Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Debian has published the official results for the 2026 GR on LLM usage

vote.debian.org

Post reply on HN
Print view View original

1–10 of 10 posts

Re: Debian has published the official results for the 2026 GR on LLM usage

#2

Post by tulio_ribeiro »

Good outcome, mostly.

Option E beat literally every other option, so it is an unambiguous Condorcet winner.

Options: https://www.debian.org/vote/2026/vote_002

The tally.txt file is still unavailable.

Bradley–Davidson analysis: https://gist.github.com/tulior/74479b28ddd1e191e5b2d89b8b68c...

The Bradley-Davidson model gives Option 5 a 99.9993% posterior probability of ranking first. Options 2 and 6 are effectively tied for second.

Option 5 also beat every alternative head-to-head. Its closest matchup was against Option 2, at 57.8% to 42.2%, so the result was decisive overall even though the closest pairwise contest was not a blowout.

Re: Debian has published the official results for the 2026 GR on LLM usage

#3

Post by parasxos »

For anyone not wanting to parse the ballot page: the winner was Proposal E, "Responsible Use of Generative AI" i.e Debian neither endorses nor prohibits AI tools. Contributors remain fully responsible for anything they submit, and disclosure is encouraged but not required. The outright ban via the Social Contract (Choice 1, which needed a 3:1 supermajority) lost to this decisively.

Re: Debian has published the official results for the 2026 GR on LLM usage

#6

Post by roryirvine »

The graph at https://vote.debian.org/~secretary/gr_llm/ is really useful when understanding the results, too.

Debian is remarkable for being such a huge project which has successfully held itself together for so long. GRs like this are an essential part of that, allowing controversies to be settled in a way that manages to be both fair and transparent.

Re: Debian has published the official results for the 2026 GR on LLM usage

#7

Post by NitpickLawyer »

> That supports “decisive win” comfortably; whether it qualifies as a “landslide” depends on where the margin threshold is set.

I think the landslide is "pro" vs. "against". The two most "against" options got the lowest rank. Top 4 are all "pro" (with different caveats for each).

  Rank Option 
  θ
  i
  1 5: Responsible Use of Generative AI 1.794
  2 2: Allow AI-Assisted Contributions with conditions 1.430
  3 6: A cautious approach to generative AI 1.424
  4 4: Accept AI contributions for Debian specific work 1.252
  5 8: Avoid LLM use: climate impact 1.007
  6 7: Debian is created by humans 0.881
  7 9: None of the above 0.762
  8 3: Reject LLMs as far as practical 0.683
  9 1: Ban LLM contributions via Social Contract 0.474

Re: Debian has published the official results for the 2026 GR on LLM usage

#8

Post by grahamjperrin »

tulio_ribeiro wrote: post #2

Good outcome, mostly. Option E beat literally every other option, so it is an unambiguous Condorcet winner. Options: https://www.debian.org/vote/2026/vote_002 The tally.txt file is still unavailable. Bradley–Davidson analysis: https://gist.github.com/tulior/74479b28ddd1e191e5b2d89b8b68c... The Bradley-Davidson model gives Option 5 a 99.9993% posterior probability of ranking first. Options 2 and 6 are effectively tied…

Is this the tally?

https://vote.debian.org/~secretary/gr_llm/results.txt

Re: Debian has published the official results for the 2026 GR on LLM usage

#9

Post by tulio_ribeiro »

grahamjperrin wrote: post #8
tulio_ribeiro wrote: post #2

Good outcome, mostly. Option E beat literally every other option, so it is an unambiguous Condorcet winner. Options: https://www.debian.org/vote/2026/vote_002 The tally.txt file is still unavailable. Bradley–Davidson analysis: https://gist.github.com/tulior/74479b28ddd1e191e5b2d89b8b68c... The Bradley-Davidson model gives Option 5 a 99.9993% posterior probability of ranking first. Options 2 and 6 are effectively tied…

Is this the tally? https://vote.debian.org/~secretary/gr_llm/results.txt

It is. Thanks for the update.

Re: Debian has published the official results for the 2026 GR on LLM usage

#10

Post by tulio_ribeiro »

NitpickLawyer wrote: post #7

> That supports “decisive win” comfortably; whether it qualifies as a “landslide” depends on where the margin threshold is set. I think the landslide is "pro" vs. "against". The two most "against" options got the lowest rank. Top 4 are all "pro" (with different caveats for each). Rank Option θ i 1 5: Responsible Use of Generative AI 1.794 2 2: Allow AI-Assisted Contributions with conditions 1.430 3 6: A cautious appr…

Great insight. Let me see what ChatGPT can cook if we classify each. The tally is now available; might be useful.

### Edit

Re-classified using permit, prohibit, materially_restrict or none_of_the_above as I think these are more accurate labels.

Conclusion: All five proposals that continued to permit AI-assisted direct contributions occupy ranks 1 through 5.

The result is:

permit: 15 wins

restrict/prohibit: 0 wins

Every single permit option beat every single restrictive or prohibitive option head-to-head.

ChatGPT’s interpretation: Debian was divided over how AI-assisted contributions should be handled, but not over whether they should remain permissible: all five permissive proposals outranked every proposal that materially restricted or prohibited them, and permissive proposals won all 15 cross-category head-to-head comparisons.

Post reply on HN