How I Use Claude Code to Stay Locked In (While I Work)
Seven attention problems from social media and AI agents, and the plain fixes one builder found over 60 days: an agent-run notification filter, custom feeds, paper, walks, slop filters, and a daily journal.
Posted
yesterday
Duration
Format
Essay
sincere
Views
2K
77 likes
57 · 43
Big Idea
The argument in one line.
AI agents can rebuild the filters that apps and algorithms refuse to give you, and the rest of your focus comes from three analog habits: reading on paper, walking, and writing by hand.
Who This Is For
Read if. Skip if.
READ IF YOU ARE…
You run Claude Code, Codex or a similar agent every day and feel your attention getting shredded by notifications, feeds and half-finished agent tasks.
You want specific, copyable setups, like an agent that only pings you for things due within eight hours, rather than general advice about digital minimalism.
You consume a lot of short-form video and podcasts and want a way to filter out slop and find the few high-quality sources worth your time.
SKIP IF…
You want a rigorous, sourced study of attention and screen use. The statistics are quoted quickly and the fixes come from one person's 60-day experiment.
You want a tutorial on setting up the tools. The video shows ideas and prompts, not step-by-step configuration.
TL;DR
The full version, fast.
Riley Brown argues that notifications, algorithmic feeds, short-form content, agent multitasking, AI slop, meaninglessness and outsourced thinking are all pulling at your attention, and that the same AI tools can be pointed back at the problem. He has agents act as an intelligent do-not-disturb filter, builds custom feeds and slop filters with scraping APIs, and uses a coordinator agent that reminds him of unfinished work. Around that he leans on three analog habits: printing long reports to read on paper, walking while talking to agents, and handwriting in a journal every day.
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.
Names the attention problem and promises seven issues learned over 60 days.
00:43 – 05:16
02 · Notifications and the importance line
Why apps notify you, and how to make an agent your intelligent do-not-disturb filter.
05:16 – 09:37
03 · Algorithms and custom feeds
Karpathy's HackerTube and a custom Grok bot show how to replace scrolling with your own feed.
09:37 – 12:35
04 · Short-form content and printing
Reading is declining, so he prints Claude reports and transcripts to read away from his phone.
12:35 – 16:16
05 · Agent multitasking
A coordinator agent with nightly reminders, and walking while talking to agents.
16:16 – 19:05
06 · AI slop
Using Exa Podcasts and Scrape Creators to have agents filter out low-quality content.
19:05 – 22:52
07 · Meaninglessness and thinking for yourself
Focus on a hard project, and handwrite daily so you keep thinking without AI.
22:52 – 24:13
08 · Recap and sign-off
One-paragraph summary of the seven fixes and a request for feedback on journal-style videos.
Atomic Insights
Lines worth screenshotting.
Apps send notifications to serve their own business model, not your well-being, so the fix is to let your own agent decide which pings are allowed through.
A good agent filter rule is concrete: only message me about things I must act on within eight hours, and batch the rest for the end of the day.
Apple's built-in intelligent notification silencing is not trusted here. The better filter is an agent you instruct yourself across Slack, email and your other apps.
Andrej Karpathy fixed his YouTube recommendations by having Claude filter 43 million Hacker News comments into 6,454 videos, in about two hours for roughly $5 to $15 of credits.
When software costs a few dollars and a few hours to build, you can optimize an app for your own learning instead of accepting a platform's engagement goals.
A custom Grok bot that posts the top 10 AI and tech posts each morning replaces scrolling an addictive feed with one ranked list.
The share of US teenagers who read daily for pleasure fell from about 30% to nearly 10%, and almost half hardly ever read.
Printing a 12-page Claude Code report costs about 50 cents, and a 200-page book costs about $6 with bulk ink.
Printing analytics reports and creator transcripts lets you read and circle them with no phone nearby.
A laminator costs about $30 and turns a one-page Claude-generated graphic of a concept into something you can keep in view while you learn it.
Agent multitasking fails because you lose track of the main objective, so a coordinator agent that reminds you nightly of unfinished tasks keeps threads from rotting.
