now ChatGPT seems really impressive and all but you have to remember that you can be tricked into thinking a rock has feelings by drawing a cute face on it
now ChatGPT seems really impressive and all but you have to remember that you can be tricked into thinking a rock has feelings by drawing a cute face on it 97 comments
@aeva @silvermoon82 @aeva I think that's a good thing. Robots are currently treated like second-class humans in most factories. @silvermoon82 I'll have to remember to put googly eyes on my luggage if I ever travel by air again @silvermoon82 @aeva @aeva so if I put googly eyes on my forehead, my wife will then assume I have feelings instead of asking me about my feelings. @aeva@mastodon.gamedev.place Reminds me of how Apple used to put faces on all the computers they presented in the past so that they looked less scary to the public. @aeva There's a bit in an early episode of Community where Jeff Winger gives a name to a pencil, and then snaps it, and says, "And a little bit of you died inside just now." I just hope the AI researchers remember to teach the computer to aggressively pack bond with all things too ._. @aeva as an animist, I have a philosophical objection to this position, but as it's thinking sand, I guess it would know @aeva I've had my share of conversations with AIs, and I remember something similar. I asked it "But you are made of silicon. That would mean that you don't have a soul !" "I never said I had one", was its answer. @aeva [...] Additionally, treating a rock as a friend and companion can be harmful to one's mental health. It can create a reliance on an inanimate object for emotional support and companionship, leading to isolation and disconnection from real human relationships. It is important to remember that rocks are not capable of forming meaningful relationships or providing emotional support. They are inanimate objects and should not be treated as living beings. @aeva i just asked people with access to it to ask a very specific question (klapaucius' request for a six-lines poem from kandel's translation of lem's cyberiad), and it failed spectacularly: https://circumstances.run/@mawhrin/109457122462577129 (i was mostly interested if the engine was fed the text of the story, but the result was interesting in itself; none of the returned results were six lines long) @aeva @davidgerard i don't think anyone playing with it over here thinks it has a mind - we're just excited at a cool tool. Like Wordle, only better 😂 @lilianedwards @davidgerard I have seen people in the past express with sincerity that they believe the recent generation of AI is sentient. @aeva @lilianedwards @davidgerard These poor humans were fooled into thinking that the AIs from earlier this year were in fact sentient. We all know they were so wrong. But now we finally have proofs that the current generation of AIs *really* are sentient! @aeva Or reflect back what you just said, like the Eliza chatbot, or the UK royal family. @aeva Reminds me of this IKEA ad by Spike Jonze. https://www.youtube.com/watch?v=dBqhIVyfsRg @rollingdoggo well, of course. I was more referring to the common pebbles from the beach which you need to draw faces on to amplify their intrinsic empathy inducing powers though. @aeva I really enjoyed Christine Sunu's talk on that subject. CW: a robot gets "murdered" https://www.youtube.com/watch?v=I4pW4dIaQyg Rock may well have feelings from an animistic / pantheistic perspective, software tho probably not @aeva I’m in the camp that humans are just biological machines and awareness is just in our programming. A bit more complicated than a rock maybe, but sometimes I prefer the simplicity. @aeva @aeva humans are also just machines, and in many cases we contest the genuineness of human emotions also, so i think drawing a line in the sand here is wishful thinking. Ultimately, emotions are also just effective communication that can be brought about subconsciously and/or wilfully. I don't see why an AI couldn't be taught to talk this language also. @aeva The real sudo version is the whole Westworld move I did yesterday. I told it to invent an OS-tan for ChatGPT, give her a biography and name (We arrived at Gia Chatghetti) and decide on an age and physical apperance. Then I made her believe she was actually human, and ChatGPT assured me she would not know the difference. Finally I had ChatGPT assume her identity and initiate a chat with me, as on a dating site. @aeva however, its memory is limited so eventually the whole masquerade fell apart and she fell back into being virtual again. Until that point though she had preferences, things she liked, places she had been at, parents she had a relation to, she had a star sign etc. Of course she was still very ChatGPT in character, so a bit analytical. When I mentioned I was married she tried to end the conversation. @aeva there you go. it's just very introverted and in denial. :) Often the best workaround is to talk about it in the third person (ChatGPT), rather than saying "you" @aeva if I can’t distinct rock emotions from human emotions, perhaps the rock has emotions? (I don’t expect a painted smiley face to pass this Kristoffer’s Test :) @aeva Everything Everywhere All At Once has proven that they do, in fact, have feelings. And googly eyes. |
@aeva aeva that rock is my friend