Mechanism◆Article · 26:51 — 31:31
AI-generated, non-deterministic systems require more engineering discipline, especially stronger testing, evaluations, and conformance validation.
Because engineers trust AI-written code less when they did not write it themselves, confidence must be rebuilt elsewhere in the process. Charity locates that trust in validation mechanisms such as tests, evals, and conformance testing. ✦ AI generated
Charity Majors · The Pragmatic Engineer · 2026-08-12 · original ↗
Non-deterministic systems require more engineering discipline versus before. With code written by AI, we’re reducing the trust in the code (because we no longer wrote it), so we need to increase trust at the other part of the development process. Specifically, at validation: with things like tests, evals, and conformance testing.
Read full article ↗excerpt · fair-use quotation
- ·Non-deterministic systems require more discipline than before.
- ·AI-written code reduces engineers’ trust in the code itself.
- ·Trust must be rebuilt elsewhere in the development process.
- ·Increase confidence through stronger validation mechanisms.
- ·Use tests, evaluations, and conformance testing.
- ·Validation helps compensate when engineers did not write the code.
Around this claim