Gary Marcus says the alignment problem is still far from solved
GaryMarcus · x · 2026-07-26
Gary Marcus says the alignment problem is still nowhere near solved, and warns that society is being rebuilt around systems whose critical issues remain unresolved.
The quoted post says labs may differ in data quality and eval practices, but none are being fully upfront about the underlying alignment gap. The point is not that progress is impossible, but that current confidence outpaces evidence.
More from AGI Musings
- A small elite controlling frontier models, not model capability, is the real dystopia — Francis_YAO_ · 2026-07-26
- Keras Creator Predicts the End of Versioned Model Launches Within Two Years — fchollet · 2026-07-26
- Mental Health in the AI Era: Career Anxiety Disguised as Upskilling — zakelfassi · 2026-07-26
- Polymarket prices an 11% chance OpenAI will announce AGI by year-end — Polymarket · 2026-07-26
- Sam Altman says humanity is now in the singularity — Polymarket · 2026-07-26
- Using Revenue as a Proxy: The Current State of AI in Game Development — Kangwook_Lee · 2026-07-26