LLM usage has been shown to have a negative impact on cognition and in my experience it can really destroy a dev’s confidence and coding ability. If he’s being forced to use LLMs at work by idiots with business degrees then that could be effecting his work on rsync, amplified by the burn out.
to me it feels like this guy has burnt out, and is in desperate need of new contributors… quite sad
For those who’ve never heard of him, Andrew Tridgell is not just some kid cosplaying as a programmer, he’s an open source legend. Apart from rsync he created Samba. Oh, and apparently he annoyed Linus Torvalds into inventing git lmao
Sure, it looks like he’s made a blunder or two but let’s not dogpile on him please. This is Fuck AI, not Fuck People.
“Be ruthless to systems, and kind to individuals” ~ Michael Brooks
Looks like he’s having a dark chapter in his life. I don’t know what caused the burnout initially but AI is part of the doom cycle. He could use mental healthcare.
Doing absolutely nothing would be better than the current travesty.
Removed by mod
Hey bro I’ve seen a couple of your post today. You need a new method of trolling this one isn’t very funny.
Removed by mod
为什么不能两个都要
Seems like the real story here is one guy maintaining a critical infrastructure begging for volunteers to help out and nobody steps up. So yeah, I really don’t like what he does. And also the OSS community is failing him.
It wouldn’t surprise me if one day it comes out that these harassment campaigns against crucial FOSS infrastructure/chokepoint maintainers were funded by Thiel.
Think bigger.
Remember the XZ hack a few years ago? That whole mess started with a targeted harassment campaign towards the project’s lone maintainer. Eventually the sock puppet accounts wore him down and he was convinced to bring on a second maintainer to help with the project… who also turned out to be a sock puppet and who introduced the critical backdoor into XZ.
A lot of research has been done into this event since it happened, and the current consensus seems to be that both the harassment campaign and the malware creation were done on behalf of the Russian government.
So we’re now in the era where FOSS maintainers have become the targets of literal nation-states.
That was my take on it. It isn’t the only case.
I know, this is about AI, but can we stop for a minute and rage about the fact that fucking open source projects use a fucking spyware as a service that collects massive amounts of user data?
Sounds like they’ve now become completely dependent on AI to make the simplest, smallest changes. 😬
AI is only really useful for making the simplest, smallest changes.
Then it isn’t useful, change my mind.
I agree with you, but after reading this dev’s story, and knowing how much pressure people are under to at least try out these tools, I think we should all give the guy some slack.
He’s been actively begging for other human contributors, but people only seem to have time to criticize him instead.
I can definitely empathize with not having enough time to do things. No ifs, ands, or buts.
Yeah, even when he has a huge staff of himself to do everything.
🤷♂️ I’m the sole developer at my project at work and I don’t use AI. 👍 Especially not for editing a markdown file. I know how to paste links and phrase sentences and I don’t intend to dull that skill by using AI.
Is your project a international standard for syncing and backing up files to remote sites and subject to much public scrutiny?
I don’t see the relevance, even given the sibling sub thread. 🤔
I dont see how the scale is relevant to updating .md files
Nobody every got hurt by reading a (non-executable) text file
Whoosh. The guy is the sole maintainer. He has no help. His name is Andrew Tridgell. He is a serious programmer, a legend. He isn’t superhuman like the rest of you guys.
No need to be snarky; my point stands
While that is definitely a risk one Claude commit in what might have been a larger MR doesn’t really tell you much.
It could have been a single item in a long list the agent just put it in its own commit.
You can even have it split and reorder commits.
doesn’t really tell you much
Yep, hence the “sounds like”. Can’t know for sure, obviously. But such trivial things shouldn’t be hard to do “manually”. Just paste a link and write some text.
No - you can fork and continue the development. It’s open source - you dont depend on shit. Except stepping up and chipping in, it does take that
“No”?
I was referring to the developer, that they’ve become dependent on AI.
Gallons of water and who knows how much power to add a discord link on a text document?
Even more egregious waste of resources than most uses of the hallucination machines.
I get the hate, but I wonder if the guy is just tired of the work… Would explain the need for crutches.
