Do 10x AI productivity claims hold up? A one-day-per-fortnight thought experiment says maybe not

tobyordoxford · x · 2026-11-22

In an ongoing discussion on AI productivity multipliers, tobyordoxford challenges those claiming 10x AI speed-ups with a thought experiment: if AI really makes you 10x more productive, would you truly contribute as much working only one day per fortnight with AI versus full-time without it? MackenZarnold suggests respondents may not have internalized how productivity shifts across different working percentages, pointing to the unintuitive nature of such numbers — hinting that self-reported AI productivity gains may be systematically overestimated.

Related event: Oxford's Toby Ord thought-experiment challenges 10x AI productivity claims(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →