Nice read!

I also believe the difference between expert level and amateur programmers will be a lot easier to spot. Especially around the software or system architecture

Due to the inherent GIGO behavior of any software, it'll be a lot easier to tell that the code or software wasn't written by someone with deep knowledge since LLMs' output is just a good reflection of your prompts and creativity while at it(prompting). For instance, a prompt from Ken Thompson can't be exactly similar to your average Joe programmer. I've seen it a lot with most new open-source projects nowadays.