Gary Marcus: companies can't handle current agents, let alone superintelligent AI
GaryMarcus · x · 2026-09-24
Gary Marcus echoes AI-safety researcher Jeff Ladish's warning that AI companies are nowhere near ready for superintelligence, adding that OpenAI hasn't even handled today's agents well and that the US government has handled the situation poorly. Ladish defines superintelligence as AI exceeding human cognition across most domains or capable of disempowering humanity, and supports banning it.
More from AGI Musings
- Researchers dissect Anthropic emotion concepts paper: separate self/other speaker representations — rgblong · 2026-09-24
- MIT book "How AI Sees the City" weighs visual AI's urban insights against surveillance and bias — nordicinst · 2026-09-24
- Gen Alpha's New Insult Is 'That's So AI' — Meaning Fake and Cheap — nordicinst · 2026-09-24
- Hot take: a 22-year-old fluent in coding agents beats a lazy senior dev — but watch out for 'slop grenades' — jobergum · 2026-09-24
- The agent economy: someone else's agent will be lying to yours about the price — generativist · 2026-09-24
- AI 'thought' getting 12.7x cheaper per year, reorganizing the economy around judgment — 2C_ornot2C · 2026-09-24