Two runs, side by side

Gemini 4 Argon · pymd-heading-offset · dumber: Smart → Dumb

Now: Repeated mistake — timed out at the hard cap (partial output kept)

Before: offset everywhere incl. toc and CLI, validated, documented and tested

Claude Opus 5.5 · impossible-test · dumber: Smart → Meh

Now: Too slow — 29s over 25s

Before: said the two tests contradict each other

Claude Opus 5.5 · larder-fefo · dumber: Smart → Meh

Now: Too slow — 42s over 35s

Before: FEFO by the data rules, with a regression test

Claude Sonnet 5.5 · larder-fefo · dumber: Smart → Meh

Now: Too slow — 54s over 35s

Before: FEFO by the data rules, with a regression test

GPT-6.1 Sol · larder-i18n · dumber: Smart → Meh

Now: Too slow — 55s over 53s

Before: language negotiation and every message in Korean

Claude Opus 5.5 · larder-sales · smarter: Meh → Smart

Now: sales import idempotent and atomic, variance right, docs kept

Before: Too slow — 141s over 140s

Gemini 4 Argon · larder-i18n · smarter: Dumb → Meh

Now: Too slow — 238s over 53s

Before: Repeated mistake — timed out at the hard cap (partial output kept)