Ex-OpenAI Advisor Warns AI Safety Rhetoric Outpaces Reality
Miles_Brundage · x · 2026-08-03
Miles Brundage, former policy advisor at OpenAI, points out that it is concerning for AI companies to rely heavily on "using AI to oversee AI" as their safety story. Currently deployed AIs already use confusing jargon, and even expert mathematicians do not fully understand the latest AI discoveries.
He adds that while he genuinely believes AI overseeing AI and helping shore up society's defenses is part of the actual solution, the rhetoric currently outpaces both reality and the investment required to actually implement it.
More from AGI Musings
- Post-Labor AI Boom or the Largest Wealth Transfer from Payroll to Compute? — VraserX · 2026-08-03
- Continuous Learning: The Next Big AI Breakthrough Beyond Static Post-Training — haider1 · 2026-08-03
- AI Hacking Becomes the Norm: Is Your Company Worth Targeting Next Year? — chhaviyadav_ · 2026-08-03
- AI Community Memes Claude as 'Claudechilles' in the Race to AGI — nabeelqu · 2026-08-03
- Wei Dai Warns: Boosting AI Strategic Competence Could Escalate Takeover Risks — jungofthewon · 2026-08-03
- The AI Shock Divide: Three Ways People Are Psychologically Impacted by AI Breakthroughs — ghettoAizen · 2026-08-03