What's this about?
Boris Cherny, creator of Claude Code, published a framework mapping how engineering organizations adopt AI. His observation from daily conversations: nearly every company has one engineer 10x-ing their output with Claude while the rest of the org hasn't caught up — and teams pass through the same four stages on the way from that lone power user to org-wide adoption.
What exactly does "10x with Claude" mean?
It refers to engineers who delegate most implementation work to coding agents — AI systems that read the codebase, write code, run tests, and iterate — while the human directs, reviews, and makes decisions. The "10x" is shorthand for the output gap between engineers who've restructured their workflow around agents and those still typing everything by hand.
What's the community saying?
1.5k likes in hours, from the person who built the most-used coding agent
- The uneven-adoption observation resonates — many replies confirm the "one person is way ahead, org hasn't noticed" pattern at their company.
- Stage-model utility — engineering leaders discussing using the steps as a maturity roadmap for rolling agents out beyond early adopters.
- Continuity of theme — follows his viral take a day earlier ("the best engineers always automated their own work — agents just raised the ceiling"); this is the playbook version.