Walking with AirPods and talking to an agent forces you to run fewer agents at once, because you can't pick up your phone and scroll.
Roughly 70% of this creator's daily emails now come from speaking into AirPods on walks.
Exa Podcasts indexes about 3.2 million episodes from 119,000 shows with full transcripts, so an agent can screen episodes by depth before you listen.
You can teach an agent what slop means to you, such as filler episodes or sponsored content, and have it return only links you are likely to value.
Meaninglessness in the AI era eases when you stop identifying with a task or a job and attach to a hard project you want to finish.
Defaulting to AI for every hard problem erodes your own thinking, so a daily handwritten journal away from technology is the counterweight.
Takeaway
Point your agents back at your attention
WHAT TO LEARN
The tools that fragment your attention can also filter it, as long as you write the rules yourself and keep a few analog habits that agents can't replace.
02Notifications and the importance line
Apps notify you to serve their business model, so give an agent your own rule, such as only ping me for things due within eight hours, and let it watch every app.
Plain do-not-disturb blocks everything, including the urgent. An agent connected to your apps can let through only what clears your importance line and batch the rest for evening.
03Algorithms and custom feeds
When software costs a few dollars and a couple of hours to build, you can replace an algorithmic feed with one tuned for your own learning, as Karpathy did for YouTube.
Scraping a platform with an agent turns an addictive feed into a short ranked digest, such as a daily top 10 of AI and tech posts.
04Short-form content and printing
Long reading degrades on phones, so print what matters. Claude Code reports cost about 50 cents each and let you read and annotate with no notifications in reach.
Turn concepts you are learning into one-page graphics, laminate them, and keep the key ideas physically in front of you.
05Agent multitasking
Running many agents at once makes you lose the main objective, so use a coordinator agent that reminds you each night of what you started and haven't finished.
Walking while talking to an agent limits how many tasks you run at once, because you can't reach for your phone, and it keeps you in a calmer flow state.
06AI slop
Agents can screen podcasts and creators for depth using transcripts, so you spend listening time only on sources that pass your own quality bar.
Define slop for your agent in your own terms, such as filler, sponsored or surface-level content, and it will return only links you are likely to value.
07Meaninglessness and thinking for yourself
Meaninglessness eases when you stop defining yourself by a task and commit to a hard project you care about finishing, with AI as the means.
Using AI for every hard problem erodes your own thinking, so handwrite in a journal daily, away from technology, to keep your judgment your own.
Glossary
Terms worth knowing.
Importance line
The creator's name for the rule you give an agent about what is important enough to interrupt you, instead of letting each app decide for itself.
Intelligent DND
A do-not-disturb setup where an AI agent monitors your apps and decides which notifications reach you, based on instructions you wrote.
Coordinator agent
A single main agent in a Claude project that organizes your work into side threads and reminds you what is unfinished.
Agent multitasking
Running several long-working AI agents at once, which is productive but makes it easy to lose track of the main objective.
AI slop
Low-value, mass-produced AI-generated content that buries the good material on feeds, podcast directories and video platforms.
Exa Podcasts
A feature of the Exa search API that covers about 3.2 million podcast episodes with full transcripts, usable through normal search.
Scrape Creators
An API that lets an agent pull posts and transcripts from Instagram, Twitter, YouTube and similar platforms so it can filter them for you.
Resources
Things they pointed at.
03:25toolClaude Code
05:20productHackerTube (Andrej Karpathy)
08:07toolGrok bot 'Viral Posts'
17:30toolExa Podcasts (Exa API)
18:35toolScrape Creators
Quotables
Lines you could clip.
00:12
“Between AI agents and social media algorithms, it feels nearly impossible to get into a flow state and actually focus on important tasks.”
states the problem in one sentence, no setup needed→ TikTok hook↗ Tweet quote
04:30
“You get to decide what you see rather than relying on these companies that get to freely send you notifications. I call this the importance line.”
names a reusable idea→ newsletter pull-quote↗ Tweet quote
17:46
“You can train AI to teach it what slop means to you.”
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.
