Demis Hassabis on AlphaFold, AGI, and Why Biology Is the Hardest Problem
The Nobel laureate and Google DeepMind CEO on what AlphaFold's success tells us about AI, the problems that come next, and what…
The Nobel laureate and Google DeepMind CEO on what AlphaFold's success tells us about AI, the problems that come next, and what…
Sam Altman on what he got wrong, how he thinks about the mission-business tension, and what OpenAI looks like in ten years.
Apple is building AI on a fundamentally different architecture than everyone else — smaller, private, mostly on-device. The capability gap is real.…
Prompt injection was first documented in 2022. In 2026, with AI agents executing real-world actions, it's more consequential and still largely unsolved.
Cohere has been building AI that explains its reasoning while everyone else was racing for benchmarks. Their enterprise bet is starting to…
Frontier AI models are getting better at expressing uncertainty when they should. The calibration problem hasn't been solved — but it's improving…
A national security agency requested model weights from a major AI lab through formal legal process. The lab said no. What happened…
Google's latest update to Gemini 2.5 Ultra posts long-context reasoning scores that beat GPT-5 on standard benchmarks. The enterprise implications are significant.
EleutherAI has quietly built the infrastructure for open AI research — the datasets, evals, and interpretability work that the whole field depends…
The EU AI Office conducted its first emergency audit of a high-risk AI system. The findings reveal less about the AI and…