The AI productivity trap: Why your best engineers are getting slower

But it would inevitably miss a crucial, unstated context — perhaps that empty miles are billed at a different rate only when the carrier is part of a specific dedicated program. Finding that omission in 100 lines of AI-generated code is exponentially harder than writing the 100 lines myself. I spend more time debunking the AI’s plausible hallucinations than I would have spent just doing the work.

Context rot and flow state

There is also the cost of context switching. Deep work, or flow state, is the essence of high-level engineering. It takes time to load the context of a distributed system into your brain.

AI tools, in their current chat-based forms, encourage a fragmented workflow. You stop coding, you prompt the bot, you wait, you review, you reject, you re-prompt. The flow is gone.