The biggest mistake, AI or not, is dropping a 10K+ PR. 300~500 LOC is how far one should be going, unless they're doing some automated refactoring. E.g. formatting the entire StaticCompiler.jl source. This should've been a distinct PR, preferably by a maintainer.
I've seen this in other places as well.
The bottleneck is not coding or creating a PR, the bottleneck is the review.
This ought to be automated using AI.
It could first judge whether the PR is frivolous, then try to review it, then flag a human if necessary.
The problem is that Github, or whatever system hosts the process, should actively prevent projects from being DDOS-ed with PR reviews since using AI costs real money.
> This ought to be automated using AI.
When the world is telling you to fucking stop, maybe take a moment and listen.
It's been stated like a consultant giving architectural advice. The problem is that it is socially acceptable to use llms for absolutely anything and also in bulk. Before, you strove to live up to your own standards and people valued authenticity. Now it seems like we are all striving for the holy grail of conventional software engineering: The Average.
I mean this with all sincerity, try doing this yourself.
The established projects are resistant to YOLOing their projects and running them on complete LLM autopilot.
You are proposing a completely different development style.
Fork Ocaml to Oca-LLM and Julia to Jul-AI and see how it goes.
I'm not trying to say that this is now projects ought to work right now.
I do think this is where we are heading, though.
No, existing open source projects are not ready for this and likely won't ever be.
It will start in the corporate world and maybe already has.
> This ought to be automated using AI.
...
> I'm not trying to say that this is now projects ought to work right now.
which is it?
It's both but the focus is on the future.
I agree with you.