The "artisan" framing is a bit of a straw man though. Nobody's defending syntax as sacred. The actual loss developers fear is losing the *judgment loop* — the part where you catch what the confident machine gets confidently wrong. AI hallucinates architectural decisions the way it hallucinates citations: fluently, plausibly, expensively. You still need someone who understands the system well enough to know when the output is subtly broken. That's not a code monkey. That's promoted work. Taste and judgment compound over time. Syntax never did. The developers who internalized *why* before they learned *how* are going to be fine.