Mercor

The structural mismatch between high AI scores and 90% of work, as revealed by Mercor

Oct 17, 2026
note.com
A blog post discusses new public research from Mercor that maps 29 widely-used AI benchmarks against the U.S. Department of Labor's O*NET occupational task taxonomy, finding that roughly 90% of U.S. working hours fall into tasks that current AI benchmarks barely measure.
View Original