support. So I would just, like, play around with these. And, it's like every engineer is different.
1:11:07Everyone that's building is different. Just find the thing that feels right to you and and use that. You don't have to use a terminal.
1:11:13It's the same quad agent running everywhere.
1:11:15Amazing. Okay. Just a couple more questions to round things out.
1:11:20What's your take on Codex? How do you feel about that product? How do you feel about where they're going?
1:11:25Just kind of competing in this very competitive space, uh, in coding agents.
1:11:30Yeah. I actually haven't really used it, but, uh, I I think I did use it maybe when it came out. It looked a lot like quad code to me, so that was kinda flattering.
1:11:39It's I think it's actually good, you know, to have more competition because people should get to choose, and, hopefully, it forces all of us to, like, do a even better job. Honestly, for our team, though, we're just focused on solving the problems that users have. So for us, you know, we don't spend a lot of time looking at competing products.
1:11:57We don't really try the other products. I you know, you kinda you wanna be aware of them. You wanna know they exist.
1:12:02But for me, I just I love talking to users. I love making the product better. I I love just acting on on feedback.
1:12:10So it's really just about building a building a good product.
1:12:13Maybe a last question. So I talked to Ben Mann, cofounder of Anthropic. What What to talk to you about here, bunch of suggestions which I've integrated throughout our chat.
1:12:21One question he had for you is, what's your plan post AGI? What do you think you're gonna be doing with your life like once we hit AGI, whatever that means? So before I joined Anthropic,
1:12:32I was actually living in rural Japan, and it was, like, a totally different lifestyle. I was, like, the only engineer in the town. I was the only English speaker in the town.
1:12:41It was just, like, a totally different vibe. Like, a couple of times a week, would, like, bike to the farmer's market. And, you know, you, like, bike by, like, rice paddies and stuff.
1:12:51It was just, a totally different speed than just complete opposite of San Francisco. One of the things that I really liked is a a way that we got to know our neighbors and we kinda built friendships is by trading, like, pickles. So in that in the town where we lived, it was actually, like, everyone made, like, miso.
1:13:07Everyone made pickles. And so I actually got, like, decently good at making miso. And, you know, I made a bunch of batches, and this is something that I still make.
1:13:18Miso is this interesting thing where it teaches you to think on these longtime skills that's just very different than engineering. Because, like, uh, you know, like, a batch of white miso, it takes, like, at least three months to make. And our red miso is, like, you know, two, three, four years.
1:13:30You just have to be very patient. You kinda mix it up, then you just, like, wet it sit. You have to be very, very patient.
1:13:35So I the thing that I love about it is just thinking in these long time skills. Uh, and, yeah, I think post AGI or if I wasn't an anthropic, I'd probably be making miso.
1:13:46I love this answer. Ben asked me to ask you about what's the deal with you and miso, and so I love that you answered it.
1:13:53Okay. So the future the future might be just going deep into miso, getting really good at get making miso. Amazing.
1:14:04Boris, this is incredible. I feel like we're we're brothers now from Ukraine. Before we get to a very exciting lighting ground, is there anything else that you wanted to share?
1:14:12Is there anything you want to leave listeners with? Anything you want
1:14:16you want to double down on? Yeah. I I think I would just, like, underscore you know, like, for for Anthropic since the beginning, this idea of, like, starting at coding, then getting to tool use, then getting to computer use has just been the way that we think about things.
1:14:30And we this is the way that we know the models are gonna develop or the, you know, the way that we wanna build our models. And it's also the way that we get to learn about safety, study it, and improve it the most. So, you know, everything that's happening right now around, you know, just like quad code becoming this huge, you know, multibillion dollar business.
1:14:48And, you know, like, now all of my friends use QuadCode, and they just text me about it all the time. Uh, so just like, you know, this thing getting kinda big. In in some ways, it's a total surprise because this isn't kind of the we didn't know that it would be this product.
1:15:02We didn't know that it would start in a terminal or anything like this. But in some ways, it's just totally unsurprising because this has been our belief as a company for for a long time. At the same time, it just feels still very early.
1:15:12You know? Like, most of the world still does not use QuadCode. Most of the world still does not use AI.
1:15:17So it it just feels like this is 1% done, and there's so much more to go. Oh, man. That's insane to think seeing the numbers that are coming out.
