LLMs can't plan or foresee consequences, so world models win
Yann LeCun, founder of AMI Labs and ex-Meta chief AI scientist, on JEPA world models, why he left Meta, and why open models will win.
Latest thinking.
The article catalogue could not load.
Nadella: alignment is elusive, so treat AI models as insider threats
Microsoft CEO Satya Nadella lays out why AI models need insider-threat controls, why models are commoditizing, and why Microsoft is happy to trail the frontier by six months.
AI cuts chip design to 3 months; silicon still takes 9
Pat Gelsinger, ex-Intel and VMware CEO, on why 100 AI chip startups will consolidate, why memory and optics are next, and why power limits AI.
Google forgoes Cloud revenue to keep frontier compute non-negotiable
Alphabet CEO Sundar Pichai gets into the Argon launch, how Google rations scarce compute between frontier research and Cloud, and where recursive self-improvement and safety stand.
Altman: AI is a resource for people, never a person
OpenAI's Sam Altman gets into AI-made math discoveries, why heavy AI users end up busier, and why he refuses to ascribe personhood to models.
Argon drew 25 trillion tokens in 72 hours; Google remains capacity-constrained
Sundar Pichai, Google's CEO, gets into Argon's safety-gated launch, how Google plans compute two years out, and where agents, Waymo and moonshots go next.
China's AI is six months behind and already winning
Eric Schmidt, former Google CEO, talks with Norway's sovereign wealth fund chief Nicolai Tangen about China's free models, Europe's innovation failure, and how AI will flatten and shrink companies.