Skip to content
❯
gwenneg
/
tags
/
evals
posts
tags
search
◐ auto
❯
grep -l "#evals" posts/*
#
evals
2026-08-01
Turn an AI you can’t trust into one you can
You can’t unit-test an AI. Notes from a month of building evals instead.
ai
claude-code
evals
llm
testing
← all tags
LIST
tags/evals/
1 post, 2026
/
search