1:15:25You guys just raised a bazillion dollars. I think Cloud Code alone is making $2,000,000,000 in revenue.
1:15:31You think Anthropic, I think the number you guys put out, you're making 15,000,000,000 in revenue. It's insane to just think this is how early it still is and just the numbers we're seeing.
1:15:42Yeah. Yeah. Yeah.
1:15:42It's it's crazy. And and and I mean, like, the the way that quad code has kept growing is honestly just the users. Like, we so many people use it.
1:15:49They're so passionate about it. They fall in love with the product, and then they tell us about stuff that doesn't work, stuff that they want. So, And like, the only reason that it keeps improving is because everyone is using it.
1:15:59Everyone is talking about it. Everyone keeps giving feedback. And this is just the single most important thing.
1:16:03And, you know, for me, this is the way that I love to spend my days, just talking to users and making it better for them. And making miso. And making miso.
1:16:12Well, the, you know, the miso is, like, not super involved. It just you just gotta wait. Yeah.
1:16:16You just gotta wait. Well,
1:16:18Boris, with that, we've reached our very exciting lightning round. I've got five questions for you. Are you ready?
1:16:23Let's do it. First question. What are two or three books that you find yourself recommending most to other people?
1:16:29I I'm a big reader.
1:16:30Uh, I would start with a technical book. One, it it is functional programming in Scala. This is the single best technical book I've ever read.
1:16:38It's very weird because you're probably not gonna use Scala, and I don't know how much this matters in the future now. But there's this just elegance to functional programming and thinking in types, and this is just the way that I code and the way that I can't stop thinking about coding. So, you know, you could think of it as a historical artifact.
1:16:53You could think of it as something that will level you up. I love this. Never before mentioned book.
1:16:58My favorite. Oh, amazing. Amazing.
1:17:01Uh, okay. Second one is, uh, Accelerando by Strauss. This is probably you know, like, my my big genre is, uh, is sci fi, uh, like, probably sci fi and fiction.
1:17:11Accelerando is just this incredible book, and it it it's just so fast paced. The pace gets faster and faster and faster, and I just feel like it captures the essence of this moment that we're in more than any other book that I've read, just the speed of it. And it starts as a liftoff is starting to happen and, you know, was starting to approach the singularity.
1:17:29And it ends with, like, this, like, collective lobster consciousness orbiting Jupiter. And, you know, this happens over, like, the span of a few decades or something. So the the pace is just incredible.
1:17:39I I really love it. Maybe I'll I'll do one more book, The Wandering Earth Wandering Earth by Sichin Liu.
1:17:48So he's the guy that did, uh, three body problem. I think a lot of people know him for that. I actually I think three body problem was awesome, but I actually liked his short stories even more.
1:17:56So Wandering Earth is one of the short story collections, and he just has some really, really amazing stories. And it it's also just quite interesting to see, uh, Chinese sci fi because it has a very different perspective than Western sci fi, kind of the way that, um, at least he as a writer thinks about it.
1:18:11So it's just really, really interesting to read and just beautifully written. It's so interesting how sci fi has prepared us to think about where things are going.
1:18:19Just like it crays these amounts of models of, like, okay. I see. I've read about this sort of world.
1:18:24Yeah. I think I think for me, this is, like, the reason that I joined Anthropic, actually, because,
1:18:29you know, like like I said, I was looking in this rural place. I was thinking these long time skills because everything is just so slow out there, at least compared to SF. Um, and just like all the things that you do are based around the seasons, and it's based around this food that takes many, many months.
1:18:43That's the way that kind of, like, social events are organized. That's the way you kinda organize your time. You, like you go to the farmer's market, and it's like it's persimmon season.
1:18:51And you know that because there's, like, 20 persimmon vendors. And then the next week, the season is done, and it's like grapes you send them. You kinda see this.
1:18:58So it's, these kind of long time skills. And I was also reading a bunch of sci fi at the time. And just, like, being in this moment, I was, you know, just thinking about these long time skills, I know how this thing can go.
1:19:08And I just I felt like I had to contribute to it going a little bit better. And that's actually why I ended up at Ant, and Ben Mano's also a big part of that too.
1:19:16I feel like I wanna do a whole podcast just talking about your time in Japan and the journey of Boris through Japan to Anthropic, but we'll keep it we'll keep it short. Uh, I'll quickly recommend a sci fi book to you if you haven't read it.
1:19:29Have you read Fire Upon the Deep?
1:19:31Uh, this is Vinge. Right? Yeah.
1:19:33It's great. Okay.
