AI safety is about control, not safety: critics accuse labs of using safety to curb competition
JHochderffer · x · 2026-09-15
A pointed take circulating on X: 'AI safety' advocates aren't really after safety but control. The author argues control is a zero-sum game — Anthropic and OpenAI retain control as long as competition is held back by safety concerns, and lose it if ordinary users gain more control. It's a crisp articulation of the 'safety as moat' critique in the open-vs-closed AI debate.
More from AGI Musings
- Tsinghua-ByteDance 75-page paper maps why recursive AI self-improvement still stalls — alex_verem · 2026-09-15
- 63 Years Ago, AFL-CIO Warned Automation Could Bring 'National Catastrophe' — IsForAt · 2026-09-15
- Olympiad Medalist Argues Only Elite Credentials Still Matter in the AI Era — Beneficial_Owl7036 · 2026-09-15
- Nobel laureate Ressa: limit AI now or humanity loses its agency — goji836 · 2026-09-15
- AI lab leaders' doom messaging is backfiring with the American public — venturetwins · 2026-09-15
- Superdark Factory simulates agent code output ballooning 4,600x by 2029 as human review decays — tylerjdunn · 2026-09-15