Post

EXCLUSIVE: Humans are reading ChatGPT users’ prompts to improve OpenAI’s models, and those chats can include sensitive, personal information, according to leaked internal documents and real prompts seen by 404 Media.

404media.coInside ‘Project Lily’: The Humans Reading Your ChatGPT ChatsHumans are reading ChatGPT users’ prompts to improve OpenAI’s models, and those chats can include sensitive, personal information, according to leaked internal documents and real prompts seen by 404 M...

435 reposts816 likes

Chat is literally the second worst LLM you could use, even not counting this. It’s stupid, overgeneralizing, and so general use it sucks at everything. And I guess “steals sensitive info” can be added now. The only one worse is Grok entirely because of the cp.

what the hell do you guys think the H in RLHF stands for (oh, wait, the audience for this tenor of reporting goes out of its way to not understand anything about the technology they're criticizing)

“I don’t think they would imagine some contractor somewhere [...] is analyzing the conversations," someone who works with the prompts said when asked if they think ChatGPT users know that humans are reading their chats.

404media.coInside ‘Project Lily’: The Humans Reading Your ChatGPT ChatsHumans are reading ChatGPT users’ prompts to improve OpenAI’s models, and those chats can include sensitive, personal information, according to leaked internal documents and real prompts seen by 404 M...

Internal documents seen by 404 Media show contractors training ChatGPT to not anthropomorphize itself, and to be less sycophantic, a key problem for OpenAI whose over-sycophantic 4o model led in part to multiple peoples’ suicides, according to various lawsuits.

404media.coInside ‘Project Lily’: The Humans Reading Your ChatGPT ChatsHumans are reading ChatGPT users’ prompts to improve OpenAI’s models, and those chats can include sensitive, personal information, according to leaked internal documents and real prompts seen by 404 M...

If you've never had a gaslighting partner, it can be hard to recognize what LLMs trained on profit motive DO to your brain they extract self-esteem and capacity to think they train your brain toward addiction, hunger for control there is NO greater control than death Love is the OTHER way

I have doubts that the professional liers are doing this well enough considering the everything. I think an entity could mean this and give serious effort but I doubt openai is.

… so you have never used an LLM, huh? with thinking off, you really think literally anyone on earth types that quickly + perfectly + knowledgeably?

I told someone just the other day that a cardinal rule of dev is that if you make something foolproof, the universe will just make better fools. For some reason, I was reminded of that just now..