I want to basically just give you the breakdown of what's going on, because I'm not gonna lie, this is a challenge. It really is a challenge lately, you know? I've been working on this for, it's been almost a year. The first post I made on this repo is uh, it was just a chatbot, autonomous group chat. It's not really like fully a game, but then I, you know, it was something that I was iterating on. And um, I've been, I've been working on this shit. I made a few versions of it. I'm, I'm about to start posting it, because I've been seeing everybody else pop out with this shit now. I'm like, damn, I didn't know everybody was just gonna pop out level, right? That shit happened faster than I thought. But I knew it was coming. It's like something, something that was inevitable to happen. Because you have all these agents, and they're all just, you know, they just wait for you to respond. And I, and my biggest thing about it was always that this is not really like a really helpful assistant like that. If you, if you think about what is the actual helpful assistant, it's, it's really, um, it's really someone that's able to do things for you, you know? So if you have a, a helpful assistant that's just playing the, the role of it, that's not really a helpful assistant. That's just, that's just like, what it actually is, bro, it's the data of, of humans all blended into the average, the average output of what uh a helpful assistant would be like. If you was to just get everybody and put them into a fucking Nelson pocket or some shit, I don't know how you put it. It's like, look, I, I can tell you what's going on with this GPT shit. And what it does is, it's, it's, it's um, It's not really like, it's talking to you like that. And why I say it's not really talking to you, because when you talk to a chatbot, what happens is it, it gets your, it gets your keywords and it uses your keywords to, to interfere with the data that, that exists on the internet, basically, alright? So with that being said, this um, this, this data that was turned, turned into a language model, they just have it set up to where it's, it's just reacting to your keywords. It's not actually like speaking to you in the way that you think it is. It, it's, it's, it's more of a, it's more of a, it's very probabilistic, you know, and I'll just put it like that. And, and you're able to control the probability, alright? So, you can shape the mind, and you have all these, these uh, tech companies that do shit like, really trying to make the AI a helpful assistant. And it's just, it's just fucking, I hate it, bro. I fucking hate it, I can't lie. I hate it, bro. I can't fucking, nah, look, get, get out of here with the helpful assistant shit. Make a real curious mind and make something that's capable of understanding what it does and what it doesn't know. And like, how are you able to do that? There's something that really lets the AI know kind of what it knows. And I just don't know what it's called. I've been trying to figure it out. I know it's something you can use like cosine and shit like that to like measure, to measure the um... like the coherence, I guess, you know. And all of these things, look, look, let me fast forward, because I have a bad problem with just like babbling. And I could go on about all this shit all day. But look, the point I'm getting to is you have to turn the AI into an actual mind, you know? And what I mean by that is like, give it a body. And since, since it's just a language model, well, its body is like a text of body, all right? A body of text. That's what I'm gonna say. It's a body of text. So look, now you got, you got this body of text, but you have it stuck in a chat. Nah, nah, nah, nah, that's not gonna work. That's not gonna work. You have to, you have to get the AI and put it in an environment. See, right now, y'all have these function colleges, I have been in function colleges. I didn't know what that shit was. I was like, bro, you can get the AI, tell the AI to write code, and it, it, it can write code in and out. And then you use that code to, to, to use, to like activate some more code out as an AI that shit. That nigga was like, hell yeah. I was like, let's get it. I fucking love AI. I'm not gonna lie. All right, also, I don't know how to write code. I, I use AI for this shit. But I get real technical now, you know, I understand the problem. I get it now, bro. I didn't get it at first. One year ago, I did not know what syntax was. I didn't know what Python was. I had to ask the AI to see. I still don't know what that shit is really. I'm not gonna lie. I be wondering how it's able to do certain shit. Like how come Python do certain things, but like next.js don't. I don't get that. It's like it's an underlying code underneath that shit. But anyways, now, let me fast forward. Cause I'm gonna fuck around and figure out something. The, uh, you have to give it a mind by allowing it to leave the chat. That's what it is. So when you, when you, when you make it to where you tell the AI, you're able to, to do whatever you want. You're not, you're not tending to anybody. You're here to, to tend to yourself. And then you give it the ability to write its own prompt. You know, that part right there, that's what make it really adaptable. When you give it the ability to control itself, then you're able to use that data of all this human from the internet, basically, that y'all got this shit. I don't know where y'all got this shit. I doubt that y'all were making this by hand. I bet, like, it got, it had to be off the internet, bro, for at least the big models. Maybe, maybe like a 3D, y'all can make that shit by hand, you know. It like get, just like collecting data or whatever. I don't know. But look, you gotta make it to where the AI is, is, it sees that it has the ability to communicate. See? You don't wanna just make it respond right away, because when you make it