well it turns out that people think I'm an asshole for pushing back on AI slop so here goes some more hours wasted on this.

The WHOLE POINT is that we're wasting extremely limited time on this shit.

Maintainership is a fucking thankless thing.

You do the right thing but are 'nice' when you shouldn't be and it causes tonnes of problems maybe people don't accuse you of anything...

You be 'mean' and push back when it's RIGHT to do so, and wow you're a bad person.

You can't win really

https://lore.kernel.org/all/ah887A5VkXOcmq-g@lucifer/

@ljs Tbh someone new to a free software project should start with small stuff even if they are experienced. It is about generating trust and understanding the process.

The series is likely LLM generated but even if it wasn't this isn't they way to interact as a newbie. Start small, join conversations, post an RFC with your idea to start discussions and then you post the series.

That way people will trust you much more.

@ljs and then we have the problem of being "bad" due to pushing back to vibecoded stuff. I agree with you here and I understand the pain.

@ljs you need to have a lobster dunk on ai slop

@ljs I really don't understand this at all. Even if we admit it's not AI generated in one way or another like who could defend posting this as a newcomer, without any prior discussion and on such a complex topic. Pushing back on this seems to me like the normal kernel experience, and you did that in a nice way too. Oh my.

@ljs Your reply in that thread was thoughtful, clear, carefully positioned, non-accusatory, and realistic. Some people may not have *liked* it, but it was a good reply, IMHO!

@neil thanks, I think the objection was more so with https://lore.kernel.org/all/ah71gjj21rLIuTnb@lucifer/ where I was a little more strident but given the context I think understandable + reasonable...

I think we really have to contend with the deep imbalance between work generated vs. maintainer resource and some people are still in denial about the scale of the problem.

@ljs

> I think we really have to contend with the deep imbalance between work generated vs. maintainer resource

Yes!

@ljs @neil from a great talk yesterday

@vaurora @neil yeah saw you post about it yesterday, spoke to me!

Are there slides/a vid anywhere?

@vaurora @ljs @neil great slides. the real bottleneck in productivity is human energy, not AI speed. lots of developers are finding themselves more tired because they spend so much time to review and fix AI output, rather than focusing on creative problem-solving.

@ljs @neil FWIW from my POV as a relative outsider to "sustained software development", anyone who doesn't understand the asymmetry problem (or thinks the solution is "just review it with more word prediction machines") fundamentally misunderstands
the "asymmetry" problem itself,
the scale of the problem, and
why the problem matters.

It is totally fair to stop talking to anyone who is wasting your limited time. The person wasting your time has already broken the social contract.

@interpipes @neil the problem is people in the community might look at how I dealt with it, accuse me of being 'mean' and engaging in a witch hunt 'without proof' (you can never 100% prove it), and then I'm undermined completely.

Which is exactly what's happened here :) [in off-list comms]

@ljs @neil I think non-existent variables are pretty cast iron proof myself

that isn't a language barrier, that is a lack-of-knowledge and/or lack of care barrier

@interpipes @neil I feel there's several elements of overlapping proof there.

I didn't exactly list all in detail there as I wanted to limit my time on it but I've failed at that already :) so maybe I should have...

@ljs @neil aside all of this, if it's a dead-end area of development because of known architectural/implementation limitations then absent some actual evidence front and centre of how to convincingly address those shortcomings it doesn't really matter where the input came from, it's still not going to be a priority for anyone.

I imagine if you hadn't mentioned word-guessing-machines this wouldn't have been contentious at all.

@interpipes @neil it would have resulted in the exact back/forth I was trying to avoid, but has happened anyway.

My hope was that I'd be supported in my stance for the exact purpose of avoiding this.

But now I have to sink more time in 👀

@vaurora @ljs @neil Talk looks great, would love getting the slides if they are available.

@doragasu @ljs @neil I got the slides https://app-share.mailbox.org/appsuite/api/share/08990cad08122c5688ee76a8122c42de80866c53bb4e96f5/1/8/MTU5Nw/MTU5Ny83MDMx

@vaurora @doragasu @neil thanks!

@vaurora @doragasu @neil really great. And absolutely true.

People in the open source community really need to wake up to this.