Why Amodei and Altman fear AI risk while Zuckerberg and Huang dismiss it: a decade of thinking apart

trevposts · x · 2026-09-22

The author analyzes the split among AI leaders over slowing down to avoid existential risk. Amodei, Altman, Hassabis, and Musk have (inconsistently) urged caution; Zuckerberg, Huang, and most VCs call it nonsense.

The author rebuts both cynical readings: if risk talk were hype, Zuckerberg and Huang wouldn't dismiss it; if it were regulatory capture, non-frontier figures like Hassabis and Musk wouldn't engage. The real difference: the first group has seriously considered these risks for over a decade—concerns dating back to Turing—while the second has engaged only superficially and resents that the narrative could dampen boom-era profits.

Conclusion: absent a solution to aligning powerful agents with human flourishing—which labs are visibly failing to do—they should simply not build superintelligence with anything like current techniques or understanding.

Original post →

More from AGI Musings

AGI Musings channel →