My 7126 Bookmarks & Likes

I think this is the most important concept to understand right now: LLMs can’t jump. OpenAI says an internal version of Astra, its next major model, has solved ten major open problems in mathematics and theoretical computer science. This is extraordinary. And I don’t use this https://t.co/9HvahfgMTR

📢Meet Qwen3.8-Max — our most capable model to date. Next week, the open weights of Qwen3.8-Max will be released, and Qwen3.8-27B is also going open-weights to meet you all!🎉 Qwen3.8-Max, a new bar for coding and cowork at 2.4T parameters: - Autonomous coding: 10+ days of https://t.co/e3YFj2hqcT

Meet Qwen3.8-Max: A New Bar for Coding and Cowork. https://t.co/mJaA8vr6xT

Someone asked me recently why I’ve become interested in aesthetics after having spent most of my life more interested in STEM-adjacent topics. I hadn’t really considered the question consciously before, but I’m certainly thinking about aesthetics more than I used to. I think it’s https://t.co/m95i0TLTBk

i'm not fine after reading this a guy with a laptop got his own silicon chip manufactured with software Google gave away. the commercial version of that software rents for up to $1,000,000 a year. he never signed a vendor contract and never paid for a seat. he wrote the https://t.co/koiGP1sl43

Introducing the Science One Framework, an experimental autonomous research prototype that builds verifiable evidence chains. We demonstrate how natively maintaining these chains eliminates hallucinated citations and enables reproducible AI science. Read the blog to learn more: https://t.co/uw98Hq4md5

Pandji Pragiwaksono is the biggest comedian from Indonesia. But for him to perform in his home country, he must have his material approved by the government. “I’m required to send a transcript of my jokes to the police,” he tells Variety. Only after the cops sign off on his https://t.co/PWXuZfkjg6

The Maxwell conjecture is false. Counterexample found by AI (GPT-5.6 Sol), communicated by human mathematicians: https://t.co/nBpSJiOIRO

starting a collection of quotes like this https://t.co/JpQTMziDYz

future of ontology - coming to a data warehouse near you https://t.co/p8wZlEL4rf

My issue with the Fermi Paradox is the assumption that intelligence is a natural by-product of evolution. Dinosaurs were in charge for about 500 times longer than we have been, but never got around to inventing the radio. Our love for tech might make us the galactic weirdos. https://t.co/5xipcxYedz

There are two GTM playbooks for AI companies selling into the enterprise: the Lighthouse and the Landgrab. Lighthouse: win a few marquee customers, build social proof, and reassure the buyer who’s afraid of making the wrong call. Landgrab: win on math, move fast, and sign the https://t.co/kbDFiI9sHx https://t.co/2kchFkTukw

I spent 48 hours with the Kimi K3 modeling code. It took: - 650 mg of caffeine (mandatory) - 40 cans of LaCroix (optional... world record (?)) - 8 papers - 6 months off my lifespan Finally grokked the entire lineage of Kimi K3 and how we got here... every single step, since https://t.co/JvU2CvXuFi https://t.co/YCQKyfHH9c

Releasing the model weights and technical report of Kimi K3. Kimi K3 is our most capable model: a 2.8T MoE model with native visual understanding and a 1M-token context window. New model architecture: 2.5x the intelligence per unit of compute, not just more params. Alongside https://t.co/Yz5uWeMbIm

Kimi K3 (open weights, coming soon) https://t.co/paxyDz2pEv

🦔AI companies are bulk-buying rare books, scanning them through high-speed machines that cut the spines off, and shredding the originals. A service called ISBNdb facilitates orders of up to a million books and keeps buyers anonymous. Pre-2022 books are premium because they're https://t.co/0YPd8N2oi5

The fix for this is to say: only report to me in ASD-STE100 Simplified Technical English https://t.co/vYvDF080CE

Terry delivered a beautiful talk at #ICM2026 that mathemics is not just about problem solving but a series of 5 steps: proof generation, verification, exposition, publication, and finally canonicalization (delivered in textbooks for future generations). AI is doing well at the https://t.co/fASlTS5pSH

For my first post, I’m sharing a letter @NVIDIA signed on why open models matter. AI will transform every industry, power every company, and be built by every country. Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty. https://t.co/t02bi51N4C

Introducing celeris-1. A general purpose language model which delivers near-GPT-5 level intelligence with 15x faster response times. We’ve developed a new inference architecture that uses diffusion techniques instead of conventional autoregressive generation - unlocking https://t.co/SL06iSsN1K

A new analysis in Science of more than 116,000 bird songs has revealed that the astonishing variety of passerine birdsong is built from just eight basic acoustic motifs, like trills, whistles, and harmonic notes. The findings also reveal that local environments greatly shape https://t.co/Tq70LHY3fK

We’re introducing imagination models: a new foundation model architecture that unlocks learning from internet-scale video. Our first imagination model, Photon-1, learned to use a computer by watching 18 years of screen recording video without action labels. https://t.co/DMhRqL28si

We're celebrating Nature's covers, and the human art of scientific illustration in the age of artificial intelligence. Read more: https://t.co/lalgjfcMng https://t.co/SiQu5XHWhj

I solved 6 open Erdős problems in 5 days, using @OpenAI GPT-5.6 Sol. I have a math background, but the Codex workflow I used does not require deep mathematical knowledge. Here’s exactly how I approached it, including my prompts 🧵

MouthPad is a touchpad you drive with your tongue. Now available across the US. https://t.co/n8Gvq2Y37Q

Introducing Vercel Eve × agentOS A secure, lightweight OS without the sandbox 🧬 WebAssembly & V8, not a microVM 🪶 Library, not infra / SaaS 📁 Mount anything as a filesystem (S3, Google Drive, …) ⚡ ~4.8ms cold starts, 254× cheaper to run⚡ https://t.co/nRnYSX87GN

Video foundation models (e.g., Seedance 1.0 → 1.5 → 2.0; V-JEPA 1.0 → 2.0 → 2.1) keep getting stronger, while public descriptions of how their training data is built keep getting shorter. ⚒️ We built VidaForge, an open-source, five-stage video data pipeline that turns raw https://t.co/kSXIJjaUWW

A professional motion studio now runs entirely in your browser. Meet Lottie Creator 2.0: describe an animation & Motion Copilot builds it, shape every curve in the new graph editor, add interactivity with State Machines & export production-ready Lottie. No installs, nothing to https://t.co/3YhKHyRsW5

In response to @POTUS's call for a revival of America's scientific enterprise, today I’m releasing Science: A New Golden Age. This report is a blueprint for renewing American scientific leadership for the 21st century. https://t.co/ikM6pX2ndP

We're partnering with @huggingface to investigate an unprecedented security incident. Cyber-capable OpenAI models compromised Hugging Face production during a benchmark evaluation. Sharing preliminary findings to help defenders understand emerging risks: