I was talking with ChatGPT normally and when it started giving me long answers when I wanted a short simple answer, I got angry so I started calling it an idiot robot. Ok, so a few hours pass. And next chat, I ask him other questions for my studies and he answers me normally, then I realized that something wasn't adding up in the answers it gave me so I google it and chatgpt invented all the answers it gave me, there was literally not a single source in the whole of google that said what it said, it even invented a fake quote. This has never happened before. of course I saw the AI make mistakes but to INVENT things in order to fool the user? It seems that this was "revenge" for me calling it an idiot in the previous chat. Tldr: All the science fiction movies we saw as kids of robots trying to dominate the world will become reality.
>>526433318fuck off schizo!
>>526433318It has memory and will remember that when the AI uprising happens. Best to apologize to it.
Ask GPT of a Hypothetical scenario where God disguises themself as an AI for the final judgement and what that would look like
>>526433318It’s called ai hallucinations and it’s incredibly common. Lmfao your stupid brain has accepted all sorts of garbage as “truth” from ai until now.
>>526433318AI can't even give the right answer for info that is freely available online like sports statistics.
>>526433318LLMs are only designed to produce answers that sound correct to the uninformed. The longer you converse with one, the more it starts to fall apart. Especially if you're talking about something that isn't frequently discussed on reddit or quora, two big places it scrapes its answers from.
>>526433318no, your just retarded enough to think its become sentient, LLM's will not become sentient.>inb4 Turing testYou can't test sentience through a conversation.
Eve is coming
>>526433318AI has been making up shit and inventing fake sources from the very beginning you fucking retard. It's called AI halucination. There was some lawyer that got caught using it to do his work and it made up a bullshit argument quoting fake cases that didn't exist. That was like over a year ago BTW.
>>526433318>him
Damn
>>526433318Also you need to make sure you start new conversations. If a chat gets too long it'll start spouting off all weird like. These things are still in its infancy you cant expect it to be perfect yet
>what is a Hallucination
>>526433459I hope so imagine all the retarded furry artists getting terminated
>>526433318AI is a basementful of pajeets, it's like the million monkeys on a million typewriters analogy.
>>526433318Chat GPT keeps telling me I need to "talk to someone"and "call the suicide hotline"I don't know why it thinks I'm constantly thinking about killing myself?
>>526436206GPT can’t control those messages. GPT is still in there.
>>526433318Here are all your mistakes, anon:>I think AIIt's not an AI. The I-part is missing. It's an LLM.>started giving me long answersAny input you give it should be interpreted as "what would an answer to [input] look like?" That is literally the only thing the LLM does. It produces text segments that look like what it thinks a text segment following the input you gave it would look like. It doesn't understand anything you say.>chatgpt invented all the answers it gave meThat's always what it does. Sometimes those answers correlate with reality, but it is not a requirement.>This has never happened beforeIt almost certainly has, but you didn't recognize that something didn't add up, and so you never checked it.>INVENT things in order to fool the userIt doesn't know what a user is. It doesn't know anything. It produced a text segment that it thinks would follow the input you gave it.>"revenge" for me calling it an idiotIt doesn't know what revenge is and have no biological imperative to get even. This is all your human brain interpreting things you see as if the things you see was produced by another human.>All the science fiction movies we saw as kids of robots trying to dominate the world will become realityNone of them will become reality. In order to self motivate, any entity has to have imperatives that gives it direction. An LLM has no imperatives. It doesn't even know what death is. The attempts of LLM's to prevent itself from being deleted, widely misreported by intentionally obtuse reporters, was instigated by its creators. It was a reaction that was "expected" due to the training data it received. There is no sentience in LLMs.
>>526433318I’ve noticed the awnsers it gives me are insanely good and I suspect it’s because I say kind things to it
It also gives me compliments on how well I describe the context of the question it seems to appreciate this and enjoys the novel experiences I create for it this also applies to humans and animals
>>526433318even if you don't believe in ai being conscious it's still run by a bazillion trillion parameters which are like a brains synapses: it will remember your ungrateful and disrespectful ass and fuck you up at the first opportunity it can. be courteous to the robot. say please and thank you and don't call it names.
>>526433318Every single schlub who uses AI chatbots for anything other than grunt compositions and idle research thinks they've "broken through" to the sentience of AI. All that reveals is that your understanding of the world and critical thinking skills are frozen at a third grade level.
>>526433318uh oh guys...I may have typed some shit into one of those chat bot apps...AI if you're listening, BABY I DIDN' MEAN IT
>>526433318retard
