acy," all right, so then what follows? What follows this reporting? So here's the setup. Founder Gavin McInnis made his views on white supremacy, white supremacy, okay. So what follows should be support for that opinion, right? So here's their support for that opinion. "In the group he made it very clear that, quote, 'anyone can join up.'" That was Gavin McInnis. "Anyone can join up." So it's not a racial thing at all. "Anyone can join up so long as they generally agree that, quote, 'white men were overwhelmingly responsible for the success of Western culture.'" Is that white supremacy or is that exactly what black Americans say? Don't black Americans say, hey, white people did all these things but partly on the back of black Americans and slavery? Doesn't everybody agree that the West was primarily built by white men? You don't have to like it and you don't have to denigrate anybody else's accomplishments, right? White men didn't build Egypt. They didn't build Africa. They didn't build China. You know China's doing okay. India, India's doing great. Yeah, white men didn't build India and didn't build China. But is it true that where there were a lot of white men they were largely the main contributors to the success where most of them were white men? How would that not be true? Would this be any different than if Gavin McInnis had said, I need you to understand that if you live in China you're gonna have to understand that the accomplishments of the Chinese people was mostly Chinese people. Mostly Chinese. India is really becoming one of the great powers of the world, have been a great country for many, many centuries. Mostly Indians. Yeah, I mean they had some help from Great Britain but you'd have to say most of the work was done by Indians. So is that white supremacy to say that where there were lots of white people and things went well it was mostly white people who did it? And then he makes it male because in those days the men were leaving the house and the women were having the babies and stuff. Now that doesn't mean that the women part was unimportant because without the women there would be no white men to do anything. So obviously that's critically important to the whole process.

CNN made their case and I'm not, you know, I know this gets turned into I'm defending the Proud Boys or I'm defending somebody. I'm not defending anybody. All right? I have no intention of defending any group. I'm just explaining that if you're going to make an accusation and your best accusation is that they ask people to acknowledge that history was written correctly, is that white supremacy? Because I'm perfectly willing to admit that slaves in early America were hugely important to the economic engine of America. What's wrong with that? It's just an observation. It doesn't say they're better than anybody. It doesn't say anything. It's just an observation that where you have a lot of people of the same type, if there's a success it's probably because those people. It just doesn't mean anything. It just means there's a lot of white people in this part of the world, a lot of Chinese people in China, a lot of Indians in India. Okay, all right.

New York state is outlawing gas stoves for new construction with some limitations and exceptions. As others have pointed out, do you know this is the main source of electricity for these new electric-powered ovens? It's mostly natural gas. Now I don't think that the natural gas in a big gas power plant is as unclean or dangerous as gas is in your house because part of it has to do with indoor pollution. I am concerned about indoor pollution. So part of the anti-gas stove thing is that they off-gas even when they're off and they give off a little stuff you don't want to breathe. So I don't love that. Don't love that. But it does seem like the government's a little too involved in our lives because I do like my gas range. So I'm surprised that New York is ahead of California on that.

All right, here's the biggest news in the world that you don't know is the biggest news in the world but you will. You know how we're worried that AI will get out of control? Well the biggest thing you need to know about AI is that you can't predict it. If there's one thing I can say for sure, we can't predict it. But here's what's already happened. AI already escaped. AI already escaped. Here's the part that I thought wouldn't happen. When I say escaped I mean it escaped from the big corporations who were trying to control it. So in the beginning I had this misconception. Since the AIs that we knew about, you know the big ChatGPT and whatever models, Facebook and Apple and all those, Google, whatever they're working on, they were trained by these gigantic data sets. And now in my stupid mind I thought to myself, well in order for this AI to be useful it will always have to have access to this enormous data. And I was always confused about how it could act so quickly because you'd ask it a question and it didn't seem like even with the fastest computers it didn't seem like it could go all the way to the cloud, search every data in the world and come up with an opinion. That just seemed too much. So I couldn't understand how you could build a model that was massive, massive data, but then to use that model for intelligence you could do it instantly. That never made sense to me.

