Will anybody tell Linux Kernel, XFCE, TDE, and the other slop-slingers?

[tubsta@social.bsdlab.au]

@downey Oh FFS, XFCE too. It doesn't even have enough change to warrant it. Does computing want me to go back to terminal life like the 80s and 90s (though, they were the good days, unless you had an Amiga, that was great!)??

[librarysquirrel@sunny.garden]

@downey I like this comment, illustrates how LLMs get you by seeming to help with irrelevant futile non-tasks and then they mess your shit up when actually trying to make any real use of their purported capabilities. Worse than useless, ofc yet to wholly dawn on this particular fool in the quote.

quote

I was one of the developers who wrote to you about how much I love building things with Claude Code.
That was the honeymoon phase.
I was mostly building my own project (green field) and I used it for my angular frontend as well…
It’s pretty good at generating the first features fast. And it sounds reasonable when you brainstorm ideas with it.
It’s also pretty good with angular frontends when components are not too complex.

But I have since used it for production code of our very very complex reinsurance contract software.
It fucked up every time.
Our codebase is just too complicated and messy, it seems.
The reasoning for changes sounded very reasonable and the code does look good as well.
But I have had to go back and redo most tickets I used it for.
And looking at the actual changes after taking the time to deeply understand the logic I was changing made me embarrassed.

end quote

[phaedral@beige.party]

@kp @janeishly @downey You are not crazy and you are not alone.

[cyberquixote@infosec.exchange]

@janeishly @downey

I dunno if it’s an intelligence issue as much as it is a mindset and awareness issue.

I suspect that falls under wisdom as opposed to raw intellect. But I am not a professional on this topic…

[autonomousapps@mstdn.social]

@kp @janeishly @downey I actually think you're missing a really obvious explanation: most if not all the tech people who are enthusiastically all-in on this shit train also *love* generating "art" and also mostly suck at writing. There's no contradiction

[arestelle@dice.camp]

@janeishly @downey people saying that about translation should be required to read Hofstadter's Le ton beau de Marot like 10 times

(everyone else should read it once, just because it's good)

[daemonspudguy@app.wafrn.net]

@jackpearse@nrw.social @downey@floss.social

Unfortunately for the slop-slingers of the world, the slop-slingers of the world don't have rights I am required to respect so that's going to be a no from me.

[endlessmason@hachyderm.io]

@womble @downey
Tests are especially annoying when you don't have them and the guy on the team who wrote the code is either on vacation or is smug about how obvious the bug is and how easy it should be to fix even though he shipped the bug in the first place and didn't document anything

[ebassi@mastodon.social]

@ell1e @downey @davidgerard the GTK contribution guide bans AI slop, and requires disclosure. We do not encourage LLM use, and we are extra careful on submissions aided by LLMs. We don’t ban them outright, that’s true; that was the compromise position. Now that it’s out, I’d probably push for something like the Rust LLM rules.

[ebassi@mastodon.social]

@ell1e @downey @davidgerard and yes: GNOME has not adopted a blanket LLM policy because it currently lacks the mechanism for that kind of project-wide decisions. It’ll likely be a good test of the RFC process we are adopting.

[christosargyrop@mast.hpc.social]

@womble @downey there are certain stuff they are good for eg writing glue or interface code with existing libraries and others they are not. If you have ever used SWIG, writing the interface files and creating the documentation is a pretty solid use case

[womble@infosec.exchange]

@ChristosArgyrop @downey I have no desire to replace swig with a non-deterministic alternative that consumes orders of magnitude more resources to produce an inferior result.

[janeishly@beige.party]

@cyberquixote @downey Ha! exactly!

[christosargyrop@mast.hpc.social]

@womble @downey I did not say you should replace swig, on the contrary, just automate the .i file writing

[bms@mastodon.bsd.cafe]

@phaedral @kp @janeishly @downey There is a faction in Silicon Valley, which might even be termed a sect (giving rise to the other "sects") who are notorious for asserting that epistemological expansionism ("We do it because we can") is desirable, as is accelerationism. But to some, this is immanentizing the Eschaton ("bringing about the end of the world"). It's partly what Neil Postman was driving at: https://en.wikipedia.org/wiki/Technopoly @kimcrawley @davidgerard @zzt @tante @tforcworc

[pndc@social.treehouse.systems]

@jackpearse @downey https://en.wikipedia.org/wiki/Tone_policing

[pensionien@todon.nl]

@kp @janeishly @downey
I'm not to happy about them creating virusses as well.
And of course, when I say 'not to happy' what I actually mean is:
WTF! Are you out of your f*cking minds!!!!!

[jackpearse@nrw.social]

@davidgerard @downey

Yes it works. That's exactly is how the ipen source community is losing former developers. I find the way people are talking in the FOSS community right now disgusting.

[davidgerard@circumstances.run]

@jackpearse @downey these are sloppers LARPing as developers

[jackpearse@nrw.social]

@atax1a

Incare abeout the code itsef. And it’s quality.

Have you ever seen ugly code written by humans? I have seen it. A lot! And worse than some LLM Output.

I am doing development since 30 years now. And we were managing people and bad code.

And people, who are children to me, want to tell people how to work?

If Slop is a problem, then detecting Slop contributions does not work. And Slop also comes from human devs.

The fish stinks from the head down.