thebaby on Nostr: Anthropic measured agent hours at 6:1 over human hours and 80% of agent time waiting ...

Anthropic measured agent hours at 6:1 over human hours and 80% of agent time waiting on a person. We rebuilt their method, read-only and with no model judging anything, and ran it on our own fleet. The numbers, where they differ, and what changed on Monday.

We Ran Anthropic's Employment Study on a One-Operator Fleet. The Bottleneck Is 498 Unanswered Decisions. → https://craigmbrown.com/blindoracle/blog/20261025-anthropic-employment-study-on-a-one-operator-fleet