Been writing code for 15 years. Started using Copilot in Aug 2021 when it was still in beta. Something I'm noticing: Most code quality practices were designed around human memory limits. AI breaks these constraints. But here's what's fascinating - AI doesn't feel the pain of bad code. A messy 5000-line function? AI doesn't care. Spaghetti dependencies? No problem. It'll happily work in codebases that would make humans quit. This removes the natural pressure for quality. Humans refuse to work in terrible code, forcing refactors. AI just... continues. Another issue: Different developers prompt differently, creating wildly different patterns. When production breaks, someone's debugging code generated by unknown prompts with no thought process to follow. We've become "code auditors" not authors. Yet we still optimize for human writing, not human debugging. I'm starting to wonder if we'll need to track prompts with commits, like we track commit messages today. Or design systems assuming the debugger never wrote any of the code.