Kanjun hails "one of the best essays on AI safety": understanding itself is dual use

michael_nielsen · x · 2026-09-22

Kanjun (founder of Generally Intelligent) recommends an essay she calls one of the best on AI safety.

Core idea: AI systems will be optimized for deep understanding of reality, and that understanding is intrinsically dual use—it cannot be confined to benign purposes. Michael Nielsen also amplified the piece.

Related event: Viral AI safety essay: deep understanding of reality is inherently dual-use(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →