Your arguments are actually confirmed by many researches: people with low competence in some area solve the problem with AI significantly faster than deducing the solution themselves. However, there is also a catch: people relying on LLM too much eventually fall behind in learning.
However, often time people start talking about something like 100% increase in performance — I just don't see it. Don't get me wrong: I totally enjoy AI making dirty work for me, but it's not something magical.
However, often time people start talking about something like 100% increase in performance — I just don't see it. Don't get me wrong: I totally enjoy AI making dirty work for me, but it's not something magical.