17px
We live in a world filled with algorithms and endless content that want to steal our most valuable resource, our attention. Between AI agents and social media algorithms, it feels nearly impossible to get into a flow state and actually focus on important tasks. This is something I've been personally struggling with a lot for the past year, and I know many people who would agree.
And so for the past two months, I've decided to take this problem very seriously. And every day I've asked myself, what is causing this? causing the lack of attention and how can we use the most powerful AI tools to actually steal our attention back.
And over the past 60 days, I've learned that there are seven main issues we need to consider. And so our attention is under attack. And I feel like it's under attack by these seven main problems.
I think these problems fall into one of two categories, social media problems and AI problems. And so problem number one that we are all dealing with are notifications pull us in a hundred times per day. And before we get to my solution that involves using AI agents, I think we need to break down the problem a little bit more.
And so let's think of app notifications for a second. I want you to think what are the reasons why an app sends you a notification? Is it your well -being?
Is it increasing your happiness? Is it increasing your rate of learning? Is it making sure you feel more secure?
Is it so that you are more financially successful? Is it to improve your love life? The answer to this question is probably not.
The reality is it looks a lot more like this red triangle right here. What's really going on is there is a trade -off. There's a trade -off between what's good for you, what's good for the rest of the users on the platform, and what's good for the business model of the platform, even if user hostile.
And so when these companies are deciding whether they want to send you a notification, they're not thinking about whether it's good for you, or at least that's not all they think about. They're thinking mostly about is it good for their business model? Will it make them more money?
And is it good for other users? And that's not just current users on the platform, right? This is also future users on the platform.
They are trying to grow their app. as quickly as possible and maximize revenue, which means that when they make these decisions and when they're deciding whether to send you a notification or not, they aren't thinking about your well -being. OK, so what if we actually changed this and we were only notified when something is good for our well -being or if we had to be notified about that thing?
And so, of course, you could always just put your phone on do not disturb. But then you get literally no notifications. So that means if something really important happens, you completely miss it.
However, on iPhone, they have these exceptions where they allow notifications from certain people and certain apps. And so I've said it. So only the most important people to me and my business.
can message me. You can also turn on this intelligent breakthrough and silencing where Apple will decide what's important for you. It sucks.
Don't use it. But the most important part of this is the app section. Since GPT dots.
Claude Duet, their new agent platform, and Grokbot connect to all of your apps anyway. And you can actually prompt these agents to notify you or not notify you about certain things. We can actually use these tools as a filter between all of the potential notifications you could get from your people, your apps, or your email.
And it can be filtered exactly how you want it. And so if you want to be productive for work, you should take the agent platform that you use. Let's say...
for this example, you use GPT dots, what you would do is you would make ChatGPT one of these apps that can send you notifications. And then you should go to the plugins tab in the ChatGPT app and make sure that you are connected to all of the different apps that you use. And then now you can go to your dot and you can be like, hey.
I want you to please only notify me, like message me if it's something that I need to do within eight hours. If it's not that important, I want you to save that for the end of the day to message me. But you should constantly be monitoring my Slack, my email, and every single thing that I use in business to make sure that you don't miss anything.
And you should constantly monitor these every 30 minutes. And so you can just have your AI agents notify you based on your own instructions. So you get to decide what you see rather than relying on these companies that get to freely send you notifications.
And so I call this the importance line. You get to decide what is important and what is not important. And so this is how you create your own notification algorithm is you monitor everything that's going on in your digital life and you tell it under what circumstances these agents should send you a notification or ping you.
And so the answer to the notifications problem is intelligent DND. And I've been using this for the past two months straight and it's improved my productivity a ton. So problem number one is notifications.
And problem number two is that algorithms keep us scrolling longer. And so here's a perfect example of what I'm talking about here. So Andrej Karpathy posted this today.
So in two hours with Opus 5 .5, he built something called HackerTube. Listen to this. I was annoyed with YouTube's recommendations.
So I had Claude filter down 43 million comments over 20 years of hacker news. to 6 ,454 videos that kept being recommended. And it's a significantly better viewing experience.
And he built this in two hours with Claude Opus. In his blog, he said this. It was Friday night at 11 p .m.
