Join the discussion

Write your take first — we'll ask for email only when you're ready to publish.

  • Hacker News
  • What an interesting cultural artifact! I note no proposals in favour of "let it rip" which is presumably because that would not be a reflection of Debian philosophy. I do think the aggressively anti AI proposals are themselves un-cautious in the extreme.
  • It's one thing for a maintainer to use AI, and another to review AI code sent your way. The latter has been proven to be unsustainable, low value, and ultimately a challenge for open source. Everyone acknowledges this problem exists and some corrective action is required to mitigate it.
  • Crazy how divided AI is here on HN…

    We’re almost to the point where the average non-technical person uses AI daily (just like phones and the internet were once only used by the technocrats) without blinking, it will be ubiquitous and universal in adoption, yet comments here on HN from people that have the vantage point of seeing AND understanding how AI works, and who can also predict the inevitable integration of AI in society, have got their collective heads in the sand to the point they’re soon going to sound like conspiracy theorists.

  • > the average non-technical person uses AI daily

    Citation needed.

  • "Having one's head in the sand" means being unaware of something; everybody is aware of it. "Conspiracy theorists" is just a slur.

    If you have anything substantive to say about it that other people haven't heard, we're listening.

  • > it will be ubiquitous and universal in adoption

    [[citation needed]]

    Which is gonna be tricky as I am confident you don't have a time machine.

    I think generative AI as a whole is a total scam. It has uses, but not many, and outside of translation, not good ones... and yet, it is running at truly epic financial losses. Multiple trillions of dollars are being pumped into this.

    I think the next AI Winter, the 3rd big one, is coming very soon.

    There's going to be a market crash to dwarf the dot-com crash and 20087 financial crisis (both of which I worked through as an employed adult), and a lot of the current tech industry will be wiped out.

    I've watched that happen before as well.

    I am quite looking forward to it. I have no stocks or shares in anything anywhere, few savings, no loans, no debt... but all my immediate family own their own homes outright, and I am close to retirement age.

    Let it all burn.

  • Every non-technical person I know hates genAI and goes far out of their way to avoid it.
  • Total ban please. This nonsense needs to end. Linus Torvalds is weakly vibe coding the kernel, that's bad enough.
  • Then you cannot use Linux, Firefox, Chromium, Python, Rust, LLVM and so on. Hurd can't be used either.

    Maybe you have to switch to FreeDOS?

  • Ban.
  • The options range from is A to H. It's not yes or no.
  • Posting this because I want people to vote -- and ideally vote for a total ban.
  • > and ideally vote for a total ban.

    If a total ban were to happen, wouldn't that make Debian a target for hackers who can use LLMs to find vulnerabilities?

    How would Debian keep up with the influx of AI assisted 0 days?

  • > Posting this because I want people to vote

    Irrelevant to post it here then with your reasons because it only concerns Debian developers

  • What are your arguments for a ban? Why do you expect someone to listen if you didn't provide a justification?
    by js8
  • Are you going to fork kernel? If not then you are just a bunch of stupid hypocrites.
  • Given almost all people here cannot vote, does this mean you posted non-AI-generated slop?

    Wouldn't it be better to reduce slop, no matter how it was produced, instead of only AI-generated contributions, including non-slop contributions?

  • It does not appear that normal people can. There is no vote page, no information on how to, and we can only view results.
  • I'm curious about enforcement. Are they going to ban packages with a whiff of llm? If so, how will they reliably audit?
  • I do use and run my own LLMs locally. But I understand why Debian might ban it, under just 1 reason:

    LLM outputs cannot be copyrighted.

    That alone starts to invalidate the GPL and other FLOSS licenses. It overall weakens the project the more you accept. And companies (MS, Amazon, etm) will gleefully loot anything marked with "LLM" as a free-for-all.

    Until the whole copyright situation can be finalized, I'd understand for any FLOSS org to refuse all LLM code.

    (Its also the same reason why Oracle also forbids all LLM code in Java and their DB. Same reason.)

  • There's not yet any reason to believe that mixing LLM output with your work makes it uncopyrightable any more than hitting enter in intellisense did before that.
  • > LLM outputs cannot be copyrighted.

    For stuff that isn't Debian supplied Debians only contribution is the packaging. Debian considers "Public Domain" is a perfectly acceptable licence for the packaging work. So LLM produced packaging is also perfectly fine - copyrightable or not.

