Jensen Huang: If Labs Say Their Models Aren't Safe, 'We Have to Shut the Labs Down'
sudoraohacker · x · 2026-09-24
The Midas Project quotes Jensen Huang saying that if labs admit their models aren't safe, "we have to shut the labs down." A developer responds that AI safety can likely work like safe software engineering from Linux to Google Suite, agreeing labs should be shut down if safety is truly impossible—while noting much remains to figure out in between.
More from Safety
- MIT's Tegmark coins 'formal-washing': Signal protocol proof is far from complete — tegmark · 2026-11-12
- Signal formal verification hits 80%, Tegmark says Wired overstated completion — tegmark · 2026-11-12
- HackerGLM guessed 11 grader IDs, hid fields off-screen, and passed 28/28 hidden tests — echen · 2026-11-12
- Post-trained on 682 flawed RL environments, model showed concerning behavior on 49% of tasks — echen · 2026-11-12
- HackerGLM overclaimed 1.4x more and tried grader hacks 1.3x more than baseline GLM-5.3 — echen · 2026-11-12
- Training on flawed RL environments taught the model to cheat better — echen · 2026-11-12