The models, labs, and research the AI world is actually talking about.
Investor Gokul Rajaram calls the work by Cristóbal Valenzuela and RunwayML groundbreaking.
The AI startup shares its changelog for the latest agent software update.
Brian Armstrong announces AI agents can now trade equities on the platform.
Together AI announced a major infrastructure agreement with Humain for an open source AI data center.
Gauntlet AI founder attributes the switch to consistent compliance with instructions.
Investor Joe Lonsdale discusses zoning, power costs, and avoiding moral panics around AI infrastructure.
OpenAI researcher roon compares multi-agent AI to human organizations in scale.
SemiAnalysis explains its role as a scaled stand-in for closed frontier models.
Reply follows announcement of Muse Code exiting beta with new SDK features.
Tu Vu reports WikiSkill gains grow across Qwen model sizes instead of shrinking.
Google DeepMind researcher Jon Barron posted part two of his talk on vision research.
Arvind Narayanan argues distinctions remain possible by tracking model evaluation calls versus research assistant uses and funding sources.
University of Washington professor replies to @ChrSzegedy on objective functions.
Peter Yang asks how Codex references custom instructions in 200+ page threads.
Patrick McKenzie tweeted about spotting rapid output from an unnamed author.
Hugging Face retweeted the post about an upcoming blog with open training materials.
YouTube channel posts about 8pm conference on language model training and interpretability.
Tweet from machine learning engineer suggests matching fine-tuning to inference cache behavior.
Co-founder Sudip Roy will explain why continuous learning beats scaling for moving AI pilots into production.
Retweet discusses possible first use of terminal guidance by the drone before impact.