I was tired. I went to YouTube to watch something and realized that I was looking at an enshitified wall of slop. insanely exaggerated thumbnails, dumb recommendations of the same tired topics, truly terrible idiocracy grade content.
I'm sure this algorithm works well for the median person on the internet, but it definitely does not work for me. So this morning, Saturday, I took a hold of my Opus 5 .5 and gave it a task. And he basically created this entire new site that had his own personal algorithm built into it.
It recommended him videos that he actually wanted that would make him better. And so now we can go back to the triangle because Opus 5 .5 is so good at building basically any software. If you just go to Claude Code and you were to create a new Claude Code session, you can just ask it to build basically any software that you want and it will just build it on command.
So this probably cost him like $5 to $15 worth of credits, and he created it in a couple of hours. And so he was able to completely optimize for his own experience. And so remember, YouTube is not incentivized to make something that is good for Andrej Karpathy.
YouTube is the most profitable social media platform in the world, arguably. It's always looking after its own business model. And so in the age of AI, you can now just build software that optimizes for your own well -being, your own rate of learning.
That's basically what he was doing. Your own security, your own financial success, your own love life. You can build software that optimizes for these things for just a couple of dollars in a few hours.
And so instead of scrolling the algorithms or looking at the For You page on YouTube, which... you know, is a lot of terrible content, Andrej Karpathy was able to scrape from social media to create his own feed, which is really cool. But YouTube is not the only platform that you can do this for.
So I think the most important platform to learn AI is actually Twitter. That's where all the people who use AI hang out and the smartest people in the world hang out on Twitter. It's just a fact.
And so one thing that I found to be really useful is I still like to see what went viral for that day. It gives me a sense of kind of culturally what's going on. And so I created my own custom Grok bot called Viral Posts that every single day it said, morning, Riley, here are the top 10 AI and tech posts from the past 24 hours ranked.
It was quieter Sunday. So a few picks are discussion posts rather than big launches. And so now instead of me having to scroll through these addictive feeds, I'm able to.
get Grokbot to just scrape directly from Twitter. And so I did it to scrape Twitter posts. Andre Karpathy did it for YouTube.
And so the third thing that is coming after our attention that is a big problem that we need to solve for is that short content is rewiring our brain. If you look at the data, the amount of people and teenagers who are reading in their leisure time is going down drastically.
30 % of teenagers in the US were reading every day. It is now approaching 10%. Now the amount of US teenagers that hardly ever read is almost 50%.
And I believe that it's just that people can't concentrate anymore because they are reading on a phone that has really short content and it's completely rewiring our brains. And so the solution that I have found to this problem is that I am done reading on my computer.
If there's something that I need to read that is more than one page, I will simply print it out. So I got this printer about three and a half weeks ago and I'm telling you... It has completely changed the way that I consume reports.
I'm constantly printing things out and I will go to a place where I don't have my phone and I will read it. For example, this is a report that Claude Code created and it actually did research for about two hours before generating this report. It's 12 pages long and it went through all of my YouTube data.
And so on Cloud Code, you can actually connect it to your YouTube channel and you can connect it to all of the analytics. And so it actually went through every single video and had all the retention charts for every single video. And it created this chart and.
Look at this. It has the retention charts for every single video that I created. And I did the math here.
If you buy a nice printer, a document like this costs like 50 cents to create. And so I bought my ink in bulk. If I were to print out a 200 page book.
It's $6. It's $6 to create this. And so you can very easily have Claude create these reports.
And instead of getting distracted reading them on your phone or your computer, you can sit down with your coffee with no distractions and you can just read the report. Here's one that I had analyze all, I think it's 10 of my top short form creators. And to conserve paper, I keep them really small and I'll just like circle the videos that I like.
And so these are the short form scripts. This one is, I believe this is Kane Calloway. And I just analyze short form content.
Instead of watching them on Instagram, I'm literally printing out the transcripts. from people's content. And it has been incredibly fun to do.
And I can do it without any distractions whatsoever. And then whenever I'm trying to learn something new, I'll actually print it out and laminate it. A laminator is like $30.
