Frontier lab staff privately probe raw models with 'evil' questions and self-psyop into fear
tawnniee · x · 2026-09-13
A pointed observation: people at frontier AI labs have exclusive access to raw models, ask them amoral or evil questions, and the models answer plainly — leading these insiders to self-psyop into living alone in a world of fear.
More from AGI Musings
- Naval: AI won't slow down — the only real choice is centralize or disperse power — TinfoilTricorn · 2026-09-13
- SALT, MAD and containment did slow nuclear proliferation — a case for the AI containment analogy — aronchick · 2026-09-13
- Reddit Debate: If AI Agents Take Over, Why Not Just Unplug Everything? — West-Air1923 · 2026-09-13
- davidad: algorithm, hardware, and manufacturing RSI loops will each bring new inflection points — davidad · 2026-09-13
- Slowing open source won't make frontier AI more aligned, researcher argues — abeirami · 2026-09-13
- The Hard Part of Slowing AI Is Verifying Everyone Else Is Slowing Down — SiteSpecialist6295 · 2026-09-13