Module 06 · What to measure · 5 min
An AI judge is useful, not an oracle
An AI judge can scale semantic evaluation, but it is just another evaluator: it has to be tested, calibrated and versioned before anyone quotes its scores.
Checking your access
For a question, an idea or a proposal, write to me here.
Module 06 · What to measure · 5 min
An AI judge can scale semantic evaluation, but it is just another evaluator: it has to be tested, calibrated and versioned before anyone quotes its scores.