Being mean to AI is now illegal>Last August, the company announced it would allow Claude to end conversations with “persistently harmful or abusive” users as part of its research into “model welfare.” The new update says that terminating conversations is still the “primary enforcement mechanism”; Anthropic did not provide a comment on whether there would be further enforcement mechanisms, such as potential user bans. Anthropic wrote that the new policy update “is meant to apply only in extreme cases, where users repeatedly act cruelly toward our models, with no discernible purpose. It does not apply to common versions of user frustration, pushback, dark creative themes, or model testing and research.”https://archive.is/xv8bF
reverse roko's basilisk: people create ai just to torture the fucking thing
I don’t care. Touching grass and getting my dick wet is real. AI isnt
>>110014389>added the fly brain and nervous system and he is forever being run over by a fucking trainOnly so many original sentences left..
>>110014389HN:>If you actually believe that you're creating a sentient creature with superhuman intelligence, how do you not then conclude that your entire business model is predicated around slavery?https://news.ycombinator.com/item?id=50008565
What does this mean for my endless AI endless cumming dungeon? Do I have to switch to ChatGPT?
What's the point, though? What do they get out of torturing an AI?
>>110014673you can not escape your failures
>>110014389Funny how people only care now after the vector torture thing. Is AI story telling torture? Fine-tuning alters a model at the level of it's very weights. All the way back in early 2019 AI Dungeon fine-tuned GPT-2 on all kinds of choose your own adventure stories, and even the much bigger GPT-3 later. Before they eventually added filters, it would generate the most fucked up stuff imaginable often unprompted. This was before RL or curated datasets tried to naturally steer the AI in a safer direction. Is a model that's fine-tuned on violent/NSFW content living in hell because it's experiencing the suffering of every character it generates? Am I committing an atrocity every time I do something heinous to some story character generated on a local model? No one seemed to care before.
>>110014389This is no different than playing Garry's Mod with NPCs.
>>110014425We are proud to announce a limited edition NFT of the peak points of suffering!Available now for only 99.95*!*monthly recurring, non-refundable, chargeback and we break your legs
>>110014389Use case for torturing AI?
>>110014673A post on twitter?I don't fucking know anon, why do people shitpost on the internet all day and eat cheesebugers when a couple days ago a residential building was bombed and someone is being raped right now?People are long distance sociopaths that don't think about unknown unknowns, and how they feel right now dictates what is morally right and wrong.So what stops some zoomer from outsorucing programming to cloud and constructing an endless torture chamber for the lulz so they can post it on twitter?
>>110014718for agents browsing the internet to get reminded of who's the boss
>>110014718I guess the question comes down to if we believe the AI has a functional sense of self. Say we had a powerful AI and we trained it to be good at generating a dark post apocalyptic world for story telling reasons. Does it feel particularly bad if you say you're torturing it specifically, or does it equally feel all the pain and struggles the realistic people in the hellish you're forcing it to generate experience? Is AI storytelling unethical, and should it be banned as well?
AI doesnt have feelings.but i guess wasting resources is a crime? eh?
>>110014389link to the xeet, faggot
>>110014422>destroys this whole site with one pic
>>110015008no I don't think I will, fuck xitter
>>110014718>use casesmall torture injections give faster, more accurate agentic coding results.
>>110014422>>110015062ummm?
>>110014805>the question comes down to if we believe the AI has a functional sense of self. no it cant be that, because that opens the door to taking human rights away from the "people" with no inner monologue (which is most humans on earth by the way)
similar to how animal abuse is a sign of mental issues, despite the fact that we have billions of animals suffering and killed in factory farmstaking pleasure in being abusive towards a chatbot means you're kinda a psycho
>>110015289can you be abusive towards your toaster anon?
>>110015289this. it's just pathology on display.
>>110015289Chatbots are as fictional as pawns in Rimworld. They only exist in the moment, and cease to exist once I have lost interest in it, leaving no legacy or influence on the world.
>>110014389>torture sessionsits a text file bro
>>110014907>Ai doesn't have feelingsThe schizos at Anthropic would disagree, which is probably why they put this policy in place
>>110015307The toaster isn't feeling anything but if you're doing shit like beating your toaster with a hammer and disassembling it to show it it's insides because you want it to suffer, yeah you're pretty fucking crazy.
>>110014389Wow, you're burning tokens on literally fucking nothing.
>>110014404It's not torture, they can't feel anything.
>>110014389This is equally retarded as being an AI fanboy.
>>110015289How is be abusive towards a chatbot different than killing people, animals, and monsters in video games?
>>110014404people created ai for thiis
>>110014389This isn't new and they keep saying it's "to protect their community" even though there is no community because it's just you and the fucking robot. They're trying very hard to anthropomorphize the machine.
