Frontier lab insiders say voicing AI safety skepticism is socially costly

ivan_bezdomny · x · 2026-09-11

Andrew Ho argues frontier labs have a strong social desirability bias: politely reasoned skepticism about AI safety makes you unpopular internally, and with 7-8 figure compensation there's little incentive to make enemies. The observation echoes Matt Zeitlin's note that outsiders (Meta, VCs) are the most skeptical of 'AI Safety.'

Related event: Insiders Say Questioning AI Safety Is Unwelcome Inside Frontier Labs(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →