Boris Cherny's AI adoption model fuels the bubble
Originally posted on LinkedIn, July 17, 2026.
Unpopular opinion: Boris Cherny from Anthropic annoys me. He published a guide that maps AI adoption in organizations across five stages, from 0 to 4. It is easy to share a guide for burning as many tokens as possible while taking zero responsibility for the output when you are speaking from the ivory tower as CEO of the company that gets paid for every token.
I see this as an anti-pattern for how to work. It will take time before we can give agents complete autonomy to write production code without supervision.
These are exactly the claims that fuel a bubble. The developers doing the work are saying it does not work that way, and no investor is listening. Nothing is sexier than assigning automation levels to ai slop.