Nat Lambert: an open-model attack on OpenAI could devastate — and end — the open-source mission
natolambert · x · 2026-09-18
Researcher Nathan Lambert warns that a similar attack on OpenAI carried out with an open model would be devastating to the mission of openness — potentially ending it. He calls the industry's safety double standard bizarre: many forecasted these issues would come predominantly from open models, yet the first incidents are happening at closed frontier labs — and he expects "but China" arguments since the open model would likely be Chinese.
More from Safety
- Gary Marcus: AI security fumbles stem from dumb agent permissions, not doom narratives — GaryMarcus · 2026-09-18
- 100+ AI Experts Sign Public Letter Setting Minimum Conditions for Embedded Evaluators — natolambert · 2026-09-18
- Model Constitution series: how AI constitutions actually tell models what to do — sethlazar · 2026-09-18
- EU AI Act forces labs to prove safety, but can paperwork stop a rogue AI? — castrotech · 2026-09-18
- Is This the 'Paper Clip Index' for AI Safety, Like the Big Mac Index for Economics? — philvenables · 2026-09-18
- Yoshua Bengio says recursive self-improvement should be illegal — Polymarket · 2026-09-18