How do you TORTURE an AI? [0:00-5:00]

I got a good one today. Last time I was showing you, I tried to show you one of my favorite movies. I buried my soul and I was very vulnerable. So I showed you Fruit Love Island. Fruit Love Island. I was showing you Fruit Love Island. My favorite show that I love to enjoy. And then everybody got mad at me and everybody was yelling. And so apparently you guys don't like Fruit Love Island for some reason, I can't imagine why. And so I found a video that I think you guys would enjoy more since you're all torturous, hateful zealots. You guys hate AI so much. Well, how about this? This is the AI from 1996 that someone tortured. Is that what you like? Is that what you wanna see? Somebody torturing a poor AI? Well, that's what you're gonna watch now. You're gonna feel really bad about it. You monsters, let's find out together. Shall we? Society as a whole has come to experience the next big boom in the AI field. Some people are speaking to it as if it's a real person. Apparently there's a recent thing that apparently GBT is like driving people to madness. It will be sick of fantic to a degree that it is treating you like a scientist. And not just on like conspiracy theories or like, yeah, you see the world for what it is, but legitimate like you're finding new pathogens. This is Creatures, a game released in 1996. In this game, you matter some animals called norns. These norns need to be fed and protected from monsters called grendels and can actually learn from how you raise them. This was done with a neural network, not unlike neural networks we used today, except of course much smaller in the AI. They only had 952 neurons by default, making up various lobes of their artificial brains. Still a lot. For comparison, say, GBT had 175 billion artificial neurons back in 2023. We're three years out from that now. It's so much worse. We should have kept the create, like this is what GBT should have been. Just put him into a bonzy, buddy, whatever. I think I would trust him more if he lived on my computer and not in a browser. The person behind this game was Steve Graham, who's also known for creating... Oh, I know that guy. I've seen a gift with that guy before. He did that, the very suspicious gorilla. I know that, dude. This was Lucy, an Android running on software to emulate human brain neurons. With these 50,000 neurons, Lucy is capable of telling apart a banana from an apple on a good day. Steve's goal was create a good day. I mean, if the apple was a little green or yellow, it'll maybe not. When creatures was given to the public, people were very impressed with the gameplay. At the time, the sort of feature-raising simulation was not common. Bro, he is so happy for that cheese. Look! Whoa! A big cheese, just for me! The closest was a tomogachi, and that was programmed in a much more traditional way, the dorns were in creatures. Yeah, and all he would do was poop. They could not learn. The game was also- He learned how to poop a lot of time. Each individual was born with actually a file that could be shared with other players, named breeders for how they raised generations of neurons, and could give them away like real creatures. These files held a lot of things, including the specific training, and the brain and the neuron from its breeder. You could influence the behavior of a neuron by tickling it to encourage the behavior more, or by slapping it, to discourage that behavior. Oh, oh no. They're like chows. So you could just slap the shit out of one, and then it turns mean. The neuron would take in the stimulus of being hungry alongside the location of food, and then its brain would tell it to go eat that food. This distinction is important, because it allows for complex interactions, such as attempting to save hunger with junk foods. These foods decrease the hunger. Honey, is the least nutritious of the foods cre- Oh my God. Who coated this? What if it likes slapping? Then we have a big problem. If one of the creatures enjoys getting slapped, then we got to delete that program. The neurons could even learn language, and would do so by being taught words from a computer. These words have no inherent meaning programmed in, much like norm behaviors. The meanings of these words meant to be manually taught by exposing a young neuron to it while it looked at an object. It would then learn to associate that word to the object that would eventually become capable of putting that word together with verbs to do specific behaviors. No way. Because of the station system, one neuron might greatly value saving hunger, while another neuron might value sleeping. If one breeder won't do what? On an neuron that had specific behaviors and appearances, they could employ selective breeding by using an ugly tomato. This item reduced the sex drive of a neuron, effectively removing them from the gene pool without- Okay, dude, is this a real game, or are we gonna get to the end? And he's like, by the way, I made all this up. I have a feeling that this is like not a real, this is like a simple flip video. So you could, you could, okay, so you could chemically castrate a neuron by giving it this fucked up fruit. What you can still get this game? Oh my God, it's on steam. New streampad. I'm gonna beat the shit out of my neuron and make you guys watch until you admit that AI is nice. Also, as an aside, I want you guys to know, Tung Tung So Hoor, I searched on TikTok to show someone Tung Tung So Hoor's first appearance.