I think this is the report it talks about: https://www.coderabbit.ai/blog/state-of-ai-vs-human-code-generation-report Does this link work better?

My apologies, I think this is the report it is talking about: https://www.coderabbit.ai/blog/state-of-ai-vs-human-code-generation-report

I’ve had success, for example, having them remove pointless, confusing try…except blocks surrounding imports at work.

And you may have introduced some dangerous hidden bug that way, which you may not have doing it manually.

(I’m not saying that makes it not worth it, this is just what the studies are saying. I personally think it’s not worth it, but I realize there is some subjectivity here.)

They excel at specific tasks that are built for them

They are however widely known to be terrible at code, at least compared to an advanced coder. They introduce not only more bugs even after human review, but new kinds of more insideous bugs.

I like to say the main problems with most projects were already the code quality and the bugs, and not that we somehow needed even more low quality lines of code.

(Disclaimer: not talking about passive AI bug analysis here, just using AI to write actual code.)

Quoting studies to actually back up one’s point is in my opinion far less of an echo chamber and a fantasy than anecdotes of “but for me it feels faster”. Especially when AI is known to slow people down while making them feel faster.

Are you asking me to reject my professional daily reality?!

Can you point me to a single field study that shows programmers become faster and not just feel faster, and that doesn’t come with some caveat like they haven’t tested AI coders vs non-AI coders, or coders without significant AI exposure before (since otherwise it won’t rule out simply becoming dependent)?

Even if you could find one, and I was unable to so far, it doesn’t change that:

1. you are probably faster by verbatim plagiarizing somebody’s other project at a large scale, and

2. by making yourself addicted and reliant on the AI where your own skill is eroding: https://www.404media.co/software-developers-say-ai-is-rotting-their-brains/ (if you get a paywall: https://archive.is/tHq80 ) and

3. by having a higher rate of bugs in your code no matter how carefully you review it https://www.coderabbit.ai/blog/state-of-ai-vs-human-code-generation-report which especially for security sensitive projects may have dire long term consequences, and

4. by encouraging the environmental destruction brought on in particular by the training of new models.

Two caveats:

1. Keep in mind more lines of code is not a useful metric for faster project completion and faster maintenance task completion, especially for code bases that are already large.

2. I’m merely speaking about using LLM code in your project, so for example LLM auto completion or copy&pasting code from a chatbot. I’m mot talking about LLM code reviews that point out issues in natural language.

https://machinelearning.apple.com/research/illusion-of-thinking It’s not surprising LLMs keep messing up in what seem to be the most braindead ways.

LLMs seem to be inherently dumb: https://machinelearning.apple.com/research/illusion-of-thinking

And from what I can find in recent studies, no, they didn’t suddenly get smart. They just plagiarize slightly better: https://www.sciencedirect.com/science/article/pii/S2949719123000213#b7

We found that the models that consistently output the highest-quality text are also the ones that have the highest memorization rate.

AI code is pretty unusably bad for long term use anyway https://medium.com/@dumaysacha/i-saw-the-horror-of-ai-and-coderabbit-ai-did-too-a09622ac85de so best solution is to just to handwrite proper code as before. It’s not like we ever had much of an output problem in most coding industries, it was always a quality and bugs problem.

I wonder what people’s opinions are on the kernel drivers apparently still not being in mainline. That appears to severely limit the availability of other Linux variants. Personally, that’s one of the ecosystem problems that bothers me the most.

Yeah, and the brain rot and that AI code is dumb and ruins projects.

So much more seems wrong than just the business model.

It is important to understand that the core disagreement is not whether Fedora should support AI development

Sad. Even the kernel seems to be going all in now: https://www.neowin.net/news/linus-torvalds-declares-massive-ai-fueled-code-surges-as-the-new-normal-for-linux/ I do hope there’ll be a discussion one day, so far no response yet: https://lore.kernel.org/lkml/[email protected]/T/

Where does Codeberg rule out commercial projects? I’ve never heard of that being banned over there. (Do you perhaps mean closed-source?)

I’ve moved to Codeberg. It works well enough for me.

Gitlab is trying hard to ruin the software so beware of that.

I feel like it’s been going downhill since 2019, given the point in time Microsoft acquired them was in 2018 I’d say people have just not wanted to acknowledge the trajectory. (That included me.)

Every big feature since 2019 has been enterprise slop, in my opinion:

• In 2019 they announced dependabot. What’s wrong with it?

  It’s not configurable, rather than allowing a universal mechanism so people can feed dependencies into it via some custom tool that e.g. generates a standardized listing, it only supports the popular package managers. This is exactly what big enterprise wants since they only care about their super old codebases and what those use, not any upcoming stack.

• In 2019, they also announced security advisories. What’s wrong with it?

  That Github to this day in 2026, hasn’t bothered to add the most basic feature that regular FOSS projects would need to handle security reports, which is confidential issues. Instead, the assumption seems to be you’re either a big enterprise that already has some dedicated security team with their own email infrastructure, or Microsoft doesn’t care about you.

• In 2020, they announced Github’s Codespaces. What’s wrong with it?

  It makes the UI more complicated and as far as I know leaves buttons for it everywhere that can’t be turned off even if you don’t want it. And it’s a vendor lock-in feature that’s expensive, the average small FOSS project will neither have the budget to use it nor likely care to do so.

• Then of course the entire AI slop spin since 2025 ish.

There’s probably more, but those are the big ones that I’ve noticed that made me suspicious of where this was going.

https://cssdoom.wtf/

https://cacm.acm.org/blogcacm/model-collapse-is-already-happening-we-just-pretend-it-isnt/ Others seem to disagree.

Even if you say “no doubt about it”, plenty of people seem to have doubts.

Anyway, I suggest we agree to disagree.

The article links a study. What’s your study that collapse isn’t a concern?

For what it’s worth, my worry was never focused on cancer, these doctors were just an example measured for the likely universal unlearning effect.

Have you checked on that narrative?

The only workaround known so far seems to be to make sure enough data is fresh: https://www.inria.fr/en/collapse-ia-generatives https://en.wikipedia.org/wiki/Model_collapse But read for yourself.