natewizz / nkcom Personal
tags/process/ 2 files
Name Message Updated
evals-are-the-new-unit-tests.md Tweaking system prompts at 2 AM without a benchmark harness is just superstitious typing. Here is how to build deterministic eval suites that keep AI pipelines honest.
one-page-or-it-doesnt-start.md An internal project or tool that cannot be described on one page is not ready to be built.