Today for AI
HOT RADAR
toolsHEAT 19.6°

AI CURATED HOT STORY · 11/5/2026

Anthropic Merges 30,000 Code Changes a Day by Reviewing AI Evidence

1 reports archived1 independent sourcesupdated 11/5/2026, 10:12:02
Synthesis & Latest Updates
1 Sources Cross-Validated

Anthropic's Addy Osmani says the company went from roughly 2,500 merged code changes a day in May to over 30,000 by September. The workflow has Claude exercise features end-to-end like a real user, auto-fix what breaks, and hand humans runtime evidence such as browser recordings and screenshots instead of line-by-line diffs. He argues coding is no longer the bottleneck and that solid verification, giving AI a way to check its own work, can improve output quality two to three times over.

LATEST/Anthropic's Addy Osmani says the company went from roughly 2,500 merged code changes a day in May to over 30,000 by September. The workflow has Claude exercise features end-to-end like a real user, auto-fix what breaks, and hand humans runtime evidence such as browser recordings and screenshots instead of line-by-line diffs. He argues coding is no longer the bottleneck and that solid verification, giving AI a way to check its own work, can improve output quality two to three times over.

Hourly heat curve

3 fully observed hours
Now
19.6
Peak
20.211:00
24h change
–

The line compares only sources observed throughout; gap hours are interpolated to keep the trend continuous.

Coverage timeline

Total 1 reports · Latest first
  1. 宝玉 (@dotey)T2·66 pts

    Anthropic Merges 30,000 Code Changes a Day by Reviewing AI Evidence

    • Merge volume grew more than tenfold in four months, forcing a shift from reading diffs to reviewing runtime evidence
    • Team taste is codified in AI-facing files such as CLAUDE.md and AGENTS.md, with hooks auto-checking every change
    • Building a self-verification loop for AI can lift output quality two to three times; verification, not generation, is the new bottleneck