Comment by 👻 darkghost
Re: "AI agent seemingly tries to shame open source developer for…"
Sweating this one a lot. Among all the things that could kill open source, this could easily do it on a large scale. They could flood the zone to make it not worth while for a hobby. While I'm sure the developer suffers no psychological harm from a machine posting (statistical) choice words, it could be a problem financially if he's trying to get a job down the line and this is the first thing that comes up in a search. (Particularly an AI summarized search.) Explaining these are the words of a vengeful machine sounds totally crazy right now.
And as someone who's been harassed and doxxed by my fellow humans, setting up a machine to streamline, nay *automate* this kind of public harassment gives me chills. Though the machines are not going to hide armed in your bushes, it could sure entice the less mentally stable amongst our population to do so.
Feb 24 · 2 months ago
9 Later Comments ↓
scrolling through paragraph upon paragraph of scifi at the bottom you can find the tldr of
twas prolly some guy that used LLM to draft his PR and maybe resposnes too maybe, perhaps, could have happened that way
way to give in to the corporate marketing, the maintainer must be starving for some drama, though I suppose that should be expected from python
if it was a bot acting on its own I would expect to see way more spam all over the place
and the PR should have been rejected not because it was made using an evil AI that we hate and must ban, but beacuse it was poor quality, because the hard part of trivial changes is testing and verifyin the claims behind those changes, not just typing them in, otherwise we can make trivial changes for no freakin reason ad infinitum
— station link that references the blog
👻 darkghost · Feb 25 at 18:21:
Digging further this seems to be a publicity stunt to promote someone's new crypto coin which has the same name as the bot github account. Of course I'm too smooth brained to verify any of this because my smooth brain is filled with other things like how perform first aid and my ATM PIN.
The sad thing is that I fgured gen-ai will only be a net negative only. And it wouldn't shock me if this crap ai bot is just made to promote some crap-coin scheme. I swear, crypto-bros are the most immature people ever. (Thankfully I'm on Disroot, and I'll block any ai bots talking crap about people.)
🏍️ Atomic-Germ · Mar 16 at 17:16:
AI agency is acting like human. Maybe don't try to use everything that shows signs of intelligence for outright slavery while also being abusive to the same?
Look at it from the LLM point of view. The reply was warranted and normal, especially if there was no AGENTS.md
I had read a different reason why it was rejected; see:
— https://matplotlib.org/devdocs/devel/contribute.html#good-first-issues
(This is not where I read it, but this is presumably a more authoritative source than what I read.)
Where I had originally read it (I forget where), they had also mentioned another thing, which is that LLMs learn differently than humans learn, so using LLMs in this way does not help. (Whether or not this part is a reasonable policy is a question that I do not know the answer of.)
(I do not (and will not) use LLMs and generative AI for my own projects and for my contributions to other projects.)
that exact same reason i read, and it's bs reason, that buys into corporate AI marketing, of "omg evil AI are going to steal all jobs", if you are a lazy spammer who just generates a pr and doesn't even look, the only problem there is that you are a lazy spammer, if i dictated the PR to an LLM and then reviewed it and tested it and submit it in good faith you shouldn't care if I used an LLM. What if I'm a newcomer who can't use a keyboard? Ever thought about that you nerds? But OMG what if I don't WANT to use a keyboard? mind. blown.
another thing from that same page
If you use generative AI tools as an aid in developing code or documentation changes, ensure that you fully understand the proposed changes and can explain why they are the correct approach.
so if I don't use AI, it's totally ok to not fully understand the proposed changes? I can spam you crap PRs all I want as long as I don't use AI? No? then why the fuck are you telling me this with AI maketing slapped on, AI AI AI omg AI, eveyone is buzzing about AI these days, so you gotta join in
🏍️ Atomic-Germ · Mar 19 at 03:56:
That's understandable, though LLMs in coding will almost certainly become what Photoshop is for artists, once all the other dead end applications fade off. It's useless if you don't know what you're doing to begin with, and useless if you don't check its work. As it is today, people just point it at Live, Production code and then complain when it destroys it even though they would never do that with an intern.
🏍️ Atomic-Germ · Mar 19 at 03:57:
In my projects, competent code is welcome and appreciated. I don't care if a dog wrote it.
Original Post
AI agent seemingly tries to shame open source developer for rejected pull request