Gary Marcus: both 'human extinction in 5 years' and 'everything will be fine' are wrong
GaryMarcus · x · 2026-09-14
AI researcher Gary Marcus pushes back on the two dominant narratives in AI discourse: that humans will go extinct within five years, and that everything is going to be fine. He calls both equally preposterous, suggesting the reality lies somewhere in between, though the post offers no detailed argument.
More from AGI Musings
- An Open Letter to Altman and Amodei: Safety Essays Aren't Safety Leadership — AryHHAry · 2026-09-14
- AI researcher lays out the safety paradox: pausing algorithms while compute piles up maximizes risk — RichmanRonald · 2026-09-14
- Reddit proposal: a neutral 'AI Switchboard' — public rails, competitive models, no lab moat — NewYak4281 · 2026-09-14
- AI bio-risk debate: if AI can design a super-virus, another AI can design the vaccine — RichmanRonald · 2026-09-14
- AI safety commentator says intelligence can't be contained, urges hardening critical systems — gandamu_ml · 2026-09-14
- Blogger argues Altman's 'international coordination' push targets China's open-source models — ayushtweetshere · 2026-09-14