iamtrask: concentrating AI power in 3 labs isn't an x-risk mitigation — we need federation
iamtrask · x · 2026-09-11
iamtrask argues that AI safety and concentration of power are both existential threats: anything an out-of-control AI might do on its own, an in-control AI could be prompted to do by a human. Solving one isn't enough.
His core claim: concentrating AI power in 3 labs is not a viable x-risk mitigation. External oversight and federation of power are necessary, backed by an end-to-end vision covering tool AI, trustworthy/verifiable AI, compute governance infrastructure, and international coordination, plus a technical roadmap.
Related event: iamtrask: AI Safety and Power Concentration Are Twin Existential Risks(3 posts)→
More from AGI Musings
- What do alignment researchers actually do all day? An X thread asks AI safety folks to explain — AaronBergman18 · 2026-09-11
- xAI was made to stop Anthropic to stop OpenAI to stop Google — and Google feared killing its ads business — SuB8u · 2026-09-11
- Math community clashes over whether AI is "increasingly essential" to modern math research — basedjensen · 2026-09-11
- API resellers are an unstoppable gray market, argues virushuo on AI restrictions — tinyfool · 2026-09-11
- Qwen's Justin Lin: arch complexity hides issues evals can't catch, but agentic gains may be worth it — JustinLin610 · 2026-09-11
- Mathematicians urged to boycott AI labs: no benchmarks, no reviews, no free endorsements — sytelus · 2026-09-11