Nina Schick: If you believe in AI safety, go long on AI compute

NinaDSchick · x · 2026-09-24

Nina Schick offers a contrarian take: if you truly believe in AI safety, you should bet long on AI infrastructure and compute, because curtailing AI misuse will require a lot more AI — and the defensive side will likely remain predominantly human-led.

Her argument: build for capability, which delivers both innovation and security, rather than relying on "vapid unsubstantiated statements about certain existential death." A direct rebuttal to doom-flavored safety narratives.

Original post →

More from AGI Musings

AGI Musings channel →