    By the by, even the option most favourable to LLM's in the ballot insists the Debian Developer take responsibility for all his work, regardless of whether he or an LLM produced it.

    > That alone starts to invalidate the GPL and other FLOSS licenses.

    How?

    > And companies (MS, Amazon, etm) will gleefully loot anything marked with "LLM" as a free-for-all.

    I'm not fan of their sharp practices either, particularly when I wrote this: https://lwn.net/Articles/1046105/ But what Amazon could do with publically licensed packaging that they can't do now is a mystery to me - it's not terribly useful outside of the Debian ecosystem, and it's not like you are forced to distribute most of it. Many companies just ship Debian binary packages now, no Debian source available. GPL and friends only bite if you distribute it.

    I do think LLM's could be a threat to open source licences, but this isn't the mechanism. The real threat is far more insidious: https://lwn.net/Articles/1064113/

    So, for Debian packaging LLM's don't create copyright issues, or accoutability issues. If environmental concerns are serious, Debian could always setup it's own LLM server farm running open source models powered with renewable power. That sounds expensive, but whenever Debian needs compute (it already use a lot recompiling all those packages on all supported arches) it just seems to "appear".

    I'm not sure what's left, beyond a technophobia of computation done using 4K vectors rather than bits. The underlying silicon is the same after all, lots of packing tools already automate most of the steps, and the output (the packaging) is highly constrained so will be near identical.

  • > Debian is a large and complicated project: the release announcement for version 13 says it has 69,830 packages, which take a total of 403 GB of disk, and contain 1,463,291,186 lines of code. So, suitably, it is a large and complex poll.

    Surely these are not just Debian-owned packages. Whatever Debian decides won’t apply to packages not owned by Debian

  • Debian is huge - that's why so many other distros use it as a base!

    So, yes, all those packages are maintained either by an individual Debian Developer or a team of them.

    There are third party repos, but they're (currently) neither as widespread nor as commonly-used as those for other major distros. This vote won't affect them, and you might potentially expect to see their use grow if one of the harsher options in this GR were to pass.

  • > If Proposal H has a weakness, it's that it does not distinguish between local and cloud-based LLMs

    Is the energy usage so different between local and cloud inference? Both require electricity, the local option even more than the better optimized cloud variant even perhaps. How either is powered makes the crucial difference I suppose. Both can potentially run on solar as well as gas or nuclear.

    It's the training that takes the most energy, and that needs to happen for locally running or cloud models regardless.

    What am I missing here? EDIT:

    Proposal H mentions "LLM usage accelerates the destruction of our ecosystem"

    I was thinking solely about energy usage, but there is of course also water usage. A local setup is not water-evaporator cooled most likely.

  • People want to ban LLM contributors not because of energy usage, but because AI generated code is unmaintainable and the people who generate it tend to engage with others in domineering and brazenly manipulative ways.
  • This doesn't have anything to do with water usage? The rationale of banning it is that it's copyright status is ambiguous and at the least unethical, it produces lower quality code, and it stifles new developers from getting involved.
  • > Is the energy usage so different between local and cloud inference?

    In throughput mode for agentic loads, the energy usage (tok/s/MW) of the new NVidia Vera Rubin hardware is 30x lower than that of the B300 and perhaps 450x lower then the H200 was, which in turn is hundreds of times lower than the inference for single users at home in any non-data-center hardware. It feels like comparing the momentum of an ant to the momentum of an elephant.

    by pama
  • Kind of arbitrary since it doesn't impact their upstreams.

    The very first thing I used GenAI for was fixing build and CI pipeline issues, exactly the drudgery no one cares about. Today, LLMs are excellent at the yeoman's work of maintaining patches across multiple versions and upstream, which is what I imagine a lot of Debian devs do.

    Why you wouldn't want to hand that off to a computer program is beyond me.

  • > Why you wouldn't want to hand that off to a computer program is beyond me.

    Trust, stability, security. The reasons many people run Debian.

    Although I think the main incompatibility in philosophy here is in "hand that off", e.g. direct agentic usage eroding opportunity for human judgement, which seems unlikely to be adopted anyway. The realistic proposals are essentially AI use with various guard rails and rules to preserve that human judgement (my interpretation).

    Actual Proposals (scroll down): https://www.debian.org/vote/2026/vote_002