Lars Faye’s essay argues that AI coding tools are quietly preventing the next generation of developers from ever becoming experts. The catch: the skills needed to manage AI coding agents — steering, auditing, verifying — are the same skills heavy agent use erodes. Veterans benefit most because experience lets them judge the output; novices are handed expert-level tools with none of the underlying judgment, a situation he calls the “expert novice.”

The evidence is mostly studies, and it’s consistent:

  • A widely cited ACM study found heavy AI users skipped crucial planning steps and finished with an “illusion of competence”; the best-performing learners mitigated or ignored the AI entirely
  • A UPenn study of 1,000 students learning math with an LLM found they scored 17% worse than the textbook-only group — while believing they were excelling
  • The same study’s “tutor mode” (ask for help, then solve it yourself) improved practice performance 127%; Anthropic’s 2026 study agrees that getting “painfully stuck” is what builds mastery

His explanation is “inverted learning”: the student steers the mentor, which falls apart exactly when you don’t know what you don’t know. The model becomes “a compass that always points north, wherever you suggest north might be.” Friction, he argues, is a feature — the applied struggle of debugging, tracing errors, and failing is what forms developer intuition, or “fingertip feeling.”

The practical prescription: use LLMs as Socratic sparring partners — interactive documentation, tutorial generators, question-and-answer — not as answer machines. His checklist separates cognitive debt (abdicating your judgment) from cognitive offloading (delegating the mechanical). If the industry keeps optimizing for lines of code generated while the expertise pipeline dries up, he warns, Sam Altman’s vision of selling intelligence “on a meter” becomes self-fulfilling — and companies mandating AI-only coding will learn their own lessons.

A useful companion to the expertise-and-AI canon — the same “you have to do the work yourself” argument as Don’t Be a Meat Proxy, aimed at the juniors being pushed through the AI pipeline today.