For me, the exception has always been made at what happens after the first interaction on a PR:

1. Just fix it or correctly (<---- very important) argue why I am wrong - this is peak collab
2. Argue nonsense and/or get angry - this used to be about 80% 
3. The bot replies - this is nowadays 80%

If you're lucky, you have to close less than 80% of all PRs (or just wait for the bot operator to move to greener pastures where fame, hookers & dope take less effort to attain)

optimism

Even before LLMs, writing the code was not the main bottleneck for me. But writing code did take time, so a solid, working, easy-to-review PR was often worth the small extra risk and inconvenience.

With LLMs becoming quite good at implementing things, that tradeoff is almost never true anymore.

I could get huge amounts of code written, but I'm bottlenecked on:

The code in your PR doesn't help me much with any of these.

The one exception to all this is PRs from someone you trust to understand/design/review changes well. Their code still needs review, but I can review their code knowing that it's worth it. I know I won't find some naive flaw buried in layers of slop, and I can trust they'll understand and fix any flaws I do find.

devs

> Even before LLMs, writing the code was not the main bottleneck for me. But writing code did take time, so a solid, working, easy-to-review PR was often worth the small extra risk and inconvenience.
>
> With LLMs becoming quite good at implementing things, that tradeoff is almost never true anymore.

> I could get huge amounts of code written, but I'm bottlenecked on:
>
> * understanding — reading the existing code to be able to reason about it;
> * designing — coming up with the right changes and architecture;
> * reviewing — ensuring that the code is doing what I wanted.
>
> The code in your PR doesn't help me much with any of these.

The one exception to all this is PRs from someone you trust to understand/design/review changes well. Their code still needs review, but I can review their code knowing that it's worth it. I know I won't find some naive flaw buried in layers of slop, and I can trust they'll understand and fix any flaws I do find.