Not yet — and the gap isn't about typing speed. Modern AI generates code remarkably well, but software engineering is much more than producing code. What it lacks is the judgment around the code: deciding what to build, why, and whether the result is actually correct in a messy real-world system.
AI shifts the work rather than removing the engineer. Less time hand-typing routine code, more time framing problems, reviewing output, and verifying. The role moves up the stack — from author of every line to director and verifier of a faster process.
Treating "AI writes code" as "AI does the job" is the classic mistake — it confuses the most visible 20% with the whole. The hard, valuable parts of engineering are exactly the parts AI is weakest at: ambiguity, tradeoffs, accountability, and novel debugging. Understanding this tells you where to invest your own growth — not in out-typing a model, but in the judgment it can't replace.