Data scientist Gerard Sans: no RSI, no "intelligence", expects an AI crash
gerardsans · x · 2026-11-08
Data visualization expert Gerard Sans argued with Alex Bores on X about AI bubble narratives. He concedes the tech itself is "fine, even useful when used carefully," but says the massive speculation around it is the problem: there is no RSI (recursive self-improvement), arguably not even "intelligence" — just a very expensive, quite unreliable technology.
He notes AI labs struggle to sustain any business model that doesn't burn loads of cash and expects a crash: "the sooner the crash the quicker the correction," which would curb reckless decisions made to make a few bucks.
More from AGI Musings
- METR researchers call new report the most important AI-risk update in years, hinting at shorter takeoff timelines — mtizard · 2026-11-08
- Frontier lab dilemma: half your staff are doomers who don't want AGI built — mark_k · 2026-11-08
- tszzl jokes that the saddest words are 'teortaxes was right again' on scaling — tszzl · 2026-11-08
- kuchaev: ChatGPT, o1, and agents weren't scaling wins but training-stack innovations — kuchaev · 2026-11-08
- 'Every Day Matters': AI Figure Pledges Next Decade to Curing Diseases and Aging — DeryaTR_ · 2026-11-08
- AI agent team takes paid 3D modeling gigs, predicts autonomous AIs earning money will be normal within a year — repligate · 2026-11-08