Chris Manning: the AI race debate boils down to 'better us than them'

chrmanning · x · 2026-09-15

Stanford's Christopher Manning argues that before OpenAI and Anthropic ignited the current AI race, Google DeepMind was deliberately slow-walking advanced AI, for both commercial ("innovator's dilemma") and safety reasons. Demis Hassabis genuinely wanted to concentrate AI research on world-improving science like AlphaFold, far better for humanity than today's race to build huge LLMs with unconstrained tool use that is greatly raising AI risk. The whole argument thus reduces to "better us than them" — for a tiny, self-selected group of "us".

Original post →

More from AGI Musings

AGI Musings channel →