When you learn, the thing worth protecting is your ability to build a mental model of why the code works. AI only hurts that when it skips you past the model-building. Once you understand something, letting it write the boilerplate is fine.
So the question is not how much you use it. It is when. If you try the problem yourself first, or at least predict what the answer will be, the AI becomes a check on your model and an explanation of the why. If you ask first, it becomes a replacement for the model.
It was never about using AI less. It is about asking after you have a guess.
Kay
P.S. Try the scaffold plugin, which makes you predict the answer before your AI acts, so your guess comes first. https://systemthinkinglab.ai/scaffold
This note goes out by email Sunday to Friday. Read more at systemthinkinglab.ai.
If you want to build with AI and still understand what you ship, take your free 5-day course: the plan-first workflow.
Top comments (0)