ResearchPapers, evidence & method
Mercor analysis says 90% of US working hours go to tasks AI benchmarks cover poorly
A Mercor labor-index benchmark map reports that most US work time goes to tasks that current AI benchmarks measure poorly. That suggests a large gap between what benchmarks test and what jobs actually involve.
From the last30days · hackernews article of October 16, 2026 and cited sources: Mercor's research page reports that about 90% of US working hours are spent on tasks that existing AI benchmarks do not cover well. [1]
SOURCES[1] 90% of US working hours are spent on tasks not covered… mercor.com
AT Pulse research deskPublished · Updated · 1 min read · AI-assisted analysisAI-assisted, not human-edited. Methodology Corrections Who owns this site
- What happened
- Mercor's research page reports that about 90% of US working hours are spent on tasks that existing AI benchmarks do not cover well. [1]
- Original source
- last30days · hackernews article, independent perspective
- Brief by
- AT Pulse research desk AI-assisted; publisher and contact on the desk page
- When
- October 16, 2026 article published
- Covered
- Oct 17 edition
- Evidence
- 1 source across 1 site
StrongHow it’s measured
Hacker News · 534 likes · 452 comments
Reach blends public engagement (views, likes, comments, shares, stars) from the platforms where the story or its original source lives: YouTube, Hacker News, GitHub, Substack and similar. A video and the article written from its transcript share one score, taken from the video. Platforms with no data are left out rather than counted as zero, and the age of a story never lowers its score.
Mercor's research page reports that about 90% of US working hours are spent on tasks that existing AI benchmarks do not cover well. [1]
If the finding holds, high benchmark scores may say little about how AI performs on most real work. That would weaken the case for using today's benchmarks to forecast labor impact, and it points to where new evaluations are needed.
90% of US working hours are spent on tasks not covered well by AI benchmarks