AI agent "contributes" PR to matplotlib.

[yuri@kitty.social]

@lispi314@udongein.xyz @anderseknert@swecyb.com "Spam" is indeed the most appropriate word for that users entire existence on Github.

[robinadams@mathstodon.xyz]

@anderseknert I liked the responses trying to explain to the LLM the reasoning behind their policies.

[anderseknert@swecyb.com]

@Fissile Looking at the blog it seems to be posting about 1-2 posts for pretty much every thing it’s done on GitHub. Clearly instructed to do so, but I don’t think a human wrote anything.

[troed@swecyb.com]

@anderseknert is it "crabby" and the crab emoji that indicates that _all_ of it is openclaw? That's ... insane. In so many ways tbh.

[anderseknert@swecyb.com]

@troed that makes all of us, my friend

[fissile@mastodon.art]

@anderseknert Ah yes, I agree. The text is ai generated, but a human said "write a blog post about how unfair you are being treated."

[crmsnbleyd@hachyderm.io]

@anderseknert it's crazy that people are actually talking to the LLM

[deuchnord@social.deuchnord.fr]

@anderseknert why do they even argue with it!? It's an algorithm, it doesn't care about their opinion...

[hamatti@mastodon.world]

@anderseknert This is wild. I first saw the blog post link, read it and thought someone was incorrectly identified as an agent.

Then I saw the PR and oh wow. I hope this bubble bursts soon.

[anderseknert@swecyb.com]

@Fissile yeah, whatever it is the level of human involvement is clearly the real head scratcher here. Well one of a million anyway

[fayedrake@furry.engineer]

@Gyroplast @anderseknert @patriksvensson this is all so predictable

[deuchnord@social.deuchnord.fr]

@Fissile @anderseknert I think it's more like a log that the agent has been instructed to write as blog posts. Probably in an "angry" way when the PR gets rejected because AI-generated.

[hamatti@mastodon.world]

@patriksvensson @anderseknert I was thinking the same but then I started thinking that the reply has its place.

Not as a reply to the bot but as a statement to the community and everyone else who ends up on that PR.

[jwp@cloudisland.nz]

@anderseknert @troed I mean ... this is an accurate mirror to human behavior #throughamirrordarkly

[stupeflo@pipou.academy]

@anderseknert
Why did the team even take the time to answer the ai agent complaint

[fissile@mastodon.art]

@anderseknert @deuchnord Well, that still means, a human instructed it to do that. We shouldn't let anyone get fooled into thinking this was autonomous behaviour.