Anecdote◆Article · 50 words
Shipping AI-written code without human review leads to production disaster within months.
Dex Horthy · The Pragmatic Engineer
MechanismArticle
The key is a clear risk model that separates changes that can skip human review from those that can't, scored across six dimensions.
Each PR is scored on size, blast radius, reversibility, data/security, operational impact, and test completion; under 24 points auto-clears, over 64 goes to a human. ✦ AI generated
Claire · Lenny's Newsletter · 2026-08-10 · original ↗
The key is having a clear way to separate changes that can skip human review from ones that can't. The risk model used here scores each PR across six dimensions: the size of the change, its blast radius, how easily it can be reversed, its data and security implications, its operational impact, and whether tests and CI have actually been completed. Anything below 24 points is classified as low risk and cleared by the agent; anything above 64 goes straight to a human for review. The exact numbers matter less than turning a vague judgment call into a repeatable system.
Read full article ↗excerpt · fair-use quotation
Related moments
Mechanism◆Video · 14:32 · 2m
In a world where AI agents operate across multiple systems, organizations need more systems thinkers who build common infrastructure and 'paved paths,' rather than relying on locally-built, siloed solutions.
Elizabeth Stone · Lenny's Podcast
Claim◆Article · 48 words
OpenAI is deliberately designing ChatGPT Work to avoid the excessive refusals that hobble many general knowledge-work AI agents, while still requiring permission checks.
Andrew Ambrosino · Big Technology