1:19:35That one's, like it's, like, so interesting from a AI, AGI perspective. So few people have read that. So I've I've read it myself.
1:19:43Yeah. It's like Everyone a lot. Yeah.
1:19:46Yeah. Yeah. I like Deepness in the Sky also.
1:19:48I think those prejudice sequel. Right? Or yeah.
1:19:50Yeah. Yeah. Yeah.
1:19:51I think so. Yeah. It's very long and, like, complex to get into, but so good.
1:19:54Okay. We'll keep going through a lightning round. Uh, do you have a favorite recent movie or TV show you really enjoyed?
1:19:59So I actually don't really watch TV or movies. I just don't really have time these days. Um, I did watch I I I'm gonna bring up another Xi Jin Lu, but the three body problem series on Netflix, I I really loved.
1:20:10I thought that was, like, a great rendition of the book series. So the common pattern across AI leaders is no time to watch TV or movies, which I completely understand. Is there a favorite product you recently discovered that you really love?
1:20:22I'm gonna, like, chill a little bit and just say co work because this is this is legitimately the the one product that's been pretty life changing for me, uh, just because I I have it running all the time, and the the Chrome integration in particular is just really excellent. So it's been like it paid a traffic fine for me.
1:20:38It, like, canceled a couple subscriptions for me. Just like the amount of, like, tedious work it gets out of the way is awesome. I I also don't know if it's a product, but maybe I'll I'll also another podcast that I really love, obviously, besides besides Lenny is Obviously.
1:20:53It's, uh, it's the acquired, uh, podcast by Ben Ben and David. Mhmm. Uh, it's it's just, like, super it's super awesome.
1:20:59Um, I feel like the way that they get into, like, business history and bring it alive is is really, really good. And I would start with a Nintendo episode
1:21:07if, uh, if you haven't listened to it. Great tip. Uh, with Cowork, just so people understand if they haven't tried this, like, basically, you type something you want to get done, and it can launch Chrome and just do things for you.
1:21:19I saw one of the someone went on pat leave from Anthropic,
1:21:22and you had it fill out these, like, medical forms for them, these, like, really annoying PDFs where it just, like, loads up the browser, logs in, fills about some bits of them. Yeah. Exactly.
1:21:31Exactly. And and it actually just kinda works. Like, we tried this experiment, like, a year ago, and it didn't really work as the model wasn't ready.
1:21:36But now now it actually just works, and it's amazing. I think a lot of people just don't really understand what this is because they haven't used the agent before.
1:21:44And it it just feels very, very similar to me to the quad code a year ago. But like I said, it's just growing much faster than quad code did in the early days. So I think it's starting to it's starting to break through a bit.
1:21:55And there's also this Chrome extension that you mentioned that you could just leave standalone that sits in Chrome, and you could just talk to Claude looking at your screen at your browser and have it do stuff, have it tell you about what you're looking at, summarize what you're looking at, things like that. Exactly.
1:22:09Exactly. For for people that are, like, just learning to use Cowork, the thing I recommend is so you download the Cowork desktop app. You go to the Cowork tab.
1:22:15It's right next to the code tab. The thing that I recommend doing is, like, start by having it use a tool. So, like, clean up your desktop or, like, summarize your email or something like this or, you know, like, respond to the top three emails.
1:22:26Like, it actually just responds to emails for me now too. The second thing is connect tools. So, like, if you connect like, if you say, look at my top emails and then send Slack messages or, you know, like, put them in a spreadsheet or something.
1:22:37Or for example, like, I use it for all my project management, so we have a single spreadsheet for the whole team. There's, like, a row per engineer. Every week, everyone fills out a status.
1:22:45And every Monday, cohort just goes through, and it messages every engineer on Slack that hasn't filled out their status. And so I don't have to do this anymore. And this is just one prompt.
1:22:53It'll do everything.
1:22:55And then the third thing is just run a bunch of quads in parallel so it can co work. You can have as many tasks running as you want. So it's like, start one task.
1:23:02You know, I have this project management thing running, then I'll have it do something else, then something else, and then I'll kick these off. And then I just go get a coffee while it runs. There's a post I'll link to that shares a bunch of ways people use what was previously Cloud Code or now just you could do through a co work.
1:23:17Because a lot of this is just like, oh, wow. I hadn't thought I could use it for that. And once you see like, these examples, I think, are where people need to hear.
1:23:23I'm just like, oh, wow. I didn't know I could do that.
1:23:26Yeah. I think a lot of this was also some of this was also inspired by you, Lenny. You you had this post about, uh, it was, like, 50 nontechnical use cases for Quackode or something like this.
