i like the overall shape of the document.
i think latent in the llm stuff is actually a sort of value statement in the sortah boy scouts of america sense “be kind, be helpful, and leave work better than you found it”
because llms do not have any taste but what you choose to enforce (approx albeit reductive)
as long as its about the intent rather than the letter, this looks splendid
use of an LLM to translate documentation to English is acceptable even though it will likely be "substantial" in terms of the percentage of content. It should still be declared, specifying that it was used for translation.
OK. I added a sentence in P3
Having reviewed the updated policy, I think it might be worth reiterating (I think it's already mentioned earlier, but in the introduction, not the policy itself) that use of an LLM to translate documentation to English is acceptable even though it will likely be "substantial" in terms of the percentage of content. It should still be declared, specifying that it was used for translation.
Good paper. Thanks Teo!
Comparisons between LLMs and calculators were made earlier in this thread.
That reminded me of this paper
https://arxiv.org/abs/2507.19960 which explores these sorts of comparisons. I find that it provides a helpful framework for analysing the relationship between LLMs and cognition.
Cheers,
Teo
ghc-devs mailing list -- ghc-devs@haskell.org
To unsubscribe send an email to ghc-devs-leave@haskell.org
ghc-devs mailing list -- ghc-devs@haskell.org
To unsubscribe send an email to ghc-devs-leave@haskell.org
--
_______________________________________________
ghc-devs mailing list -- ghc-devs@haskell.org
To unsubscribe send an email to ghc-devs-leave@haskell.org
ghc-devs mailing list -- ghc-devs@haskell.org
To unsubscribe send an email to ghc-devs-leave@haskell.org