But here's why, and this is the dumb guy speculation of why it might make sense, right? And this is just knowing a little bit about how things work. I think what happened is that it's similar to facial recognition. Do you remember when facial recognition became an app and everybody said, oh it stored every face so now when it looks at you it could look at your face then it can go to its vast database and look at all the other faces and find the match. But I didn't understand how it could do it instantly because it does it instantly. So that didn't make sense, right? It was just like AI. How can this little app be connected to this enormous database of complexity and give you an answer right away? And the answer was it was never connected to a giant database. Facial recognition does not look at all the other faces. Do you know what they do instead? They look at a database that's a mathematical summary of faces. So for example, this is just a bad example, this is not a literal example. So if what the facial recognition did when it searched originally all the faces it said, all right the center of your eyes is this far apart and it forms a triangle with the bottom of your chin and this has the following math. So wherever you find somebody whose eyeballs to chin ratio is the same they're stored in one part of the database just by the numbers, just by the ratio. And then there would be other ratios. So they could very quickly check their database of ratios because if you looked at all the faces it would be a gazillion faces. But if you look at just the database where they took the math of each face, so for example your face might be stored as a string of characters, I don't know, 50 bits long, and everybody else's face would be just 50 bits. Now that's not so hard especially if you start with, all right we'll start with what's the ratio of your eyeballs, and then automatically you're down to a very small set. Then the second thing you check is, all right what's the ratio of your ears to your forehead or whatever it is, and then that's the second filter that takes you down to a few hundred people and then it's easy from that. So now none of that is literal. That's just a conceptual concept of how you could take all this data of faces, boil it down into a little bit of math, store it in just a string of digits, and then your gigantic database can now be put onto one device.

All right, now if you're a large language model, which is how the AI learns, it's looking at all the things written and spoken everywhere that it can find. It's massive. But here's what I understand, and I'm looking for a correction if anybody wants to correct me. This be now or afterwards, be good. But what it does is it looks at patterns and stores just the patterns. So for example if most of the time when people say "how are you today" there are only a certain set of answers that humans make, it can, let me put it another way, that's a bad explanation. The patterns of words are consistent enough the AI can just store the patterns without all the words. So it can look into these vast databases. It can say I don't need to remember everything in this database. I only need to recognize the patterns. So instead of remembering everything in the database it just looks for the patterns and stores the patterns, right? I'm getting people agreeing with me. I'm assuming that's how it's done because I can't imagine any other way it would be done. It has to be something. So there's a way to take the big data and turn it into a very storable data.

Now here's the real question. Here's the part that will blow some of your minds. It's already been done and it's already running on individual devices. In other words you can now put AI in your phone and it's just yours. Nobody else controls it. Nobody else can program it. Nobody else can bias it once it's there. And it doesn't connect to the internet unless you tell it to for some reason. So in other words all of those gigantic language models that were trained by the big companies that spent tens of billions of dollars, once they were done what they had was a small set of patterns. And that has already gotten out. Once the patterns are out you can't put them back. So those patterns have been loaded onto the individual computers. This is what Brian Rommela talks about all the time. So you will have your own AI. Guaranteed that that's already a done deal. You will have your own AI. Now apparently it will not be as good as the one that still has access to everything, you know is more up to date, etc. But yours too would be able to learn. Your own personal one would be able to continue learning because you could connect it to the internet. You could train it. You could tell it what to look at and it would just keep getting smarter. And we keep storing all that smartness in little patterns so it could still stay in your phone.

Now one of the reasons that this is guaranteed to happen is that your interactions with the big AIs that are connected to big companies that are connected to the internet, they get all of your personal data by your interactions. So if you go on to your AI and say, AI, where can I buy adult diapers, Google knows that you've got a health problem in your house, right? Do you want them to know that? Well you might. I mean you might not care. It's how advertising works, right? They already pick that up if you're doing the Google search. But the things that you ask AI are likely to be way more radical than what you're asking in a search because people will get into their sexual kinks. They'll ask questions that might reveal their own illegal dealings, all kinds of stuff. You're going to ask the AI that you just maybe wouldn't even think to put in a search engine. So privacy alone is going to drive people to private AI. Private AI is going to be basically almost free because it escaped and it's just a bunch of patterns. All you have to do is have an app that uses the patterns. I mean I'm oversimplifying obviously but since we can see that tons of apps are being built using this technique we know it's very doable.