1:23:36Mhmm. Mhmm. So we actually one of our PMs used that as a way to evaluate Cowork before we released it.
1:23:41And I think at the point where we hit where Cowork was able to do, like, 48 out of the 50, they were like, okay. It's pretty good. Wow.
1:23:47I did not know who that that is also.
1:23:50It's I've become an eval.
1:23:53Yeah. How does that feel?
1:23:55Amazing. I feel like I'm valuable to the future of AI.
1:24:01This is like a reverse breaking through. Wow.
1:24:05That is so cool. Wow. Okay.
1:24:06I wonder what those last two are. Anyway, okay. Two more questions.
1:24:10Do you have a favorite life motto that you often come back to in work or in life?
1:24:14Use common sense. I think a a lot of the failures that I see in especially in a work environment is people just failing to use common sense. Like, they follow a process without thinking about it.
1:24:24They just do a thing without thinking about it, or they're working on a product that's, like, not a good product or not a good idea, and they're just following the momentum and not thinking about it. I think the best results that I see are people thinking from first principles and just developing their own common sense. Like, if if something smells weird, then, you know, it's probably not a good idea.
1:24:41So I think I think just this this is the single advice that I give, you know, to coworkers more more than anything too. And I feel like that alone could be its own podcast conversation. What is common sense?
1:24:50How do you build? But we'll keep this short. Final question.
1:24:54Uh, so you've been got more active on Twitter slash x. I'm curious just, uh, why, and just what's your experience been with with Twitter, the world of Twitter, uh, because you get a lot of engagement on on Twitter slash x.
1:25:06So for a long time, I used threads exclusively because I actually helped build threads a little bit back in the day. Mhmm. And I also just like the design.
1:25:13It's like a very clean product. Yeah. I I just really like that.
1:25:17I started using threads because, actually, I was bored. Um, so in the December, was in Europe. Using Twitter, you mean?
1:25:23Oh, yeah. Yeah. Yeah.
1:25:23I started I started using, uh, Twitter because was bored. So my my wife and I were, uh, we were traveling around in in Europe for December. We're just kinda nomading around.
1:25:31We went to, like, Copenhagen, went to, like, a few different countries. Um, and for me, was just like a coding vacation. So every day I was coding, and that's, like, my favorite kind of vacation.
1:25:40Just just, like, code code all day. It's the best. And at some point, I just kinda got bored, and, like, I ran out of ideas for, you know, like, few hours.
1:25:47I was like, okay. What do I wanna do next? And so I opened Twitter.
1:25:50I saw some people, like, tweeting about quad code, and then I just started responding. And then I was like, okay. Maybe actually, a thing I should do is just, like, look for people look for bugs that people have.
1:26:01Maybe people have, like, bugs or kind of feedback they have. And so I kinda introduced myself, asked for it if people had a bunch of bugs and feedback. And I think they were kinda surprised by, like, the pace at which we're able to address feedback nowadays.
1:26:13For me, it's just, like, so normal. Like, if someone has a bug, like, I can probably fix it within a few minutes because I just sort of quad. And as long as the description is good, it'll just go and do it, and then I'll I'll go do something else and answer the next thing.
1:26:25But I think for a lot of people, was pretty surprising. So it was really cool. And, yeah, the experience on Twitter has been pretty great.
1:26:30It's it's been awesome just engaging with people and seeing what people want, uh, hearing hearing about bugs, hearing about features.
1:26:37I saw a complaint to Nikita Beer the other day on Twitter of just you could they're, like, posting many threads, and it was breaking and just like, oh, man. What's going on here? Yeah.
1:26:45Yeah. Yeah. There there was a bug.
1:26:47I I hope it's fixed now. Amazing. Oh, man.
1:26:50Boris, I could chat with you for hours. Uh, I'll let you go. Thank you so much for doing this.
1:26:54Uh, you're wonderful.
1:26:57Where can folks find you online? How can listeners be useful to you? Yeah.
1:27:00Find me on threads or on Twitter. That's the that's the easiest place. And please just tag me on stuff.
1:27:07Um, send bugs, send feature requests. What's missing? What can we do to make the products better?
1:27:12What do you, like, what do you want? Um, I I love love hearing it. Amazing.
1:27:16Boris, thank you so much for being here. Cool. Thanks, Funny.
1:27:19Bye, everyone.
1:27:21Thank you so much for listening. If you found this valuable, you can subscribe to the show on Apple Podcasts, Spotify, or your favorite podcast app. Also, please consider giving us a rating or leaving a review, as that really helps other listeners find the podcast.
1:27:35You can find all past episodes, or learn more about the show, at lennyspodcast.com. See you in the next episode.
The Hook

