ReviewDebt: Calculating AI PR Review Costs
AI Engineer · youtube · 2026-07-12
This talk introduces **ReviewDebt**, a framework scoring every pull request to measure the "hidden review burden" introduced by AI-generated code. The author argues that standard metrics like PR count, line count, cycle time, or developer satisfaction miss the real costs: diff bloat, insufficient testing, unclear rationale, ownership spread, and the extra time humans spend verifying AI code. The framework generates deterministic scores based on signals like: - diff size - test coverage changes - ownership spread - generated code traces - evidence and reasoning gaps The presentation will compare three real PRs: a clean human PR, a high-debt AI PR, and a refactored AI PR. It will also use 90 days of data from a production backend team to demonstrate how review debt scales alongside an increasing proportion of AI PRs.
More from coding & agent
- AI makes software easier to build, but it also lowers the floor on quality — paw_lean · 2026-07-21
- Fable coding run costs $6.69 for 67 lines of code in a 4-minute job — bytebot · 2026-07-21
- Cursor writes better code, but ChatGPT can still control the computer — vista8 · 2026-07-21
- Agents can remember facts, but still forget how to do the job — No_Advertising2536 · 2026-07-21
- Agent skills for project downgrade and troubleshooting tested in CLAD on LS 5.22 — stspanho · 2026-07-21
- Open-source B-roll skill turns scripts into 5-second vertical clips with Codex and Gemini — yangyi · 2026-07-21