Yeah. Kinda feels like OpenClaw and a few ‘trends’ before that.

…The heck happened?

I’ve been under a rock for a bit.

Explain to me: why is there so much hype (and controversy) around some Hyperland Arch Linux config, within like a month? Is this some astroturfing thing?

Well… IMO, there’s a whole spectrum between “very rare LLM brianstorming/troubleshooting” and “I don’t know how to code” vibe-coding, and Im not sure what the distribution looks like.

Personally, I get LLMs to generate boilerplate, formatting, and functions, but manually write plenty and audit/assemble all the pieces myself. But I would not call myself a vibecoder; I loathe vibe coding.

Yes, that too.

Incentivizing hiding LLM use would be a problem. Encouraging disclosure is good.

The title is pure clickbait, but the article is nuanced and reasonably well written, IMO.

No one wants a freaking Ubuntu chatbot on their desktop, but code submission is an entirely different issue, and I think “the submitter is entirely responsible for their code” is a very reasonable position.

And obviously, LLM-generated code policy needs to be tuned to each project and their contributor community. As an example not cited in the article, vibe-coded PRs are an ongoing plague for llama.cpp, an LLM inference engine itself, so they blanket ban them. Ironic, and perfectly reasonable. But Linus researching that Intel bug with a little help, but using his brain for the legwork and actual submission, is also reasonable.

You put that incredibly abrasively, and the analogy is questionable (as we are talking dev tooling here).

…But your point is technically correct.

Contributors who you will NOT like contribute to useful projects. And your contributions will get used for things you don’t like. That is, indeed, how open software works, and what makes it inclusive.

Whoa, that is very unlike Debian.

Titles of the proposals for those interested:

• Option 1 “Ban LLM contributions from Debian via Social Contract”
• Option 2 “Allow AI-Assisted Contributions with conditions”
• Option 3 “Reject LLMs as far as practical, update Code of Conduct”
• Option 4 “Accept AI contributions for Debian specific work”
• Option 5 “Responsible Use of Generative AI”
• Option 6 “A cautious approach to generative AI”
• Option 7 “Debian is created by humans”
• Option 8 “Avoid the use of LLM: climate destruction is a deal breaker”

And the voting system isn’t so simple, it’s really best to click through to the data/tables:

https://www.debian.org/vote/2026/vote_002#quorum

https://en.wikipedia.org/wiki/Condorcet_method

They don’t need their own LLM. They can just use existing open weights models trained far more modestly.

Maybe finetune them for some specific task, but TBH something like Deepseek/GLM flash, Qwen 3.8, or even Nemotron are going to be plenty good for automation and chores.

…But the thing is, I don’t think Debian can dictate an automation setup to contributors, just like they can’t dictate what IDE or other software they use to code. It’s up to contributors to be conscientious.

Yeah, the headlines is editorialized.

“Neutral” is right if you read the proposals: https://www.debian.org/vote/2026/vote_002#quorum

The closest to what you want is probably the Nvidia Nemotron series, which uses an open dataset. It used a sizable GPU cluster to train, but nothing on the scale of what OpenAI/Anthropic are guzzling. That, and Nvidia makes a point to advertise high utilization/efficiency.

There are smaller scale truly open LLMs like the Olmo series, but Nemotron is the most practical to use.

Then there are the Chinese “open weights” LLMs, which tend to be trained on more modest Huawei ASICs instead of GPUs, and probably with a good chunk of renewables. The dataset is closed, and who knows what is in there, but at least the training scale is much smaller

They are Apache licensed.

Another thing is that both these group pitch LLMs as modular tools to customize, not magic black boxes to rent.

Im not downvoting, I think its interesting.

…But iI suspect some are tired of the Linux community’s fragmentation? As there are a lot of distros that try to mimic Windows as effortlessly as possible.

This looks like a very clean, sleek attempt to me, IMO.

If you are just looking to anonymize YouTube, I’d browse with Cromite,download longer videos with yt-dlp, and watch them in mpv. Or use something that automates that.

You get better video quality that way, anyway.

Just FYI, it might not be sophisticated enough to beat YouTube ads. Not sure about that.

Yep, built in, with customizable lists.

Nope.

I think many Chrome extensions don’t work either, because of disabled APIs.

FYI extensions are another thing that can be used to fingerprint you. There’s a discussion on this somewhere on the Cromite GitHub, hence they don’t recommend extensions.

I mean Google can identify you with stuff like your resolution, cpu and gpu, measured execution speed, extensions, APIs, software versions. Browsers expose all that.

Check this site out. Its a great explainer: https://amiunique.org/

Now, if you run amiunique in Cromite, the results are interesting.

It may identify some stuff, but if you actually look at the readouts, you can see Cromite is spoofing data. It intenionally lies about your hardware and configurations to throw off fingerprinting, amongst other things.

This means it can actually mess up some sites that rely on the data: thats what I mean by “draconian.” Tons of features are disabled or stripped out. Cromite is a very extreme browser to daily drive, because its purpose built to defeat fingerprinting at any cost.

Use: https://github.com/uazo/cromite

My guess is YouTube is fingerprinting your browser, and no browser does antifingerprinting specifically as obsessively as Cromite.

Just read its feature list. Its quite draconian, heh: https://github.com/uazo/cromite/blob/master/docs/FEATURES.md

Okay, well, what about Nvidia Nemotron’s series as a counter example?

It’s built on an open dataset; its training regime isn’t secret, it’s reproducible. It’s Apache licensed. It’s not published to manipulate anyone; it’s mostly presented as a research tool, or a starting point for users to customize for mundane text processing tasks. It runs reasonably efficiently on a large variety of software and hardware, with a standardized architecture.

Of course AI, as it is mostly presented to the public, is an existential problem.

But you’re trying to fundamentally tie text-based machine learning to the tech bros as a root for its ethical issues. It’s not that simple. Even if Sam Altman and all his kind keel over tomorrow, the complications the existence of LLMs and other generative models present is not going away, and we will have to deal with models that can do unethical things without anything unethical in their training regime.

You’re lumping all of machine learning into the bucket of tech bros, ChatGPT, Facebook and such.

There was a thriving, small scale world of machine learning stuff that existed before they hijacked the public’s attention, and it never went away. It still exists. It will continue to exist when this bubble at least deflates some.