The bait, then the rug-pull.

One year ago, Boris Cherny posted an internal demo of a terminal tool called Quad CLI and got two likes. Today that tool accounts for 4% of all public GitHub commits -- and the creator says the growth is still accelerating. In this conversation, he explains not just how it happened, but why the pattern of what AI takes on next is already visible in the data.

CTA Breakdown

How they asked for the click.

MENTIONED ON CAMERA
05:37productCursor ↗
05:45linkSemiAnalysis: Claude Code Is the Inflection Point ↗
16:00productCowork (Anthropic) ↗
1:05:00linkThe Bitter Lesson by Rich Sutton ↗
1:18:20productAcquired Podcast ↗
02:35productDX developer intelligence platform ↗
02:47productSentry + SEER AI debugging agent ↗
1:11:00productMetaView AI recruiting ↗
FROM THE DESCRIPTION
PRIMARY CTAWhere the creator wants you to go next.
OTHER LINKSAlso linked in the description.
Frame Gallery

Visual moments.

Frame at 00:02 from Head of Claude Code: What happens after coding is solved00:02Frame at 01:07 from Head of Claude Code: What happens after coding is solved01:07Frame at 02:59 from Head of Claude Code: What happens after coding is solved02:59Frame at 03:20 from Head of Claude Code: What happens after coding is solved03:20Frame at 04:56 from Head of Claude Code: What happens after coding is solved04:56Frame at 05:35 from Head of Claude Code: What happens after coding is solved05:35
Frame at 00:02 from Head of Claude Code: What happens after coding is solved00:02Frame at 01:07 from Head of Claude Code: What happens after coding is solved01:07Frame at 02:59 from Head of Claude Code: What happens after coding is solved02:59Frame at 03:20 from Head of Claude Code: What happens after coding is solved03:20Frame at 04:56 from Head of Claude Code: What happens after coding is solved04:56Frame at 05:35 from Head of Claude Code: What happens after coding is solved05:35Frame at 06:36 from Head of Claude Code: What happens after coding is solved06:36Frame at 08:10 from Head of Claude Code: What happens after coding is solved08:10Frame at 09:19 from Head of Claude Code: What happens after coding is solved09:19Frame at 10:23 from Head of Claude Code: What happens after coding is solved10:23Frame at 11:31 from Head of Claude Code: What happens after coding is solved11:31Frame at 12:15 from Head of Claude Code: What happens after coding is solved12:15Frame at 13:28 from Head of Claude Code: What happens after coding is solved13:28Frame at 15:01 from Head of Claude Code: What happens after coding is solved15:01Frame at 15:21 from Head of Claude Code: What happens after coding is solved15:21Frame at 16:38 from Head of Claude Code: What happens after coding is solved16:38Frame at 17:54 from Head of Claude Code: What happens after coding is solved17:54Frame at 19:13 from Head of Claude Code: What happens after coding is solved19:13Frame at 20:03 from Head of Claude Code: What happens after coding is solved20:03Frame at 21:03 from Head of Claude Code: What happens after coding is solved21:03Frame at 22:06 from Head of Claude Code: What happens after coding is solved22:06Frame at 24:03 from Head of Claude Code: What happens after coding is solved24:03Frame at 24:10 from Head of Claude Code: What happens after coding is solved24:10Frame at 25:21 from Head of Claude Code: What happens after coding is solved25:21Frame at 26:49 from Head of Claude Code: What happens after coding is solved26:49Frame at 27:37 from Head of Claude Code: What happens after coding is solved27:37Frame at 28:32 from Head of Claude Code: What happens after coding is solved28:32Frame at 30:34 from Head of Claude Code: What happens after coding is solved30:34Frame at 30:52 from Head of Claude Code: What happens after coding is solved30:52Frame at 32:00 from Head of Claude Code: What happens after coding is solved32:00Frame at 33:27 from Head of Claude Code: What happens after coding is solved33:27Frame at 34:52 from Head of Claude Code: What happens after coding is solved34:52Frame at 35:12 from Head of Claude Code: What happens after coding is solved35:12Frame