n a week? A lot, right? Or there are a lot of situations where you want a service, but they're booked up so you can't get to them for a week. Well, if you add the robots, maybe it's the same workers, but they can do three jobs a day instead of two. So suddenly your human plumber is making 50% more. All right, you agree?

So there's another AI company that's being skeptical about the robots. Let's see if I can find that story. Maybe I didn't write that down.

But all right, here's a story. So I think this was in the Wall Street Journal today that the headline is even the companies making humanoid robots think they're overhyped. So you know what I've been saying for forever, so I believe I was ahead of the curve, is that if robots could do more than one thing, they would already be deployed. But we keep seeing these demos where the robot is trained to do exactly one thing, like iron a shirt, but that skill does not go to anything else. You can't just give it an AI brain today and have it figure out stuff. You would have to specifically train it for every little task.

So that's why you do see robots in warehouses because a warehouse is a very limited training set. If you get this command, go get this box, carry this box over here and put it there. So it's such a small domain that a robot could work in a warehouse. But I've been saying for a while if you have any experience with the large language model AIs, which are the dominant parts, that they don't really have any hope of becoming general intelligence. There's nothing that they're doing, the AI industry. I don't believe there's anything they're doing that could logically lead to a general intelligence robot. And I think that some of the experts are saying it too. Now that they've all been overfunded, they can tell the truth.

But there's this one guy, I guess he's a head of an AI company called, I don't know, his name is Velici and he's skeptical. So he says the same thing I do that you can train them to do warehouse tasks. But did you know that for every $100 you would spend on a robot, only about $20 of that is a robot and the rest of the money is for protecting humans from the robots. Now, I think that means physically so that the robot doesn't accidentally run you over or something.

So here's my question. Elon Musk is very pro robot and I wonder what does he know that the other top people in the industry don't know. Now, it's always a bad idea to bet against Musk when he's making a prediction of the future. You know, he might be off by timing, but it's a bad idea to bet against him. He very clearly believes that the Optimus robots will be general purpose robots. And apparently his version of AI is way ahead of the other AIs. But is he way ahead in a way that would give us general intelligence or is he just way ahead in the way that we can never get general intelligence? You know, maybe his will hallucinate less or something, but what does he know? Is he going about it in a completely different way?

Now, I know that he trained his cars with video, but again, that's a limited domain. So as varied as the possibilities are for a car, I mean there's a trillion different things that a car could have to do, you could probably get to a trillion. So I think you could train a car on a trillion different possibilities and then it would be better than a human, but it would still be trained for a narrow domain, which is driving safely. So I get how we can get to self-driving cars. That makes total sense. You just train them with video instead of language.

But how do you get to a butler? How would you ever get to a butler where there's something new every day that it's never seen? What does he know that we don't know? Does his AI have a secret skunk works that nobody knows about that's getting close to it?

And I've said this before, but it's a really good tip. If you think that you'll have a robot butler in one year, you would already see it, right? Because I can't believe that the robots would be launching in one year and they didn't know how to do it yet. Does that even make sense? What do you think? There's any possibility that a big serious high functioning company would say oh yeah in a year we'll have robot butlers but we have no idea how to do it at the moment. No big company would do that. Not Tesla, not anybody.

So if you're not hearing today that they have one in the lab that has general intelligence, and you're not, I don't believe there's going to be one in a year. I do not believe. So I want to be wrong. I want to be wrong. How many of you agree with my assessment that if they can't do it today, it's not going to be a product in one year? Woul