AI Safety Folks Were Debating Loss of Control and RSI Before NVDA Was a Gaming Company

john__allard · x · 2026-09-09

John Allard pushes back on claims that AI safety people haven't documented their fears: while disagreeing on timelines and risk probabilities is fair, few subcultures have written more in-depth about AI advancement and specific risk scenarios — with probability estimates — than AI safety folks. LessWrong posters were arguing about loss of control and recursive self-improvement back when NVDA was still a gaming company.

Related event: AI Safety Supporters Push Back: X-Risk Concerns Are Well Documented(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →