EdgeLog by Ecro
Essays Now About
EN / KO
← All Tags

#code-review

3 posts

  • Embedded Dev How Many AI Code Reviews Is Enough?

    400 LLM calls on one codebase. A single review pass finds 34% of the real defects, and the four most-repeated findings were all false positives.

    Aug 22, 2026 · 28 min read
  • Embedded Dev AI-Written Firmware: How to Stop AI from Breaking Firmware

    3-Tier Trust classifies PRs. Verification gates (spec, TDD, RAG, HIL, deterministic analyzer, more) take them to merge. Each catches a different defect class.

    Apr 25, 2026 · 6 min read
  • Embedded Dev AI-Written Firmware: Where It Breaks

    AI-generated firmware fails in predictable nonlocal patterns. 3 visible in diff, 3 that detonate after 49 days.

    Apr 9, 2026 · 7 min read
/SUBSCRIBE ยท EMAIL

The notebook, by email

A hardware product built solo with LLM harnesses, plus the benchmark runs and failures behind each call. Biweekly, no spam.

Unsubscribe anytime.

Orsubscribe via RSS.
EssaysAboutNow
© 2026 EdgeLog
GitHub·Email·NeuroTerm·RSS