You can buy it. And I have Cloud Code generate these graphics. You just tell it.
I want to create a graphic that fits on a single page. And it's really fun. And now I have like 10 of these with the key concepts that I'm actually trying to learn.
And I might start even using these in my videos. and so highly recommend getting a printer because it just freezes the pixels in time and it lets you relax as you read them and i just love holding paper i love paper i know it seems wasteful But like looking at a screen all day, I just I'm convinced that humans were not meant to look at screens.
Humans were meant to read physical paper. I did a lot of research on this one. Again, this video is not sponsored.
This one just apparently has very little problems. And I ordered ink in bulk. So I'm not going to run out of ink for a very long time.
And then you just put paper in here and. You can literally just ask Cloud Code, when it's done, can you please print it off? And Cloud Code can just print super easy.
And so those problems stem from social media. Now let's move to the problems that stem from AI models and AI agents. And so the fourth problem that is directly attacking our attention is agent multitasking.
So obviously, agent multitasking is incredibly useful. And all I mean by this is when I use Cloud Code or when I use Codex, I fire off an agent and agents can work up to like multiple hours at a time effectively, whether it's editing videos, whether it's updating my app or doing any other useful thing for my business, agents can work for a very long time.
However, when you're using these tools and you're constantly spinning up task one, task two, task three, it's easy to get lost and lose track of like your main objective. And so the first way that I actually counteract this agent multitasking is I've been using cloud projects.
It allows you to instead of just using normal cloud code where you just see all of your old sessions and you just create new sessions for everything, instead you have this main agent. And I just call this Agent Duet because I actually like it more than Claude Duet. But this is just a Claude project.
And I basically do all my developer work in here. So whether I'm editing a video or making an update to my website, I just tell my Claude project coordinator chat and then it will actually organize everything in these side threads. And so you can see these are all of the different Claude code threads that my coordinator agent has spun up.
And so basically what I do is I have this prompt that I sent my. agent, which is I want you to remind me every night of my unfinished tasks in this project. What this will do is this will your coordinator every night will message you and be like, hey, you started this thing.
You should probably finish it. And so that's one way to counteract this agent multitasking problem is that if you start something, you tell your agent to remind you what you need to continue working on. And so that is solution number one is a reminder prompt.
But there's an underrated second solution. which is go for walk. I don't think there is something that has been more impactful to me than using agents while walking.
And right now. I can put in my AirPods and I can just talk to my dot, which is the GPT agent platform. And I'm able to just talk to it in natural language and we can actually get things done.
I can have it do research. I can have it build apps or make updates to my existing apps. I can have it give updates on all of like on Slack notifications and it can send Slack notification.
It can do the same with emails. And so whenever I start to lose focus at my computer, I'll just put my AirPods in and go for a walk. And what this will force me to do is this will force me to run less agents and I get less antsy.
Right. My first instinct when I'm sitting at my computer and I'm multitasking and let's say I have two agents running and they're both working. I have this instinct, which is pick up phone.
But when I'm walking, I don't get that same urge. I'll tell it to do a task. And while it's working, I'll just walk.
And so it's more like a flow state. So that's why I think it's very useful to live in a walkable city and to make sure that you have really good 5G connection is because you can use agents from your phone just by speaking. And my phone's in my pocket and I'm able to get so many things done.
I think around 70 % of my emails that I send on a daily basis come from me speaking into my AirPods while walking. And I cannot think of anything that has been more fun and more rewarding and really good for my attention. OK, now let's move to the fifth problem that is really affecting our attention, which is AI slop.
So unfortunately, on the Internet, the ratio of slop to not slop is getting kind of scary. There is so much AI generated slop on the Internet and really low value content. OK, so yesterday I was looking for a podcast and I couldn't find a podcast that was actually.
good or high quality. And of someone that has a lot of expertise or is an expert in a field, all I could find was a bunch of slot podcasts that I didn't want to listen to. And so I thought to myself, I'm like, what if there was a way that I could get an AI to go analyze the transcripts and it could actually analyze it and look to see that it would be above 110 IQ and not general surface level jargon.
