Researcher's field note: AI aids ~60% of research tasks, but oversight is constant and critical

fredahshi · x · 2026-11-16

Researcher fredahshi shared first-hand experience using GPT and Claude for AI-aided research and engineering: the models are "super good" for roughly 60% of her tasks, but results can go badly off track if she misses anything critical in her oversight — and such oversight is needed "all the time." A practitioner's candid assessment of real usability rates and why human supervision remains essential in AI-assisted research.

Related event: Researcher finds AI boosts 60% of tasks but oversight remains essential(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →