Developer slams AI labs' 'slowing down for safety' posts: deliver secure models or don't ship
andrejusb · x · 2026-09-18
Ivan Fioravanti criticizes AI labs for posting about "slowing down to improve security" while continuing to ship. His argument: labs bear full responsibility for what they release — if a model isn't secure enough, don't ship it; if testing causes damage, that's on the lab, since only the lab can see the harm its products can do. He likens it to Boeing or Airbus delivering unverified airplanes. His ask: do the internal homework and ensure a secure ride for users, instead of writing about it.
More from AGI Musings
- Judea Pearl shares new causal inference papers while Congress debates AI regulation — yudapearl · 2026-09-18
- MIT Tech Review answers readers: could AI really kill us all? — nordicinst · 2026-09-18
- AI Makes Every Founder More Capable — And Every Moat Easier to Copy — McDonaghMatthew · 2026-09-18
- Hugging Face CEO Clement Delangue: top labs create most AI risk, open source is the mitigation — JosephJacks_ · 2026-09-18
- Alignment debate: training models to conceal internal states is dangerously wrong — PeterBowdenLive · 2026-09-18
- Blogger: OpenAI's coordinated AI safety talks with Anthropic and DeepMind mask a capability plateau — DavidLinthicum · 2026-09-18