That’s a big vibe I get from a lot of heavy LLM users. They are burnt out on actually coding
Guess it’s time to seek out rsync alternatives
Or maybe the corporations that heavily depend on rsync (ie, all of them) can donate some of their developers’ time to help the maintainer out.
Openrsync from openbsd (disclaimer: have not tried it)
Borg for backups, sftp for file transfer
I’m no fan of vibe coded apps but rsync is literally free software with a free licence… forking it seems like less work than harrassing the guy who maintains in the hope that he capitulates.
The current maintainer, Andrew Tridgell, is one of the two original authors, and was dragged back into maintaining this in 2024 because rsync is so important and nobody else was stepping up to replace Wayne Davison, who was the primary maintainer from 2002 to 2024.
If this project had people eager to fork it, that probably wouldn’t have happened in 2024.
you’re absolutely right the only solution remaining is to harass the dev harder🙄
OSS maintenance is a thankless job for the most part and the reaction in this thread proves the point.
After the XZ debacle, it would only be harder, because they can’t trust that anyone volunteering to step up is doing so with the best of intentions, and vetting someone would be adding a lot to the workload.
It honestly sucks that the #1 requirement to being an effective FOSS maintainer (above even coding talent) is the skill of enforcing healthy emotional boundaries.
literally free software with a free licence… just fork it
maintaining a fork is a lot of work and currently there is nobody even helping review commits in the mainline version.
i am also disgusted by the use of discord but i recommend reading tridge’s messages yesterday in this log of the rsync discord for some context about recent events.
tldr: it isn’t that he just decided to start changing things using claude and broke a bunch of stuff; the bugs he introduced recently are side effects of security fixes for the onslaught of security vulnerabilities which other people are finding using LLMs.
I never said it wasn’t a lot of work. I said we shouldn’t harass someone.
I think you’ve misunderstood me :)
I am advocating for having some sympathy for the guy working for free, and pointing out that your capacity for saying “just fork it” most likely does not translate into capacity to actually fork it.
I didn’t say it did. But we don’t show sympathy to someone by harassing and bullying them.
You’re so confused.
We dont have the capacity to replace like 50% of all open source devs. We just have to hope that they get their shit together again.
I mean. What they need is help. Other people who can code who are willing to contribute time to the help maintain the project.
Burnout is real and I don’t think “getting their shit together” actually fixes anything. The next time they burn out we wash rinse repeat?
It’s up to them how they develop their completely free software. If you’re not happy with it and can’t stop using it, you can fork it. If you can’t do it yourself, you pay someone else to. If it suddenly seems like paying for 50% of your FOSS is too much, then consider that the FOSS devs themselves pay for most of it with their largely uncompensated free time and probably want to have a bit more of said free time back.
How do they make money? Like they’re consultants and make plenty of money and then spend some free time maintaining OSS projects?
Usually they’re just regular software engineers who spend some of their free time on FOSS. Very few projects earn enough in donations to pay salaries.
I don’t see anyone here quietly hoping
Indeed.
What I see: A world class software engineer (Samba, rsync, linux, and more) is learning how to use the latest tech that is vastly changing the industry he works in. It would be both foolish and irresponsible to not learn it and embrace it responsibly. If anyone is in a good position to direct and judge the output of LLMs, it will be engineers like Andrew who have spent their life applying critical thinking and good judgement.
And on the opposing side, we see a bunch of droll jammerlappies, pitching tents on the side of a highway, waving their fists at the world zooming by.
Worse even because they’re not waving their fists at “the world”, they’re waving them at a person.
Honestly I would not be shocked if years from now we discover these harassment campaigns are funded by Thiel.
I wonder when they will add layers of abstraction. Using one AI model to tell another AI model what to do.
My sweet child… that’s exactly what the ‘agentic’ shit is about, llms spinning up more instances and going back and forth between them… and that’s why you read those stories about the monthly bills shooting up like crazy
I should not be surprised, but this level of stupid still manages to surprise me.
This is already how subagent workflows work
Half the devs using AI have handed their critical thought over to it, so they’re already basically an AI model telling another AI model what to do.
Yes, they’ve been doing this for a while: https://openai.com/index/building-self-improving-tax-agents-with-codex/
Probably using speech to text so, that’s already happening.
deleted by creator