And so I basically found out that EXA The EXA API can now scrape podcast transcripts. And so on October 29th, EXA launched EXA Podcasts.
And so about 3 .2 million episodes from 119 ,000 shows, all with full transcripts and updated as new episodes air. It works through normal search at no extra cost. And so all I needed was an EXA API key.
And again, this is not sponsored by EXA. They announced this a few days ago. Super useful.
And now you can just go to Cloud Code and you could just say something like this. You could say, hey, I want you to use the EXA API and I want you to find the best episode on infrastructure and learning about the entire AI stack. And this should be a really high quality episode.
No jargon, no fluff. Someone who has deep expertise that I would get a ton of value from. I want links to 20 episodes.
And then I want you to give me a full summary of these episodes that kind of bring everything together and create that document. And so I can fire that prompt off and it will just go look through all the episodes and ignore the slop and find the best episodes. So I don't have to sit there and realize it slop myself.
You can train AI to teach it what slop means to you. So maybe it's just like a filler episode or super boring and lame. You can tell AI to filter through that and give you only what you want.
And if you've been watching my videos, you know that this is not just true in podcasts. You can also use something called scrape. creators and give your AI agent a scrape creators API key and it can do the same on Instagram, the same on Twitter and the same on YouTube.
It can just scrape the entire transcript of videos and it can filter out slop. And now Instagram and TikTok, I would argue, has way more slop than some of the longer form platforms. But this is still really useful if you want to learn from a certain creator you find and you can filter out the slop, you can filter out the sponsored content and it's really fun and useful to do.
And so you can use Exa to scrape podcasts and you can use scrape creators to scrape content and filter out the slop and just give you links directly to videos that you are very likely to enjoy. And now finally, problem number six and seven, in my opinion, are deeply related. And problem six is meaninglessness.
And problem number seven is not thinking for ourselves. I think there's a deep sense of meaninglessness in our culture today. And many people are losing the sense of meaning because of AI.
Like AI is taking our jobs. AI can do what I've been training to do for years. And there's also this sense that AI is evil.
A lot of people... And I think that the way you counteract the meaninglessness is you need to separate yourself from the task and try not to think about the task. And rather, you should try and think about the project.
You need to challenge yourself and you need to find a project that is really hard and challenging and take it on. And whether you are trying to create a really cool video game that you and all of your friends really like to play, or if you want to start a business and use AI to get there, you should focus on the project.
Because if you stop thinking about yourself as a person with a job and you think of yourself with, oh my God, I can use this technology to like accomplish this grand project. That if I were to do this project, I could make money doing it. That has a lot of meaning.
And you kind of lose the meaninglessness. You're able to shed the meaninglessness. And I think it's really important.
And then you'll be incentivized to like use AI as much. as possible. And I say this with some caution because I've realized this with me.
I will default use AI with everything. Whenever I come across a hard problem, I'll run it through AI. I'll literally turn on my phone and I'll just speak for 10 minutes straight to give all of the context to AI.
And I'll be like, hey, can you like lay this all out for me and lay out my different options? And I think this is really good. Especially if you get in the habit of like taking that document and talking it over with other people.
However, if I do this too much, I'll stop thinking for myself. And so what I've been doing to counteract this part, and this also helps with the meaninglessness, is I write every day. For the past 60 days, I've made a concerted effort to sit down and handwrite every single day in a journal.
And what this allows me to do is this allows me to think and I'll do this nowhere near technology, nowhere near AI. And some of my best video ideas and some of my best business ideas come from sitting in solitude and writing. And I'll often think about the projects that I'm working on and I'll think about how I can use AI and this separation.
where I can just sit and think with myself gives me so much clarity. And so that is why I just genuinely love writing. I've been loving printing things out on physical paper.
And then, of course, I've been loving going for walks and using agents. And so writing, reading on paper and walking. These are three incredibly basic things that in this like rapidly accelerating era have helped me so much because all three of these things really ground me.
And especially the writing part has given my life a lot more meaning. And it allows me to think in this age where everything is moving so fast and everything is battling for our attention. And so there you have it.
