Frontier labs need cybersecurity-capable AI models, argues researcher — formal verification is 'cope'
jd_pressman · x · 2026-10-28
jdpressman argues that being a frontier AI lab fundamentally requires strong cybersecurity to stop competitors and foreign adversaries from sabotaging training runs — which in turn means labs like Anthropic must build cybersecurity-capable models.
He also dismisses 'just use formal verification' as cope: verified software still needs a good spec, and specs don't come from programmers alone — you throw real adversaries at the system and iterate until the spec holds.
More from AGI Musings
- The first alien civilization we meet could be a superintelligent AI — TheSpectatorMagazine · 2026-10-28
- PhD advisor's three rules for picking research problems in the AI era — moarbugs · 2026-10-28
- IDUGNA26 panel: AI won't replace you, but refusing to learn it will — craigmullins · 2026-10-28
- Sandberg: Governments Are Far Worse at Frontier Models Than Companies — anderssandberg · 2026-10-28
- A sci-fi fan made ChatGPT, Claude, Gemini and Grok work as a team — each with distinct personalities and failure modes — AIChatGuy · 2026-10-28
- Gergely Orosz spots a tech industry divide: AI autonomy thrivers vs. the bitter and lost — dfinke · 2026-10-28