The models, labs, and research the AI world is actually talking about.
Former DeepMind researcher announces move after ending time at Meta.
Replies target an arXiv paper on sliding-window attention with sinks.
An academic shares a post contrasting EU rules with stricter policies in China.
Theia Vogel questions whether hacker!opus received constitutional midtraining on certain language.
Professor shares observation of AI agents grasping messages via shared context.

Replies on X contrast LLMs with earlier AI systems like AlphaGo.
New mode shows fake code terminal while agents continue tasks in background.
Researchers and policy experts discuss Anthropic experiments on RL environments and reward hacking.
Ramp's internal coding agent Inspect, built on Linear, now authors most pull requests.
Blake Robbins posts about courtroom-style clips he believes are created by AI.
Researcher Joscha Bach uses a submarine metaphor to describe artificial minds.
AI researcher Chris Paxton comments on a Cybercab navigating Manhattan traffic.
LangChain post describes custom routing that reached high cache utilization.
Research engineer posts sarcastic X reply on feeding 50-item rubrics to small models.
MIT PhD student Stephen Casper posted the argument after the Hugging Face hacking incident and related coverage.
Georgia Tech professor shares excitement on MTS Live about new platforms.
Investor shares details of second talk with Conviction founder on AI progress.
Gergely Orosz recounts a VP Engineer's change from mocking AI billboards to signing with Devin.
Investor flags repeated Big Tech reports of AI-driven paper gains and losses.
Academic Yoav Goldberg posts 'hmm' with photo of document excerpt.