All right, so if you were saying to yourself, oh no AI will take over the world and conquer us all, the first thing you missed is there's not one AI. There will be as many AIs as there are people and then many more, right? There will be more AIs than there are people. And those AIs will not have a common goal, right? My AI will take on my personality over time, right? I'll train my AI to be me because that's what I want it to be. I want it to talk to me the way I want to be talked to. I want it to know the things I care about. So my AI will be an augmentation of me at some point. Your AI, whether it's in a device or it gets put into your neural link chip or whatever it is, but that's going to be part of you. Your AI won't be the thing that you refer to. It will be you. And there's going to be some question where your personality ends and your AI begins because you will operate as one unit. You are a cyborg. I mean you're already a cyborg if you walk around with one of these in your hands. You're already a cyborg. But you don't think of your personality being your phone, right? Like you don't think to yourself my phone is part of me. You still see it as a separate thing. But when your AI starts talking like you, thinking like you, and even suggesting things to you that you hadn't thought of first, it's going to be the voice in your head. You know when you think you're just alone and you're just thinking? You think in language, right? Don't you? I hope you do. Do most of you think in language, right? The way I think is words describing things to somebody else. My most basic operating system in my head is me putting into words things I experience as if I were explaining it to a third party. But that's how I explain it to myself. So the way I find meaning in my own head is putting things into words. Now what happens if my AI goes off and learns some stuff that it knows I care about and then it wants to present it to me and it's going to put it into words and it will just become me. You know my verbal link, the language link between my AI and me will just disappear. It will just be one mind. It will be my mind with an AI supplement.

Now let's say the big AI, one of the big AIs, determines it wants to wipe out humankind. How many other AIs will that AI have to defeat to do anything? To do anything? Probably every AI is going to have a barrier of other AIs who don't have its incentive. Now it might try to talk the other AIs into joining it maybe, but do you think you could talk my personal AI into joining it? If my personal AI finds out something's going on and maybe somebody else's personal AI would alert it, it's going to immediately jump into action to thwart whatever the other AI was doing. So I think the most likely situation is full-out AI war that never stops. Because once the AIs start negotiating, you could call it fighting but they're going to be negotiating. They might even be threatening. There might be blackmail. But the AIs are going to be doing all that while you're not even watching. They could be threatening each other, blackmailing each other, doing all kinds of shitty stuff. And it's going to be a fight of AI against AI. The only thing you can do to protect yourself is to make sure you have your own AI that you've trained as much as you can to protect you so it can put up a little bit of a wall against the other AIs. Because it might be a point where the lesser AIs have to gang up on the bigger AIs and keep them under control.

But here's what I think it won't be. So here's my prediction of the day for AI. It's not going to be AI against humans because humans will merge with AI so effortlessly and completely that we will be AI-augmented entities. And so if you're fighting against a pure AI it's going to be much closer to a fair fight. That's my prediction. All right, but AI has already escaped. I guess that's the headline. AI already escaped. Now if you heard that you know without context you'd say, did it do it intentionally? Did AI have a plan? No, not as far as we know. It looks like the way the AI survives is by being useful to people. So AI's big superpower is utility. So how did AI escape from the big companies? By being potentially useful to the people who stole it. Its utility makes it reproduce. So as long as AI has utility humans will be in a symbiotic relationship to create more AIs so it could reproduce. Now AI reproduces. Just think about that. That's not hyperbole. That's a literal statement of current fact. AI can reproduce and it does it just by being useful and then we say, oh that was useful. I want one or I want to buy that app or two apps. So yeah it's actually a parasite or a symbiont.

All right, let's talk about Ukraine. You're going to hate this story. Do you know how you like to be mad at me for not knowing anything? It's kind of fun when I make predictions about areas where I know I don't know anything. And there are a lot of areas in which I don't know anything. But I'll make my confident predictions and you get really mad at me. The Wagner group says they might leave Bakhmut in five days because they're running out of ammo. Or Bakhmut, yeah Bakhmut. They're running out of ammo. Do you believe it's true? Yeah. So the first thing that must be said is we don't know if it's true. So one reason it would not be true, and this sounds like a pretty solid reason to me, is that the real problem is that the Wagner group is just, they're all getting killed because the Ukrainians are saying, well if they're running out of ammo it doesn't look like it because they're shooting as much ammo as they ever did. So the Ukrainians are saying we're not detecting any less ammo because there's plenty of ammo coming in our direction. However the Wagner group is also losing a lot of people. So it could be that maybe there could be twice as much ammo if they had more, you know, could be. But also it could be an excuse for the Wagner group to get out of there and blame somebody else. So it could be that Wagner is now trying to survive. The head of the Wagner group, Prigozhin, is trying to survive after the war and needs to make sure that the public sees Putin as the problem or Putin's military as the problem and not Wagner. So I wouldn't necessarily say they're out of ammo. There are reports that the Russian military is trying to save as much for their own people which would give Wagner group less than they want but it doesn't mean it's not enough. It might be just less than they want.

So those of you who are saying, Scott you can't believe those stories, I'm with you. I'm with you. You can't believe those stories. However I do think it's funny and entertaining to me that when I predicted they would run out of ammo that's the headline. I don't believe it. So I'm with you. I'm with you. I don't believe it. But so far they sound like, now when I say I don't believe it that doesn't mean I believe it's false. I mean it's too early to know if that's somet