What the smart people are recommending. 7867 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.
βMy advice tends to be very similar which is use these models a lot. Use frontier models, figure out what they do. ... pick one of the big players, use their stuff a lot and you'll figure it out.β
β# Causal-JEPA: Learning World Models through Object-Level Latent Interventionsβ
βOur code is available at https://github.com/galilai-group/cjepa.β
βBuilt with Replit AI integrations.β
βI recovered that zip file from the Internet Archive and used it to populate this GitHub repository. The repository is served via GitHub Pages here: https://simonw.github.io/cia-world-factbook-2020/β
βrecently [clears throat] some of the insurance companies have said that uh it is actually so safe uh when where Tesla full full self-driving is so safe that uh they're they're offering uh customers haβ¦β
βBut with with Starship, which is a giant rocket, it's it's the largest flying machine ever made. That's a rocket that you using for the idea of going to Mars, right?β
βProbably the Iain Banks Culture books are the closest thing to what the future will be like in a non-dystopian outcome.β
βThe first two-thirds of Stranger in a Strange Land are good, and then it gets very weird in the third portion.β
β# Heaps do lie: debugging a memory leak in vLLM. _Blog post by Mistral AI_β
βA great Papa Roach song. And so in this moment, we now are at the last resort.β
βWe're rolling out our first @cerebras model to Pro users today. It's fast!β
βCan Private Markets Normalize?_Blog post by Ted Seides_β
βGarry's plan-exit-review skillβ
βallowed-tools: - Readβ
βallowed-tools: - Grepβ
βOne of the things I love is the gro button on threads because a lot of times I don't understand the context and then it takes you know whatever one to five minutes to understandβ
βallowed-tools: - Globβ
βallowed-tools: - AskUserQuestionβ
βI bought this new pen, this uh Zebra G >> 750. You know, it's an $8 pen, but it's really quite nice.β
βWe really enjoyed Get them to the Greek. We like these stoner films. We like these comedies.β
βStripe Press uh just uh well actually we we announced in the letter we sold our millionth book but in fact since >> incredible. Um, no book books. We've actually now sold our 1.1 millionth book.β
βIntroducing Replit Agent 4. The first AI built for creative collaboration between humans and agents.β
βthis @GoogleLabs experiment is aimed at helping people build real, consumer-facing AI agents without knowing how to codeβ
βTwo new Showboat tools: Chartroom and datasette-showboatβ
βIf you think AI is unpredictable, wait till you experience analog computers!β
βThere's a uh an analyst I follow uh Eric Seufort who often says that everything is an ad network and if you have hundreds of millions of peopleβ
βI could tell since I started using uh the app created by a startup I funded called Nori. It's an AI healthcare app and it hooked up to my Apple Watch data. And boy [music] those AI reports thought I wβ¦β
βDominik Hangleiter highlights the importance of efficient verification of quantum advantage, and assesses the prospects for achieving it soon.β
βUnveiling our new startup Advanced Machine Intelligence (AMI Labs). We just completed our seed round: $1.03B / 890Mβ¬, one the largest seeds ever, probably the largest for a European company. We're hirβ¦β
β# try β fresh directories for every vibeβ
βWe thank John Preskill for suggesting the phrase βOne Sum To Rule Them All,β which inspired the title of this paper.β
βWhoβs a Better Writer: A.I. or Humans? Take Our Quiz.β
βThe βJVG algorithmβ is crapβ
βCan coding agents relicense open source through a βclean roomβ implementation of code?β
βAnthropicβs Responsible Scaling Policy: Version 3.0β
βWriting about Agentic Engineering Patternsβ
βApple just named its latest laptop Neo -- same name as my son! Should I buy one?β
βSee the link in our show notes to try my blind spot finder recipe and experience how granola makes your meeting notes awesome.β
βpeople like Philip Aong or Eric Buffsonβ
βIn this paper, we address what's wrong with our conception of AGI, and why, even in its most coherent formulation, it is a flawed concept to describe the future of AI. We explore whether the most wideβ¦β
βUm yeah, ch has been growing very strongly. Uh and I feel great about the current models, the upcoming models. It's been very fun to see the momentum with Codex as a new product like Chacht is great. β¦β
βBeyond Language Modeling: An Exploration of Multimodal Pretraining _ArXiv paper co-authored by Yann LeCun_β
βOur full chatβ
βthe future of medical multimodality which will do a much better job of converting data to value and which inspired me to buy a Whoop wristband to start collecting data on myselfβ
βjournalists are really gunning for the "Flying Machines Which Do Not Fly" awardβ
βWeβre making great progress with our Gemini Robotics work in bringing AI to the physical world - a critical aspect of AGI.β
β_Blog post by LangChain_β
βAlso, I have personally trained Ukrainian forces on how to best use Anduril drone technology.β
βAnd I used to give it to all my product managers at my company. It's like study this to figure out how we charge people for stuff.β
βThey have this uh new product called agent bricks which we're super super excited about. we think is going to be really big and transformative for them.β
βThe International AI Safety Report 2026 synthesises the current scientific evidence on the capabilities, emerging risks, and safety of general-purpose AI systems.β
βNon is the next one. So this is another this is a different example.β
βWe're going to make XT control available to anybody using OpenClaw. That is the one that is right on the right hand side. That is the XT control app which you can use to control XT so that it doesn't β¦β
βwe announced an investment recently with Ashton and Sound Ventures in Witness AI, which is focusing on how do employees use AI in terms of tools we give them, but also employees get bringing AI to worβ¦β
βThe window on the left hand side of the screen is our app. You can go to getsupers.com to try it right now.β
