Ex-OpenAI researcher mocks labs' U-turn on recursive self-improvement

So8res · x · 2026-09-02

So8res, formerly of OpenAI's Superalignment team, mocks how quickly the industry flipped from dismissing recursive self-improvement (RSI) as "crazy sci-fi nonsense" to openly embracing it — quoting the framing "we've developed agentic AI, which is great because it allows us to accelerate recursive self-improvement."

Related event: Safety researchers blast labs' sudden embrace of recursive self-improvement(4 posts)→

Original post →

More from AGI Musings

AGI Musings channel →