What the smart people are recommending. 7861 books, tools, and products endorsed by the thinkers absorb.md tracks. Ranked by how many times each has been recommended across compiled podcasts, papers, posts, and tweets.
βRead, extract tables, fill forms, merge, and split. The highest utility skill for anyone who works with documents, period.β
βExcel formulas, data analysis, charts, all from plain English. You don't even need Microsoft Excel installed anymore. Just describe what you want, and Claude builds the spreadsheet. For anyone runningβ¦β
βSlide decks from natural language. Layouts, charts, speaker notes. For anyone who's ever spent 4 hours making a pitch deck for a meeting that lasted 12 minutes, this will make you question your life dβ¦β
βIf you're in the content game like me, this is real collaborative AI writing.β
βSocial graphics, posters, covers. You put text in, PNG or PDF out. If you're posting content regularly and you're tired of opening Canva every 5 minutes, this handles it inside of Claude.β
βThis is for my vibe coders. I'm talking calculators, dashboards, interactive widgets, all from natural language. Zero front-end skills needed. You describe what you want, it builds a working artifact.β
βMarketing skills by Corey Haines. 16,000 GitHub stars, over 20 sub skills. CRO, copywriting, SEO, email sequences, growth engineering. This is the most popular non-coding skills package for Claude codβ¦β
βIt encodes your brand's voice, colors, tone, and rules into a skill. Then Claude auto applies them to everything it creates, every email, every landing page, every social post, all on brand without yoβ¦β
βIβm building a project internally called Ultron and Ultron inside of my firm launch that is going to basically with the Slack API weβre pulling every single message from Slack into our OpenClaw.β
βFor anyone making content or building products where the research needs to be solid, this saves you literal hours.β
βIf you use Obsidian for notes like Coda, and you don't, you probably should. This bridges the gap between your knowledge base and your AI agent.β
βYour wallet will thank you even if your agent won't.β
βI wouldn't dramatically uh change some of the basic advice on STEM uh getting good at even for things like coding I would still recommend because I think whatever happens with these AI tools you'll beβ¦β
βI would also say immerse yourself now. That's what I would be doing as a teenager today in trying to become a sort of ninja at using the the latest tools.β
βSo, instead of manually writing skill files, which, let's be honest, you don't even know how to do anyway, you just tell Claude what you want to do differently, and this skill creates the skill for yoβ¦β
βpeople that know the language of computers, be able to understand coding will be able to do that much more effectively than people that don't... I think we look back on that as some of the worst careeβ¦β
βI use AI as a brainstorming companion much more than even my friends know... of a brainstorming I use multiple models.β
βActually go and build stuff. I think every one of you this is a wonderful time to build... There's so much cool stuff you can now build that just was not possible before.β
βyou can have maybe lm.txt txt file which is just a simple markdown that's telling LLMs what this domain is about and this is very readable to a to an LLM.β
βuse the LLM as a judge and well and just go through that process to get it done much faster.β
βI think everyone should learn to code... everyone in AI fund actually knows how to code... it's actually driving meaningful productivity improvements across all of these job functions.β
βdevelopers that use AI assistance in our coding is so much faster than developers that don't... I just hate to ever have to code again without AI assistance.β
βI mention one thing that people have talked a lot about, but I think is so underrated is the voice stack... I'm actually very excited about voice applications.β
βThen these AI agents tools like GitHub copilot cursor or anthropics claude code they ingest that plain English prompt and just generate the functional underlying code for you.β
βI think maybe the main thing is I would say read history. Read history and know these things and try to get that balance right.β
βinvesting in open source and open way models. Building these models is so expensive. What most nations need is to make sure that no one else can control your infra. But when there's open-source openweβ¦β
βRead the recommended schema.β
βThe repo link is on screen and it is worth reading the readme, the skill pack, the schema and the verify runbook together.β
βThe repo link is on screen and it is worth reading the readme, the skill pack, the schema and the verify runbook together.β