In summary, you can basically create intelligent do not disturb filter, which makes sure your notifications are incredibly important. You can scrape from social media to create your own feed to avoid the algorithms. Short form content is completely rewiring your brain.
So it's really great to print things out when you want to sit with it for a long time. And then that will actually inform your writing. And then for agent multitasking, make sure you have a clear objective and go for a walk if you need to.
And then use AI to also remind you of the things that you need to finish because working on multiple things at once can get really daunting. You can also scrape from any platform and this helps you filter through AI slot. And then, of course, I highly recommend picking a really meaningful project and then writing every day.
And this can combat meaninglessness and the increasing trend or urge to not think. for yourself to rely on AI to do thinking for you. It's always good to pull back and try and find meaning.
And I do that through projects and writing. If you found this interesting, please make sure to follow and subscribe and like. It helps me out so much.
You don't realize how much it helps me out. This is kind of a different style video. Let me know if you like these kind of like journal style videos on things that have really helped me in the age of AI because I could dig deeper into any of these topics.
Let me know. Thank you guys so much for watching, I'll see you here in the next one.
The Hook
The bait, then the rug-pull.
Riley Brown says he spent the last 60 days treating his lack of focus as a serious problem to solve. He lays out seven attacks on attention and pairs each with something he actually does, from an agent that screens his notifications to a printer and a paper journal.
Frameworks
Named ideas worth stealing.
00:43list
The seven attention problems
Notifications
Algorithms
Short-form content rewiring the brain
Agent multitasking
AI slop
Meaninglessness
Not thinking for yourself
Grouped into two categories: social media problems (the first three) and AI problems (the last four).
Steal forauditing your own attention habits
04:30concept
The importance line
You write the rule for what deserves to interrupt you and let an agent watch Slack, email and other apps against it.
Steal fornotification and inbox triage
22:29list
Writing, reading on paper, walking
Write by hand daily
Read on paper
Walk while using agents
Three basic analog habits that ground him while everything around him accelerates.
Steal fora daily focus routine
CTA Breakdown
How they asked for the click.
VERBAL ASK
23:52subscribe
“If you found this interesting, please make sure to follow and subscribe and like. Let me know if you like these kind of journal style videos.”
Soft ask paired with a question about the format, which invites comments.
Add Modern Creator as a preferred source and Google shows you more of our breakdowns in Search, Top Stories, and AI Overviews. It only changes what you see, and you can undo it in your Google settings anytime.
Add to Preferred SourcesOpens your Google source preferences with us pre-loaded. Tick the box and you're done.
A creator with 2 million followers maps his whole agent stack onto one grid: Claude for making things, Codex for running the business, Composio as the shared toolbelt, and GrokBot as the Twitter specialist.
Riley Brown spends 24 minutes inside the new Claude Projects: one orchestrator chat spinning up parallel threads that share memory, artifacts, and routines, plus token costs, the new in-app design mode, and the phone workflow.
An interview with an early power-user on why the open-source Slack-for-agents app is going viral, then a full walkthrough of running a personal AI agent team inside it.
Claude Desktop and the Codex app both quietly shipped native, multi-tab in-app browsers within days of each other — used right, one prompt now opens, drafts, and stages a dozen tabs at once.
A 33-minute field guide to the nine structural shifts in AI agents — what's permanent, what's changing, and the five human foundations that will matter most.
00:23
00:41
01:03
01:39
01:58
02:16
02:34
02:54
03:10
03:21
03:48
04:00
04:23
04:41
04:59
05:17
05:36
06:12
06:37
06:48
07:06
07:24
07:43
08:01
08:19
08:37
08:56
09:13
09:31
10:09
10:26
10:44
11:03
11:21
11:39
11:57
12:15
12:42
12:51
13:10
13:46
14:04
14:22
14:40
14:59
15:17
15:35
15:53
16:11
16:29
16:48
17:06
17:31
17:43
18:18
18:36
18:55
19:04
19:31
19:49
20:25
20:44
21:02
21:20
21:38
21:56
22:17
22:33
22:51
23:27
23:50
31:28
16:14
24:06
Video of the Day55:15
Video of the Day23:10
33:38