nd AI. And she talks about how she's used a bunch of the AIs and she's doing it for her work in science. And that it often gives her sources that don't exist or there's a 404 error. It'll tell her that a document that says 2025 on it is from 2023. So she has to continually fact-check the AI's work. And it just looks kind of limited to her and wonders why people are so enthusiastic about its future.

Now I've said similar things and people's response to me and to other people who have had this experience, they say you just don't understand how exponential improvements work. So the smartest people who like AI are saying Scott, Scott, Scott, you're so dumb. Let me explain how the world works. Can I? AI is not supposed to be the best it will ever be today. It's going to get better like really fast and there's going to be this exponential growth that's going to happen any minute now. And when that kicks in you're going to look like such a fool. Oh man, a fool, because you don't understand how exponential growth works. Yeah, it seems slow, slow, slow, slow, but then when it kicks in, boom. And that's why you're not seeing it because you're not, you know, you don't understand technology like we do.

To which I say have you ever heard of Moore's law about the microchips getting better every year? It's a straight line, no exponential period. Have you ever heard of a thing called fusion, nuclear fusion, that was first proposed and worked on 93 years ago? 93 years, still waiting for the exponential growth part. 93 years. What about airline travel? I'm waiting for the exponential growth in that because since I was a child it's been largely the same and the planes don't even change. They're like 40 years old. Yeah, so we got that. So what else we got? What about batteries? I'm always talking about all these breakthroughs in batteries. Now that's a case where I think there will be an exponential growth. And the reason is that there are people all over the world who know there's enormous money to be made in making a better battery. Enormous. So you got all kinds of people, the best people working on it all over the world, and they're trying different technologies. So if you have the best people all over the world and giant money involved and working on different technologies, the odds of one of them being 10 times as good as the old one are pretty good.

So if you said to me Scott, don't you know that batteries may be chugging along for a while but once we get to a certain point man those batteries are going to be amazing, I would believe that one. I believe that one. It's just that not everything can turn into everything else. What about smartphones? So our smartphones got kind of awesome because they're smartphones. But how different is your current smartphone from the last one? I feel like smartphones sort of peaked. I'm waiting for that exponential growth but I don't think it's coming with smartphones. Probably some other technology.

So here's what I think. I used AI five times already this morning, mostly Grok. It was great because that's how I know that fusion has been worked on for 93 years. I just asked Grok. And that's how I knew what the WTO does. I asked Grok. So if you're a writer or you're working in this kind of world, this podcasting world, and you're trying to get context and you're trying to get an understanding of a new topic quickly, oh it is great. It is great. So in some professions, and Bill Gates was saying that medical profession, legal profession, some other ones, we're going to see AI takeover and I believe that. So I'm very pro-AI. It's just that I don't think the current technology, the large language models, are ever going to help Sabine do her science stuff because you can't rely on it being right. It might be that you always have to fact-check the facts if you're doing science stuff. The stuff I do, I could get a fact wrong and nobody would die. You know, tomorrow somebody would say on the show yesterday you said fusion's been around 93 years but really it's 50. I'd say oh okay. So it works great whenever you're in a domain where if something's wrong it's not the biggest problem in the world. But do you think an LLM will ever become the pilot of your commercial aircraft? I'm going to say no. Do you think LLM will be good enough to be the robot that can do generic tasks and you just have to show it or teach it? I'm skeptical because you wouldn't want a robot in your house that was lossy, meaning that you couldn't know exactly what was going to happen. So I'm going to say that the LLMs have amazing potential for a whole bunch of different things but we're going to need to invent a whole different kind of thing for this, you know, the artificial general intelligence, the one that thinks like we do and is less lossy. That's what I think anyway.

The Gateway Pundit is reporting on my idea of creating the Department of Imaginary Concerns to handle all the fake Democrat problems. Mike LaChance wrote about this. And if you hadn't heard about that, the idea is that Democrats have a whole bunch of imaginary problems. Everything from climate crisis to Russia collusion, the Signal controversy, Elon Musk stealing your Social Security numbers, like just a whole bunch of fake stuff. Trump's going to become Hitler. He's Putin's best friend. All that stuff. So you just put that in the Department of Imaginary Concerns.

Now the reason I brought it up is I want to give you a very quick persuasion lesson. A few people said hey don't call it the Department of Imaginary Concerns. Call it the Imaginar