Ask HN: Has using ChatGPT/LLM's improved or worsened your coding skills?

I think the integration of AI tools into the workflow of all white-collar jobs will be as significant a change as the advent of the PC. Because of this, the long-term effects on our competence in cognitive work is something worth consideration.

On one hand LLM's offer everyone a personal tutor with a range of knowledge far greater than any human. This could mean LLM's help humans learn faster and become smarter. However, in using them we outsource cognitive work that would otherwise be instrumental in the learning process. Our mental scaffolding atrophies and become helpless without AI tools.

For anyone who in coding, writing, or any kind of work use ChatGPT or other LLM tools, do you believe your overall skill in the realm it helps you in has improved, stayed the same, or worsened as a consequence?