βThe other really important thing they did, which I kind of give I give as a recommendation for parents to think about, which was very useful for me becoming a scientist, is you know, when you come homβ¦β
βif you just use the tools, if you actually go spend 10 hours using these tools as you recommend and just ask, 'Hey, be nice to me.'β
βI am a big fan of CUDA graphs where the entire thing that's running there is one CUDA call to the graph and it does everything.β
βI want every researcher to add a delay Q into your uh right before your environment. Just put a little flexible delay Q where you can tune in up and down how many frames you delay.β
βSprinting is one of the most underappreciated longevity therapies.β
βThis is exactly the problem we talk about in building a second brain community.β
βthis is why I've been like pushing for social recovery, multiigs, account abstraction, all of this stuff non-stop for the last 10 years.β
βif you're if you're going to spend if you're early in your career that's I don't know there's some things I think ZK biotech things like that haven't had their massive breakout moment yet and ZK was dβ¦β
βYou could watch a YouTube video on it. There's some YouTube videos that explain it pretty clearly. Takes a bit of time to understand it.β
βfor people to fully appreciate the power of quantum computers and to accelerate their use in different fields, people should be learning the fundamentals of quantum, even at a young age.β
βToday, I'm excited to share a second brain experiment where we will explore whether it's possible to create a fully functional app using only my second brain and claude cowork.β
βA pattern I've seen that works pretty well that I actually recommend if you're a large company is to kind of make a strike team or strike force that's separate from the rest of the companyβ
βI wish I'd stuck with physics more because then I would have had that same feeling about advanced physics and math.β
βI wish I'd stuck with physics more because then I would have had that same feeling about advanced physics and math.β
βSo computer networking, computer architecture, semiconductors, all that stuff was really useful even though I was doing software programming when I got out because just understanding the layers of absβ
βSo computer networking, computer architecture, semiconductors, all that stuff was really useful even though I was doing software programming when I got out because just understanding the layers of absβ
βSo computer networking, computer architecture, semiconductors, all that stuff was really useful even though I was doing software programming when I got out because just understanding the layers of absβ
βIn fact I would say you know go even lower level study chips, study physics, study whatever you can to understand how computers work because that's the greatest form of leverage in modern societβ
βWhat I do instead is something called [music] an annual life review. And honestly, I think it's the piece that most people are missing. If you've ever felt like a failure because [music] you can't stiβ¦β
βThe first thing I would say as a researcher is you have to really look for the truth. And that's the number one thing you you can do.β
βFirst of all, I think it goes back to what I talked about earlier. You should be asking questions, and making sure that you don't feel intimidated by asking questions all the time. And you know, make β¦β
βI remember I learned in in graduate school, I couldn't make these Josephson junctions. They didn't work at all. After months of trying different things, I finally went into the clean room and spent thβ¦β
βHaving other hobbies and focusing on that is quite quite good, and I think I tend to be very intense about doing the physics. So, you know, having having some other outlets, and I would say especiallyβ¦β
βThis project was 99% vibe coded as a fun Saturday hack because I wanted to explore and evaluate a number of LLMs side by side in the process of [reading books together with LLMs](https://x.com/karpathβ¦β
βBasically the idea here is that if you want someone to like the same analogy works for like a human teammate. If you want someone to go off and do a lot of work, you might want to agree on what work tβ¦β
βI'd encourage people here this week to like try at least once paying for a coffee and ETHβ
βIf you don't have a wallet, you know, install a install a wallet.β
βlike if you haven't used a DAP, use a DAP.β
βas well as other popular enterprise AI approaches like retrieval augmented generation, rag.β
βIf uh, you know, like you have never written a smart contract, try like writing a uh, a smart contract.β
βOh, I see one hand uh go up. I'll try to find you afterwards. Uh so lisp is a um well many people say that lisp is a dead programming language, but the lisp defenders say that it's not dead. It just sβ¦β
βForget New Year's Resolutions and Conduct a 'Past Year Review' Insteadβ
βAnd then what I want also to mention is it's not by chance that we um took this topic as the end of the AI festival of the quantum year because especially the toin the faculty of physics faculty of elβ¦β
βWell,, I, think, um, what, Anthropic's, doing with code is very interesting with their claw code. There's a lot of excitement around that in the developer market. We're pleased with the the performancβ¦β
βWe therefore demonstrate that future investigations of the dark matter properties of galaxies using rotation curves need to account for varying M/L across individual galaxy profiles and between galaxiβ¦β
βyou'll leave this course with the theoretical foundations and the hands-on experience you need to get the most out of retrieval augmented generation.β
βGarry's plan-exit-review skillβ
βSee the link in our show notes to try my blind spot finder recipe and experience how granola makes your meeting notes awesome.β
βSo, here is the red skill that Jeff put together. Jeff, for folks who don't know, Bird, explain to them what it does and why we need this.β
βSo, Bird uh for OpenClaw just basically lets you search and reply to people uh on on Twitter or Xβ
βAnd we could develop this technique called MUP for how we set the initialization. And that actually gives you a much more straight line as you go.β
β- Build persistent memory stores for different agent memory typesβ
β- Implement a Memory Manager that orchestrates how your agent reads, writes, and retrieves memoryβ
βTreat tools as procedural memory and retrieve only relevant ones at inference time using semantic searchβ
βYou just gave Claude Code & Cursor production deployment superpowers Running πππ‘ πππππππ πππ gets you every Skill and keeps them updated.β
βWe call this Generative UI (eg: https://t.co/j52qMrIOts).β
βExperimenting with Starlette 1.0 with Claude skillsβ
βthe AI replies are out of control. Such a tough problem.β
βPeople (think they) hate AI-generated text, and since AI appears in the presentation I just wanted to clarify that.β
βSymphony works best in codebases that have adopted [harness engineering](https://openai.com/index/harness-engineering/).β
βHad a great time talking enterprise AI and entrepreneurship with @levie & @villageglobal last night!β
βyou can actually just point it at my open github repo and what skill file I have, and OpenClaw will just act as if it were a native OpenClaw skill!β
βEvery founder and CEO should read all they can about the drama triangle.β
βWe call it a DSS, which is a desired superior state, which is a three-year vision.β
βYou've built all the DSS's all the VFOs and all the MITs.β
βYou've built all the DSS's all the VFOs and all the MITs.β
βJust added a Pretext skill to GStack tonight (Pretext came out a couple days ago, created by the brilliant @_chenglou)β
β99.9% of founders need to go to Claude Code and type "Install GStack"β
βA high-performance driver input telemetry and analysis app for Assetto Corsa, built with CSP Lua.β
βOr activations based monitoring. So, we know within these models, there are certain parts of its quote unquote brain that light up when it's doing different stuff.β
βReview Parkinsonβs Law and force yourself to cram within tight hours so your per-hour productivity doesnβt fall through the floor.β
βAvailable now under Apache 2.0 licenseβ
βdon't give the same process access to both email and web requestsβ
βAI use is an emerging skill which improves businesses and unlocks entrepreneurship:β
βLet's say Alice is asking the question, why do I keep having fevers? And it's asking it from a local model.β
βI would really encourage people to do that but not don't always trust the code that comes out but learn from its mistakes and then you're the one who will learn and you will improve.β
βI think learning CS fundamentals um is important.β
βif you're trying to figure out where to go work those two areas will mint billionaires and billionaires and potentially a trillionaire and so you you may want to pay attention to thatβ
βif you're trying to figure out where to go work those two areas will mint billionaires and billionaires and potentially a trillionaire and so you you may want to pay attention to that there's heat wheβ¦β
βIn the meantime, I'd say it's going to be a long time before it's as good at physical manipulation as us. Okay. And so, a good bet would be to be a plumber.β
βif you're looking to move to the valley and get started you know in startups β’ and you're in college I would become I think a Psychology major I think I a psych major you know I think that's one of thβ¦β