Do the simplest thing that could possibly work

Honestly, this is how all computing works.

For example, chips just barely work.

If they work too well, you could shrink the chip until it barely works making it cheaper or faster or use less power.

That said, although this exercise is kind of interesting - like playing jenga - it might not be fun or satisfying.

better faster cheaper - sometimes you need to choose better.

> For example, chips just barely work.

This seems a bit overstated (except perhaps for certain recent Intel fabs) considering that they do billions of operations per second and a single error could completely invalidate a complex system in the worst case.