As AI writes more and more of the code, it's quite important to maintain sufficient skill and knowledge to even review or judge AI's output.

Here is a little project I whipped up - I personally found deliberate practice quite useful. It focusses on non-coding skills (as I think LOC production is basically solved by AI, IMO). Rather I start with two tracks: general technical fluency (CS / systems / practical fundamentals), and system design.

And yes it was built primarily with help of AI tooling.

Would love to get some feedback - in particular, do others feel the same way about skill atrophy risk on-the-job nowadays, and whether this tool is helpful potentially or not. What else do folks do instead, if anything.