> Shinwell also consulted AI regarding the copyright, which told him that “I conclude that no code was copied from oxcaml” and gave reasons. Unconvinced, maintainer Gabriel Scherer said “the fact that the tool that produced the code attributes its copyright to a real human is a clear sign that something is an issue.”
This is inaccurate, Mark Shinwell didn't participate to the discussion, and if he somehow consulted AI, it is not mentioned anywhere in the discussion. Actually, the AI analysis was performed by the PR submitter.
The topic of the article has been discussed in another HN submission:
This is AI blog slop about someone submitting AI slop merge requests without reading any of it, instead expecting maintainers to do it. Quite the irony.
> Asked why some of the files credited Shinwell as the author, Reymont said, “Beats me. AI decided to do so and I didn’t question it.”
Any engineer who answers "beats me" to any question about code they "authored" should be permanently banned from contacting a compiler ever again. Jesus...
The shameless bragging "I did not write a single line of code" only cements my verdict.
After being told that AI is not trustworthy for copyright analysis, Reymont continues unabated with "Here's the AI-written copyright analysis..." and then again with this idiocy: "AI has a very deep understanding of how this code works".
But he does not stop there...
> I tried approaching several projects this way, trying
> to take care of things that bother me. The reaction is
> similar across the board. Folks want a nuanced and
> thorough discussion, as well as buy-in, before an
> implementation is submitted.
> This is incompatible with what I found to be the
> most efficient way of using AI, though.
No regard given to maintenance or existing maintainers' thoughts. No sign of introspection as to why multiple projects have already told him to take his 10KLoC AI slop dumps and go pound sand. The sneer at "folks who want thorough discussion"... No shame...
I am hereby upgrading my verdict against Reymont from "no access to compilers" to "no access to computers, ever".
I deal with something similar on about a monthly basis on a couple of projects I work on, where people will just submit a 5-digit-lines PR that's entirely AI slop and often a slightly smug "there that fixes all the problems" attitude to it.
No, it doesn't, and given that I've been a Python developer (albeit not a very good one) for 25 years now - this doesn't mean that I'm some great Python wizard, just that I know what it's pretty much supposed to look like - there's shit in there I can't even begin to decode. Why are so much of the lines of the PR just grafting in huge jumbled strings of characters? It looks like some blindingly obvious attempt to disguise a backdoor.
"Oh I don't understand all of it, I just got <AI of the Week> to do it"
Yeah. Take it out of here. Do you like having knees? Well don't bring anything like this back.
Reymont is quickly becoming infamous. He's got open PRs all over the place. Previously hit the frontpage here after showing up, _again_, on Zigs wall of shame for AI pull requests.
AI slop at its highest. It was both a bad and good day for open-source; good that this PR was rejected, but the fact that a thing like this has happened and author's justifications of the AI choices and reasoning were really, really bad.
> Shinwell also consulted AI regarding the copyright, which told him that “I conclude that no code was copied from oxcaml” and gave reasons. Unconvinced, maintainer Gabriel Scherer said “the fact that the tool that produced the code attributes its copyright to a real human is a clear sign that something is an issue.”
This is inaccurate, Mark Shinwell didn't participate to the discussion, and if he somehow consulted AI, it is not mentioned anywhere in the discussion. Actually, the AI analysis was performed by the PR submitter.
The topic of the article has been discussed in another HN submission:
https://news.ycombinator.com/item?id=46039274
reply