Mastodawn

@dalias It's sad. LLVM has actually been officially tainted: https://github.com/llvm/llvm-project/commit/632151fbeea972f4aa3c14921eca1e45c07646f3

While digging, I found they actually *allow* AI slop: https://llvm.org/docs/DeveloperPolicy.html#ai-generated-contributions

Guess it's time to drop LLVM everywhere. I would consider using *any* of it a complete risk.

InstCombine: improve optimizations for ceiling division with no overf… · llvm/llvm-project@632151f

…low (#142869) Fixes #142497. Alive2: https://alive2.llvm.org/ce/z/CeaHaH The contents of this pull request were substantially written using claude-code. I've reviewed to the best of my abil...

GitHub

meek cynics Jul 1

@c8h4 @dalias By the way, you can also ditch Linux: https://lwn.net/SubscriberLink/1026558/7c46a42a4e8002f7/

At least I'd expect most BSDs to just ban it on sight, like due to not wanting a repeat of the BSDi vs. AT&T lawsuit.
But LLVM… yikes, goodbye GPU drivers.

Supporting kernel development with large language models

Kernel development and machine learning seem like vastly different areas of endeavor; there are [...]

LWN.net

@lanodan @dalias One of the articles closing sentences really says it all: "Levin said that he has not really thought about the problem, and assumes that, if an LLM produces code, he is free to make use of it.".
Guess that's the harsh reality we are heading with everyone hopping on the AI slop train - no one is thinking for themselves anymore about problems.

But at least it's only AUTOSEL *for now*, it seems.

Jul 1

@c8h4 @dalias It's not just AUTOSEL, first mention is https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=391dda1bd7c5

Making sure you're not a bot!

@lanodan @dalias What strikes me here the most is that it isn't even mentioned in the commit message.
At this point it's is just negligent and intentional code stealing and copyright infringement.

Cassandrich Jul 1

@c8h4 @lanodan TBF that one looks like it's purely formulatic following a pattern used elsewhere in same project, and may have been checked by a human author of analogous changes to match "exactly what I would have written" (except the ridiculous comment).

In which case it just burned a barrel of oil to save nobody any time.

@dalias @lanodan It is also setting a precedent for others, that's what worries me the most.

But yeah, I'm fully with you, it's a change that could have been made without further boiling the oceans.

Jul 1

@c8h4 @dalias Yeah in fact, I feel like coccinelle could have done that kind of change.

@lanodan @dalias Think so too.
Using coccinelle would have also resulted in a reusable script to detect similar patterns in other parts of the kernel tree, maybe proving useful for others in the future.

+bonifartius 𒂼𒄄 Jul 1

@lanodan @c8h4 @dalias the irony that they had to front the cgit with anti scraping measures.

@bonifartius @lanodan @dalias Yup, that's just the icing on the cake.

@lanodan @dalias And mentioning lawsuits; the infamous SCO battle comes to mind.
History is really bound to repeat itself ..

But oh well, I always wanted to dive into the *BSDs. Maybe it's time to try it out, getting a head-start on things is never wrong.

Howard Chu @ Symas Jul 1

@c8h4 @lanodan @dalias ah yes the SCO battle. I used to read groklaw.net religiously every day.

Sad how so many people fail to learn from history.

Andrey Rys 🇨🇿Jul 1

@lanodan @c8h4 @dalias what a nightmare timeline we live in... 😨

sarah tonin Jul 1

@lanodan please tell me the kernel doesn't take in ai generated code

xrvs Jul 1

@lanodan @c8h4 @dalias oh no

meek cynics Jul 1

Ah yes, @xarvos and @lanodan, let 2025 be the year of GNU/Hurd on the desktop! Given the impossible-to-determine nature of LLM-generated works' copyright, I hope @freedo will strip them off GNU Linux-libre though.

Cc: @c8h4 & @dalias

xrvs Jul 1

@cnx @c8h4 @dalias @freedo @lanodan i’d try some BSD bfore hurd