Poll: 127 self-identified AI safety researchers answer yes to sadism question
mimi10v3 · x · 2026-10-02
mimi10v3 contextualizes aella's viral poll: of 244 respondents identifying as AI safety researchers, 127 answered yes to having sadistic tendencies vs 117 no. Citing ChatGPT's estimate of 130-280 Bay Area AI safety researchers in 2022 (350-600 worldwide), the author argues a significant share of the field likely voted, with a slight majority saying yes — while acknowledging trollish or misclicked answers. Part of an ongoing community debate linking cnc culture and AI safety.
More from AGI Musings
- AI Researcher: Google Lacks the Independent Governance Needed to Safely Develop ASI — BlackHC · 2026-10-02
- Luiza Jarovsky: AI Firms Are Pushing Society to Accept Unacceptable Risk for a Non-Essential Tech — LuizaJarovsky · 2026-10-02
- antirez amplifies take: small LLM-armed teams will beat big collaborative projects — antirez · 2026-10-02
- NYT Opinion: Stop Talking About A.I. Like a Human — nytopinion · 2026-10-02
- Eric Schmidt: If I could invest in only one AI theme, it would be long reasoning — rohanpaul_ai · 2026-10-02
- The 'Waymo effect': how frictionless AI is quietly making research less collaborative — steipete · 2026-10-02