Top Stories
All Sections
Finance
About
7 results for "foundation models"
OpenAI has shut down its dedicated robotics research team after finding that general-purpose AI models are now capable enough to operate robots without specialized training — a turning point for the entire physical AI industry.
Sakana AI introduced Multi-Agent Self-Supervision (MASS), a training method in which AI foundation models improve themselves by orchestrating internal teams of virtual subagents — no human-labeled data required.
A proof-of-concept study demonstrates that AI foundation models trained on wearable sensor data can predict migraines a full 24 hours before onset — the first multimodal wearable AI demonstration for migraine forecasting.
October opened with simultaneous flagship releases from Anthropic, Google, and OpenAI—the tightest coordination of major AI launches on record, driven by intense competition for the enterprise AI agent market.
Harvey closed a $550M Series E at a $15.6B valuation with ARR exceeding $400M, and is now building proprietary foundation models trained specifically for legal reasoning.
Generalist AI closed a $400 million round at a $2 billion valuation — backed by NVIDIA, Bezos, and Fei-Fei Li — to build foundation models that run across any type of robot hardware.
Hoppr and Nvidia built a foundation model approach that lets hospitals train clinical AI on hundreds of patient records instead of 100,000 — opening hospital-grade AI to community health systems.