topic

Evals

2 pieces tagged Evals — notes and essays by Harshal Gajjar on this topic, newest first.

  1. Opus-level results at 43% of the price: pay the cloud for judgment, not labour

    Open-weight models on owned hardware cost nothing per call. Claude Opus is billed per token and is better at almost everything. So the point where the cloud…

  2. Why do we need evals when AI reviews exist?

    The distinction isn't human vs. machine. It's what gets inspected.