How do I balance between using AI tools and developing my software engineering skills?

You should absolutely be using AI as much as possible. In my opinion, if you're not learning to code for the first time using AI rarely hurts your growth

In fact using AI has helped me learn 10x faster. So many times when I see the AI do stuff and I'm like wow I had no idea I could do X,Y,Z

In my opinion in this age it's incredibly important to know how to use AI to 10x your velocity for the simple reason that your time is limited, but the AI's time spent writing code is infinite. So knowing how to best use AI will pay off huge. I also think there's a fundamental shift people need to learn to how to best use AI.

The most common failure mode I see is people trying to get the AI to work for you instead of trying to change your workflow to work with what the AI is best at - for the simple reason that AI is getting better and better and time spent running AI has no upper bound

Given that I still think it's important to use AI to 10x your learning. When the AI outputs something make sure to ask it, why did you do that? how does that work? how does it handle XYZ use case. Dont just take the output and run with it. Actually drill deeper.

By doing this I've accelerated my learning so much

I also found taking the courses on debugging/coding and accelerating that process with AI to be super helpful. A lot of the stuff in Taro debugging is your approach. It's not exactly "hard skills" but the mentality around it. I've used those lessons from Taro to change my mindset around how I debug and instead of manually doing it I just ask an AI to do it.

The caveat I'll say is a lot of it depends on the underlying LLM. I am not sure about the quality of Meta's LLMs but I am guessing its not as good as using cursor + sota models.

Here are some excellent guides on how to get started with using AI: https://docs.cursor.com/guides/advanced/large-codebases