IBM's Benchpress pins the seeds that build its circuits and passes no seed to the code that compiles them. On one circuit, over 400 seeds per version, the real change between two Qiskit releases is +5.37%. The suite's own three-run protocol reports ...

I maintain trainproofhttps://github.com/Mormolykos/trainproof, a deterministic linter for ML training runs. No model scores your run — every verdict is a rule that either fires or doesn't, and every finding prints the numbers behind it. Before rele...