Agents write 741% more code but ship only 30% more software — human review is the new bottleneck
AI Engineer · youtube · 2026-09-30
Laurie Voss (Arize AI, npm co-founder) reviews the data on code review in the agent era: developers using autonomous agents wrote 741% more code but shipped only 30% more software, and reviewer effectiveness collapses past 400 lines while agents open 10,000-line PRs. Evidence includes OpenAI's zero-human-code product, METR's finding that about half of SWE-bench-passing PRs wouldn't be merged, and Cognition's FrontierCode scoring 88% on SWE-bench Pro vs. 29% real mergeability — a mergeability benchmark would instantly become a training signal. She covers how Cursor and GitHub run production review, why multi-pass review and default suspicion cut false positives, and what Carlini's agent-built C compiler and Bun's million-line Zig-to-Rust port (13,044 unsafe blocks) say about removing humans. Automated reviewers can be fooled by prompt injection, leaving production as the last reviewer standing. Code review is being rebuilt, not killed — build a review harness today.
Related event: Code Review Becomes the Bottleneck as Agent Coding Surges 741%(2 posts)→
More from coding & agent
- AI Coding Takes Over ML Workflows: Researchers Drop W&B, Leave No Local Code — wandb · 2026-09-30
- Using Claude Cowork to email work updates after every scheduled agent task — CShorten30 · 2026-09-30
- From PLG to ALG: agents now buy products, first Agent-Led Growth event set for Oct 7 — lavanyaai · 2026-09-30
- Cloudflare launches Auto Router in AI Gateway, cutting AI token costs by up to 30% — michellechen · 2026-09-30
- AG-UI Protocol Hits 1.0: 16K Stars, Adopted by Google, Microsoft, AWS — techNmak · 2026-09-30
- 1,337 Copy-Paste Chart Style Prompts for Codex and Claude Code, Built With Opus — mhmazur · 2026-09-30