Conversation

Notices

1. Embed this notice
   [object Object] (zzt@mas.to)'s status on Thursday, 09-Jul-2026 08:46:20 JST [object Object]

   it’s very rich every time the guys who cognitively surrendered to the slopbots and can’t program without them think we won’t be able to tell their code is machine-extruded garbage if they manage to, just once, not scream about how much they love LLMs in every conversation, avoid filling the repo with AGENTS.md, CLAUDE.md, .cursor, et al, and tell claude to be extra careful to not put “one-shotted by Claude Fable!!!!” in its commit messages

   • Embed this notice
     [object Object] (zzt@mas.to)'s status on Thursday, 09-Jul-2026 08:47:04 JST [object Object]

     in reply to

     “you couldn’t tell it was slop if I lied” said the same fried brain that thinks “Key Takeaways 🚀” is a feature of a readme any fucking human wrote

     Rich Felker repeated this.
   • Embed this notice
     [object Object] (zzt@mas.to)'s status on Thursday, 09-Jul-2026 08:47:04 JST [object Object]

     in reply to

     it’s so weird that I won’t take advice on the nature and dynamics of lying from the most unpleasant men I’ve encountered in my field, as they tell me themselves that they’d rather lie to my face than respect my boundaries and my consent

   • Embed this notice
     [object Object] (zzt@mas.to)'s status on Thursday, 09-Jul-2026 08:47:55 JST [object Object]

     in reply to

     anyway I’ve been looking up packages lately for specific purposes and I keep getting very new, very confident packages that seem to solve my exact problem!

     the repos hide it but I can tell within seconds they’re slop by how their readmes look, and when I dig into the commit log it becomes even more obvious

     and then the repo owner just admits it in conversation because again these embarrassing fuckers can’t ever shut the fuck up even when they’re trying to trick you

     they want a pat on the head for fooling you into thinking they did a good job, that’s the point of an LLM

     Rich Felker repeated this.
   • Embed this notice
     [object Object] (zzt@mas.to)'s status on Thursday, 09-Jul-2026 08:47:55 JST [object Object]

     in reply to

     “boy don’t you feel foolish, you fell for this LLM generated code/generative art for a few minutes”

     no fucker, I feel angry

     why in fuck would this ever be a shame on me situation

     you’re a conman running a con, and like all conmen you got really good at tricking people for long enough to mug them and run

   • Embed this notice
     [object Object] (zzt@mas.to)'s status on Thursday, 09-Jul-2026 08:47:56 JST [object Object]

     in reply to

     you know, maybe the fuckers unabashedly telling us they’d lie at the first inconvenience are, in fact, misrepresenting the situation in some way in order to extract an advantage over us

   • Embed this notice
     [object Object] (zzt@mas.to)'s status on Thursday, 09-Jul-2026 08:48:37 JST [object Object]

     in reply to

     besides everything else, a ban on LLM code is a ban on low-quality code

     the sealions doing their work yet again to DARVO us about an LLM ban can’t process that, because they think LLM code is good

     we don’t in fact have to accept low-quality trash

     we don’t have to accept good code from dickheads either

     LLM slop happens to be both very bad and contributed by someone you really don’t want as part of your community

     Rich Felker repeated this.
   • Embed this notice
     RRB (rrb@infosec.exchange)'s status on Thursday, 09-Jul-2026 08:48:37 JST RRB

     in reply to

     @zzt Harry Frankfurter (cite On Bullshit) defines BS as statements that may or may not be true, the person saying it "does not care." They only want it to be plausible. The mixture of truth and lies in a palatable wrapper makes it worse than real lies and harder to distinguish.

     LLMs are BS generators by their nature.

     The tech lords are solving the low quality BS penury problem, <sarcasm> since we all have noticed too little shoddy BS in our lives <\sarcasm>.

     Let's all keep pointing out that AGI and superintelligence will be more LLM BS by their very nature

   • Embed this notice
     [object Object] (zzt@mas.to)'s status on Thursday, 09-Jul-2026 08:48:38 JST [object Object]

     in reply to

     and while we’re here, I’ve covered this before but the sealions have been arfing and clapping about it again

     “aren’t you worried that a ban on slop will discourage legitimate contributions”

     if your contributions are so bad they look like slop, you should be discouraged

     LLM code is uniquely godawful in a way that only a machine remixing a bunch of plagarized crap can be

     it’s really not that hard to spot it at all, and the only people I’ve seen try to weaponize a false accusation are slop fans trying to make a point. because they’re fucking dickheads and everything is a weapon to them.

   • Embed this notice
     Daniel Quinn (danielquinn@mastodon.social)'s status on Thursday, 09-Jul-2026 08:49:44 JST Daniel Quinn

     in reply to

     @zzt "We don’t have to accept good code from dickheads either" -- solid advice.

     Rich Felker repeated this.
   • Embed this notice
     Maya Granade (kevingranade@mastodon.gamedev.place)'s status on Thursday, 09-Jul-2026 08:50:01 JST Maya Granade

     in reply to

     • Daniel Quinn

     @danielquinn @zzt I've been fighting this one for years and years. With bystanders telling me I'm not allowed to moderate who has access to submit issues and PRs no matter what toxic behavior they bring with them because "progress is progress". Guess what not all issues/PRs are created equal in the first place, and more importantly chasing off the other contributors has a much higher impact than *anything* one person is going to contribute.
     I've been training for waves of AI slop for years :P

     Rich Felker repeated this.
   • Embed this notice
     [object Object] (zzt@mas.to)'s status on Thursday, 09-Jul-2026 08:50:38 JST [object Object]

     in reply to

     • Daniel Quinn
     • Maya Granade

     @kevingranade @danielquinn exactly! somehow the “you can’t kick out toxic contributors, look at what they’ve done for us” guys never pay attention when their toxic faves bully marginalized contributors out of projects, and they never do the easy analysis or extrapolation to determine the damage caused by the loudest voices on the project being toxic

     I keep watching this stuff kill projects, but somehow the idea that we should do anything at all to prevent or remedy it is unthinkable for a lot of these guys