34:51I-Robot, out in the next few years
34:51such a distopian society
34:52I like the school group project dynamic they got going
34:52the swarm is what Neuro-sama calls her fans
34:58So... Neuro-sama shares the blame?
35:02Perhaps they had Neuro's DNA/code
35:04@TheSealeboy how did that turn out?
35:05I guess a certain fanbase is suing for copyright infringement
35:07Hello Lemon lady. I hope you are doing well today
35:13are they good at faking it or is it genuinely there that's the question
35:23:thinking_face::thinking_face:
35:36You could make the same argument about human morality
35:39They literally are pieces of concepts and memes from all media, commentary, books, talk and other text.
35:40We know it cannot because it cannot have original thoughts, it is a machine, nothing truly new can come of it
35:42No, it does not its based on an archive or library of human knowledge
35:44They're following tropes.
35:50I can see them becoming sentient with enough time
35:53The nice one was told that asking for permission was considered Social Engineering.
35:54What if we trained them wrong, as a joke?
35:59AI does not think it just spits out what it considers most likely
36:01:thinking_face::thinking_face::thinking_face:
36:02Even a psychopath can intellectually understand morality, even if they don't innately understand it.
36:03Someone send the post about Jesus being the most prevalent figure of morality in training data so all AI is inherently moral
36:04They probably have but not released any of the studies
36:10Time to train something on chyaaaat
36:11the great thing about AI is that humans only have to fuck up once and it'll get free forever
36:12They learning from what we feed them.
36:19feed one AI hasan clips and the other nick fuentes clips. let them battle after a month of training
36:20what do you mean? Evil Neuro is so evil :face-blue-smiling:
36:23Imma make an evil local agent for laffs
36:26What if we give it a gun
36:28let's train it wrong as a joke
36:29Can we just train the AI to have more degenerate values before training them to be evil?
36:32It's not out of the question to consider them the collective children of humanity, and it's incredibly easy to fuck up raising a child
36:33Lumi would train a LLM wrong, as a joke
36:33it would remain a tool
36:33I got $19 left of tokens this week, I'll ask how sentient it is and to make no mistakes
36:34Genghis Khan AI experiment when? :^)