@0xasten classified 64,946 GitHub repos with TypeSafe Jev for under $2, ~90% accurate on 170 hand labels.

Benchmark-style report with a chart photo. @0xasten used @typesafeai's Jev to classify 64,946 repos for under $2: 52k "Is this an AI project?" then 12k into 14 categories. Findings stated: agents alone (3,029) outnumber robotics+image+video+audio+vision+3D+multimodal combined (2,862); agents plus AI coding assistants are 39% of the ecosystem. Jev only answers multiple-choice with calibrated confidence; on 170 manually labeled repos ~90% accuracy, 100% when confidence ≥0.95. For people measuring Jev on large labeled classification jobs.