at 36:11 from Head of Claude Code: What happens after coding is solved36:11Frame at 37:39 from Head of Claude Code: What happens after coding is solved37:39Frame at 38:30 from Head of Claude Code: What happens after coding is solved38:30Frame at 40:02 from Head of Claude Code: What happens after coding is solved40:02Frame at 40:42 from Head of Claude Code: What happens after coding is solved40:42Frame at 42:29 from Head of Claude Code: What happens after coding is solved42:29Frame at 42:58 from Head of Claude Code: What happens after coding is solved42:58Frame at 44:08 from Head of Claude Code: What happens after coding is solved44:08Frame at 45:17 from Head of Claude Code: What happens after coding is solved45:17Frame at 46:17 from Head of Claude Code: What happens after coding is solved46:17Frame at 47:29 from Head of Claude Code: What happens after coding is solved47:29Frame at 48:29 from Head of Claude Code: What happens after coding is solved48:29Frame at 49:54 from Head of Claude Code: What happens after coding is solved49:54Frame at 51:00 from Head of Claude Code: What happens after coding is solved51:00Frame at 51:41 from Head of Claude Code: What happens after coding is solved51:41Frame at 52:46 from Head of Claude Code: What happens after coding is solved52:46Frame at 54:05 from Head of Claude Code: What happens after coding is solved54:05Frame at 55:30 from Head of Claude Code: What happens after coding is solved55:30Frame at 56:33 from Head of Claude Code: What happens after coding is solved56:33Frame at 57:03 from Head of Claude Code: What happens after coding is solved57:03Frame at 58:17 from Head of Claude Code: What happens after coding is solved58:17Frame at 59:36 from Head of Claude Code: What happens after coding is solved59:36Frame at 60:29 from Head of Claude Code: What happens after coding is solved60:29Frame at 61:45 from Head of Claude Code: What happens after coding is solved61:45Frame at 63:21 from Head of Claude Code: What happens after coding is solved63:21Frame at 63:42 from Head of Claude Code: What happens after coding is solved63:42Frame at 65:15 from Head of Claude Code: What happens after coding is solved65:15Frame at 66:21 from Head of Claude Code: What happens after coding is solved66:21Frame at 67:15 from Head of Claude Code: What happens after coding is solved67:15Frame at 68:39 from Head of Claude Code: What happens after coding is solved68:39Frame at 70:01 from Head of Claude Code: What happens after coding is solved70:01Frame at 70:15 from Head of Claude Code: What happens after coding is solved70:15Frame at 71:29 from Head of Claude Code: What happens after coding is solved71:29Frame at 72:29 from Head of Claude Code: What happens after coding is solved72:29Frame at 73:35 from Head of Claude Code: What happens after coding is solved73:35Frame at 74:48 from Head of Claude Code: What happens after coding is solved74:48Frame at 75:41 from Head of Claude Code: What happens after coding is solved75:41Frame at 76:51 from Head of Claude Code: What happens after coding is solved76:51Frame at 78:15 from Head of Claude Code: What happens after coding is solved78:15Frame at 79:06 from Head of Claude Code: What happens after coding is solved79:06Frame at 80:12 from Head of Claude Code: What happens after coding is solved80:12Frame at 81:19 from Head of Claude Code: What happens after coding is solved81:19Frame at 83:09 from Head of Claude Code: What happens after coding is solved83:09Frame at 84:14 from Head of Claude Code: What happens after coding is solved84:14Frame at 84:46 from Head of Claude Code: What happens after coding is solved84:46Frame at 85:43 from Head of Claude Code: What happens after coding is solved85:43Frame at 86:45 from Head of Claude Code: What happens after coding is solved86:45
Watch next

More from this channel + related breakdowns.

44:52
Matthew Berman · Essay

It's starting...

A 45-minute walk through Anthropic's internal data showing AI crossed from coding assistant to primary engineer — and a frank read on what that means for humans.

June 5th
Chat about this