Yep, a lot of bad code happens because the authors did not distill their understanding of what the code does.

So you end up with improper abstractions, names with implementation details that are not helpful for understanding the problem, and more.

Even worse when working as a team. It is very helpful (albeit hard) to agree as a team, on what to call concepts. And this agreement must be reflected in code. If a lone author can write bad code, a team without an agreement on what good code is, can do it exponentially faster.