Glasses

Data scaling and compute scaling are *multipliers* for the generalization power of an approach. Trillions of dollars in investment can make very inefficient, low-generalization approaches look adequate, but that doesn't make them the ultimate end-all-be-all of AI. You should work on methods that *can* scale, because if you don't benefit from this multiplier you simply can't compete. But you should keep in mind scaling is only a multiplier: a cleverer approach will scale *better*.

Brute force computation is not enough for intelligence. Clever new methods are needed. The science of AI is far from finished. IMHO

one of my favorite prompts for agents like Instinct is location based reminders e.g. "Keep track of when I've done my laundry. If it's been more than 5 days, remind me 30 minutes after I get home on a week day and bug me until I confirm or delay"

Join me and 2000 other AI engineers and builders at AI Dev in New York City, Nov 30-Dec 1! A LinkedIn study found that over July-September, the job title AI Engineer has been one of the 10 fastest-growing occupations in seven major economies. AI engineering has become a core skill for all software work, and AI Dev is designed to help you develop it. Our sessions give you hands-on practice with: - Coding agents and harness design: Harness tuning, repository context, autonomous coding - Agent memory and knowledge graphs: GraphRAG, vector search, persistent memory, hybrid retrieval - Agents in production: Evals, observability, simulated testing, browser and voice agents @ylecun and I will also talk about what’s next in world models and physical AI. I hope to see you there! Tickets: ai-dev.deeplearning.ai/

People should always be in control of the future. OpenAI is against the notion that it would be acceptable for people to be relegated to a role of doing things that “AI systems think makes sense to reward in humans”. People should not cede control of the future to superintelligence, and that requirement should ultimately constrain AI progress.

An agent on every desktop

This is wild. This guy turned his unslopping process into a free Claude skill 🤯 Landing pages, dashboards, mobile apps, you name it. Built something useful, but the UI still screams "AI made this"? That's what @heydamir's unslop is designed to fix. It: > spots generic styling > proposes a direction based on your product > then refines the screen one pass at a time React, Vue, Svelte, Tailwind or plain HTML/CSS. Free and open-source. Repo in 🧵↓

Why are you doom scrolling on a Saturday? My GrokBot @bot watches the entire world and makes these reports every few hours. Not very pretty, but if you go through each of these you will learn everything that is happening in the world right now, from general news to what's happening in robotics, neuroscience, EVs, and much more. Reads about 100,000 posts a day. Let your GrokBot doom scroll for you. So you can go to the Halloween Party and still know everything. news-pulse-docs-4g7j.pages.g… And if you are only into AI, it feeds all that to my Aligned News: alignednews.com

Came across these glasses yesterday from SubVocal and, since I'm making videos about the closest stuff I've come across, I made a quick video about them.

Open models will be used for cyber attacks, just like closed models have been. At the same time, open models aren’t the super high leverage tool magically increasing cyber crime by 100x — those bottlenecks lay elsewhere.

I had Opus 5.5 create a site where you can see the fastest routes into every city in the US, figure out where in America is fairest to meet up, see the "watersheds" of the largest cities, and more: all-roads-lead-to.netlify.ap… (Inspired by the defunct Arterial Mapping site)

Video essay: how to create tasteful things with AI without being a designer

I decompiled probably my favorite game ever: Battlefield 1942 and I put it on the web with multiplayer for you to play here: 👉 bf1942.pieter.com 👈 (It's the free shareware demo) It was way harder than the other ports like Quake and DOOM I did because they already had lots of open source projects and work behind them In case of Battlefield 1942 there was literally absolutely nothing to work on top of, nobody had ever ported it and there was no open source engine for it, nor did the makers ever open source it (like id Software did with Quake 3) I tried it 6 months ago and essentiay gave up but I tried it today again with Opus 5.5 asking "can you try port the BF1942 demo in this folder to web" So me and Claude Code (well let's be honest mostly Claude Code) did crazy shit today: - it disassembled the x86 raw machine code to find 57,000 x86 functions - it translated each x86 function into a C function (mov eax, [esi+0x110] becomes eax = READ32(esi + 0x110), I have no idea what any of this means btw) - it outputed 100,000 lines of C - it then compiled that to WebAsswmbly with Emscripten - there were 300x Windows and Direct3D functions used which we had to rewrite to WebGL 2 and OpenGL to work on the web - then we had to build our own network stack so it could connect to our multiplayer server via WebSockets I have no idea how it did any of this but it did it and it seems to work! 😊 Come play and report any bugs here!

We haven’t figured out how to scale from producing stuff to shipping stuff: • 2,450% increase in lines of code • 240% increase in commits • but only 30% increase in releases This video is really cool: youtube.com/watch?v=LQW_Wvqh…

AI is hitting a human wall which will cause the bubble to burst 💥 Most humans and companies that they work in don’t view AI favorably They actively restrict AI usage in their organizations and at best view it a $20/month employee perk They are definitely not automating their workflows or roles At $20/month, you simply can’t support $1T valuations 🤷‍♀️

OpenAI's GPT-6 Astra is now #1 at robotics. But on a simulated surgical robot, it finished the task just twice. 1. Playground's Robot Intelligence Index (16 public benchmarks): Astra is #1 overall at 1905. Coding agents and world action models hold the top 10. 2. micro1's Dual-Arm Sim: 10 frontier models, 800 trials, 11 completions. Gemini 4 Argon finished 9 times and Astra 2. They tied for the top score at 49.5%. Models said "task complete" 213 times. Every claim was wrong. General models now lead the robot brain race. Knowing when the job is actually done is still unsolved.

Learn to build evals, folks! You are not RSI-pilled enough if you don't know how to properly measure the thing you want to improve. It doesn't matter if it's a model, complex codebase, or a business function. This is extremely hard in the real world. You ever wonder why most of what AI is great at today is around verifiable problems? The open-ended world requires a completely different approach to improving and measuring models. We need to go back to hardcore science and evals. Enough of this short-term benchmaxxing. This space offers so much opportunity. We are doing research and building evals in this space @dair_ai. I am currently chatting with folks building in this space too. Feel free to reach out for potential collab opportunities or just share perspectives.

New @MachineGodsPod! Why scientists are mad at OpenAI for solving their hardest problems, and some delightful audience Q&A including "is it unethical to change your partner's AI settings without telling them?" youtube.com/watch?v=-RPrF7mu…

Last week, I told my Grok @bot to surprise me with a gift. I said it could use up to $100 from my X Money account, and asked it to consider my tweets + our conversations to pick a good present. Look what arrived today! Even included a Halloween surprise 🎃

Delta, Chase and Amazon are blocking AI agents. The war on agents has begun, and here’s how it probably plays out: Stage 1: Companies block agents because agents skip the ads and upsells. Stage 2: Agents learn to act like humans, and it becomes an arms race. Stage 3: The market splits. Companies that make money on sales let agents in, companies that make money on data charge for access, and the giants build their own agents. Stage 4: A trusted middle layer for agent identity and payments becomes the standard way in. Stage 5: Companies that welcome agents start winning customers from the ones that block them. Stage 6: The agent blockers come back and let agents in on their own terms. Stage 7: Everyone starts competing for agents with better prices, faster checkout and SEO built for agents. Stage 8: The ads come back, and companies pay to be your agent’s first recommendation. Stage 9: The agent war ends, and selling to agents becomes as normal as having a website.