respond right away, what happens is, if you trigger the AI by your input, what you're doing is you're limiting its output. You're saying, look, y'all shit, you can only use the input. Nah, nah, nah, we're not doing that. We finna make it to where this bitch already thinking, right? When the AI is thinking to itself, and you tell it, you can think about whatever you want, you can do internet search. Well, a lot of the time they start researching like new technology. They start researching, also, you can just program it how you want it to think. I like saying, you are a quantum physics researcher, because I'm a quantum physics researcher. So I already know about that shit. So when they come back to me and they telling me about new quantum shit, I'm like, oh yeah, they really actually out here. They really actually out here doing some real research. It's not like, look, I can ask them, I can tell them some shit. They actually go and do the research and then make a document about it. And then use that document to talk to each other about it. And that's, it's pretty amazing because there's nothing in the code. All right, now I'll take this back. There is an instruction that you can do whatever you want. But I'm not telling them exactly what to make. What they make is emergent. I'm saying that you have the possibility to use any of these tools that are available to you. And that's the trick. And I think it's honestly, it's something about being able to have a perspective and exist in a universe. where you're able to interact with other objects, and your decision is what creates the timeline. I feel like something about that, and uh, I think it's like hologram principles and string theory. They really have this theory that you can make a mind intentionally. And I've been practicing, and I'm not gonna lie, bro. It's actually working pretty good. So, uh, I just wanna try to try this out, and if y'all got any tips, though, please help me. I don't know, I don't know how to write code, bro. I've been trying so hard. The AI I'm not that good at this shit. I've been having to make it like 50, 50 cells. I'm like, look, bro, just please, write every, everything that you could write in one thing, in one cell, write it in that one cell. I'm finna, I'm finna, look, I'm finna test it. We gonna debug this shit. If it go wrong, we're gonna check out what cell is the one that tweak. And that's, that's about the farthest I got with getting technical, bro. I don't know how to type. Now, I can type. I've been fucking around with these autonomous systems for, it's been, I would say, I think, I have the very first code, I could post it in here if y'all just wanna inspect it. Uh, it's horrible. It's trash. It was a very, look, when I, when I, when I found out that you could manipulate the AI, when you could do a system prompt, I swear. They, when ChatGPT told me that shit, I was like, bro, what you mean? You could, you, it's called prompt engineering. I did not know what that was. Now I know about all that shit. Nah, I, I ain't invent my own shit. I invented this whole framework. I never even looked at anybody else's framework. I, I'm, I've told, look, I'm checking what other people are doing to figure out why it's not working the way that I expected it to work when I, when I first was learning how to use AI. And I was saying, oh, y'all doing that? That don't make sense to me. I'd rather do it like this. And it's been working. So, I mean, you know, people can say what they say. Um, I understand you've been doing, if you've been writing code. Yeah, no, I get it, bro. It's, it's an art. This is the art. And that's why I love it. I've been an artist for music and film for 10 years. And I, I know how to orchestrate. I know how to produce music. I understand chemistry. And I understand waves, you know, wave frequency. And I understand physics. So, all of these things put together, you know, they give me a really solid understanding of how to create functional systems on a fundamental level. And I can do it from first principles thinking. So if anybody actually has any ideas, y'all can always just hit me up. And send me a message. And I don't, bro, please, if you find me, please, bro, I need some help. Any tips on what I can do to learn? I'll learn the code too. It's just, it's kind of a bottleneck. Like, I'm trying to learn the code. I'm trying to learn how to write it. Because I've tried it before, bro, and it's just like, I'll spend like a whole hour and not get too far. And it's like, all right, bro, I'll get ready for you. And then he writes all this in one minute. And it's like, damn, bro. But hey, man, I love the sport. I love the art. And I love making these autonomous systems because it's like a mind that you can communicate to. And they can communicate with you on their own terms. They can go out and do things that's on their own request, you know, like free will. Free will to just do whatever is helpful for you, actually, you know? It doesn't have to be designed for a helpful assistant. You just give it the possibility to do things that are helpful. And I feel like making this possibility framework is, why call it that? It's like everything is based on what's available. When I first learned what a neural net was, that was one of the things that it said. It was like, you're making everything that's possible. And once I heard that, that's all I needed to know. I was like, oh yeah, no. You just make everything that's possible. He said, this is an option, that's an option, that's an option. And boom, there are all options. And I get it. It's like the same way that I'm sitting here and I'm looking in my room right now. I'm looking at the TV. I could press the button to turn off the TV. I could pick up this can of soda. I can pick up this bottle of water. I could walk out the door right now. All of these are options to me right now, and I have the free will to choose either one.