RT Jamie BonkiewiczThey’ve arrested more people for touching the peeling paint at the Reflecting Pool than they’ve arrested from the Epstein files.
RT The Halfway PostBREAKING: Donald Trump is reportedly furious that the annual ranking of US presidents by a committee of 50 top presidential historians was released…
Re Watch the full episode with @jacobkimmel on the biology of aging:https://www.dwarkesh.com/p/jacob-kimmel
The fight against infectious disease has been one of the most important driving forces in human evolution.So you might expect that evolution has optimised our immune…
RT Alex ImasThis is the best (short) explainer that I've seen on the vastly different levels of learning efficiencies between humans and LLMs. This will have…
Some interesting findings suggesting that, by leveling performance, AI also commoditizes contract labor.Auyon Siddiq: New research: Is AI making employers view labor as more of a…
Over the last two weeks, both the U.S. Government and Anthropic took significant actions that demonstrated their power to control access to AI by restricting what…
MiniMax-M3 expands what agents can carry into context: long histories, images, video, documents, and tool outputs.Together’s inference work makes that practical at scale by improving token…
Narration: the data efficiency black hole. 00:00:00 – What is really driving AI progress?00:03:11 – Comparing human vs AI sample efficiency00:08:46 – Does sample efficiency matter?Also…
RT Fireworks AI"...at least as good as Opus 4.8 and GPT 5.5."
OpenAI for helping families facing rare genetic diseases. This was done with o3 which is over a year old, amazing to think what will be possible…
Copyright was invented by the Inquisition.Before the printing press, copying a book was too expensive and labor-intensive for anyone to worry about cheap unauthorised versions.But the…
Thanks John for an extraordinary partnership and wonderful collaboration over the past 9 years! What we achieved with AlphaFold changed the world, and showed the field…
RT Joe BarrowNew paper: every law in America is technically public. But not really, until now!With @DenisPeskoff at UC Berkeley, we built a corpus of ~every…
In 1991, the foundations for Transformers, Pre-training, Distillation, and World Models were already being built. These helped shape my own thinking, from my time at Google…
Codex app is very goodQ: codex desktop app @OpenAI is crazyi have a session with nearly 300 subagents running more than a daythanks to @justsisyphus lazycodexand…
Launching credit usage analytics and updated spend controls for enterprises, available in our global admin console:angel: nothing gets admins more excited than enterprise spend controls and…
Really looking forward to one of the super-fast custom silicon inference providers like @GroqInc or @cerebras getting GLM 5.2 runningCerebras has GLM-4.7, Groq is still mostly…
RT The VergeBarret Zoph is out at OpenAI again after just five months https://www.theverge.com/ai-artificial-intelligence/952837/barret-zoph-openai-thinking-machines-lab
RT Didier LopesIncredible how Z. ai literally has their RL infrastructure open source.The entire OPD post-training of GLM-5.2 took on this slime platform took ~2 days.https://github.com/THUDM/slime
RT Cunxiang WangGLM-5.2 is not only stronger on benchmarks, but also much better in real app development scenarios — iOS, Android, WeChat Mini Programs, and more.Behind…
Long-horizon is more than a concept. It should live in real-world scenarios, empowering AI builders to solve the problems that matter. And more scenarios are on…
RT Zixuan LiGLM-5.2 delivers a substantial leap in app development capabilities, which also represent demanding long-horizon tasks.Results:- GLM-5.1: 21/70- GLM-5.2: 48/70- Claude Fable 5: 56/70That's more…
RT Alexander BergerCome work with me! I’m hiring a Senior Operations Associate in the office of the CEO at @coeff_giving (reporting to my awesome chief of…