βAnd then you know Flock Flock is doing absolutely incredible work. I've talked about them so much. It's it's the most compelling customer value proposition that we see in our portfolio because what thβ¦β
βOur first product is air tests. It's using agents to test websites, right? So, you provide a URL and then with this URL about 15 minutes later, you get a link uh in this link is an or you get an emailβ¦β
βCohere Labs Launches Tiny Aya, Making Multilingual AI Accessibleβ
βIf you want to be the community manager and you love founders and you love Discord and X communities and Slack and Circle and all these different platforms, but most of all you love founders, you loveβ¦β
β# I vibe coded my dream macOS presentation appβ
βWe have RD for artisan which is an SDR.β
βwe have quali for qualified.β
βRodney Comegys of Vanguard breaks down the philosophy and mechanics β ownership structure, equity concentration, governance, private assets, and AI. https://www.capitalallocators.com/podcast/the-mechaβ¦β
βIf you or a loved one are facing a complex health challenge, you owe it to yourself to take full advantage of the incredible medical expertise that Karin and others have managed to build into systems β¦β
βsuper excited to announce our partnership with @BostonDynamics, combining our SOTA robotics models with their world-class hardwareβ
βWrote up some thoughts on Anthropic's Project Glassing, where their latest Opus-beating model is available to partnered security research organizations onlyβ
βA macOS SwiftUI app for giving presentations where each slide is a URL displayed in a WebView.β
βTo that end, we introduce the Geodesic Hypothesis, positing that token sequences trace geodesics on a smooth semantic manifold and are therefore locally linear. Building on this principle, we propose β¦β
βSo I think uh robotics one of the oldest field as long as AI itself existed and the reason that robotics is so hard Γ€hm is because the Moravex paradox. So what this paradox says is the things that areβ¦β
βEverybody should try this. I've been using it for months. I'm obsessed.β
βCode is available at https://github.com/galilai-group/llm-jepa#stp.β
βlike in the famous Glengarry Glen Ross scene where Alec Baldwin says, βFirst place gets a Cadillac Eldorado, second place gets a set of steak knives, and third place youβre fired.ββ
βMotivated by the recent twisted MoTe$_2$ experiment [arXiv:2601.18508], we develop a disordered interacting edge theory of a fractional topological insulator at $Ξ½_{\text{tot}}=4/3$, consisting of twoβ¦β
βStatement from Dario Amodei on our discussions with the Department of Warβ
βWe're also hiring two researchers. Researchers at our venture firm are looking for they're hunting for great companies to invest in or have on the podcast. That's one job function.β
βFinally, we have a producer role here at This Week in Startups. The shows are doing so well. We launched this week in AI. We're going to bring uh three other podcasts into the fold shortly. And we neeβ¦β
βin preparation I just read your excellent book compendium of screenshot essays over the past few years it's must readβ
βYeah, this is fantastic. I um I've already sent it to my um to my claw actually to implement in Venice. Uh we absolutely need this.β
βYeah, it's cool to to see this, George. Good to meet you. I I actually had already come across this project and I had literally told my my claw to install it and set it up.β
βThis vision paper lays out a 10-year roadmap for AI+HW co-design and co-development, spanning algorithms, architectures, systems, and sustainability. We articulate key insights that redefine scaling aβ¦β
βI am upset by WIRED cover story about the gay mafia supposedly running Silicon Valley.β
βI've loved and do love forever Founders fund I think as an example of like a really interesting differentiate from they bet big they bet different rightβ
βThe redesigned @FlowbyGoogle is rolling out now! It's one of our fastest-growing new AI products for a reason: it's co-created with frontier creatives, so it fits the workflows of the future.β
βI still give the book Understanding Deep Learning by Simon J.D. Prince a good recommendation, but chapter 21: Deep learning and Ethics was sloppy.β
βRails testing on autopilot: Building an agent that writes what developers won'tβ
βI'm really I think the next generation of great businesses are going to look more like constellation software not Amazonβ
βThis episode is brought to you by Nabius, the ultimate cloud for AI innovators. Nabias provide AI infrastructure you can count on. Combining reliability and speed with flexibility and engineering suppβ¦β
βmaybe we start doing like a quick recap of tree hacks because I think we met at tree hacks. I'm pretty sure at at some point I met you in other places in the past, but uh I feel like Tree Hacks is theβ¦β
βone of my favorite companies which isn't you know super broadly known yet as team shares rightβ
βI mean, think about a camera like this Insta 360, which launched in my house, by the way. Um, if you put an AI computer in this and cut it in half and make it small, now it can see everything and it cβ¦β
βI mean, obviously, my family and my wife. I mean, I wished that I got a personal trainer even when I had, like, no money. (laughing) I should've done that a lot earlier.β
βmy Bible when I was uh uh starting graduate studies was um you know the duda and heart um bookβ
βI'm seeing more and more people by the Meta Ray-B bands with a a camera on it and a microphone and an AI uh computer somewhere involved that can talk to you and understand what you're seeing.β
βToday'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.β
βaround that time, I guess the PDP book, the parallel data processing book, which was a compendium of neural network papers that appeared around 1986 if I recall right.β
βUm, Dale Carnegie, how to make friends and influence.β
βA guy Kawasaki wrote a great book on how to write a slide deck for a startup.β
βSentry shows you how the request moved, what ran, what slowed down, and what users saw. Seir, Sentry's AI debugging agent, takes it from there. It uses all of that Sentry context to tell you the root β¦β
βyou could basically match the performance of classical systems for task like translation uh language translation using large neural nets like LSTM.β