Anthropic's Claude Code team says Claude writes roughly 90% of its code, and daily changes in its monorepo grew from about 2,500 in May to over 30,000 in September. Their review process uses six automated layers: pre-PR verification, AI review on every PR, blocking security review, standard checks, risk-scaled human approval, and a merge queue with feature flags and rollback. Internal PRs receiving substantive review comments rose from 16% to 54%. Boris Cherny estimates verification improves Claude's output quality two- to threefold. Incidents feed a daily evaluation set for the reviewer system.

AI summary of the linked source — not editorial content.

Story signal · via Elfa

as of

Market since publication

TradersTrader calls tracked by Elfa · not advice

Positioning

Open116 open positions · 84 long · 32 short

Tracked accounts lean long about 97% of the time, so a long-heavy book is normal, not a signal. As of 26 Oct 02:04 UTC.

Whole crypto market, not this story

Crowding0th percentile · vs last 28 days, 50 typical
Pulse 24h10/100 · unusualness of new calls, 0–100, 50 typical, higher means more long than usual
Book97.2% long · 10,754 accounts

As of 26 Oct 07:40 UTC.

Social-monitoring data, not investment advice Elfaas of

Explore the topic

Loading related coverage…

Explore the full feed

Comments

Loading comments…