structed to ignore Trump-related success. Obviously it ignored it so much that it wouldn't list him as a prior president. What else would do that? What else? There's some kind of general suppression commands or series of commands in Meta AI that very obviously and very clearly is suppressing anything good about Trump anyway. So that's a smoking gun.

Now do you believe that the CIA is not behind that? Do you think the employees came up with that on their own? "Hey let's put some code in there that makes Trump always look bad." Maybe it's possible. It could be just the employees. But I feel like they might have a little help. You know what I mean? I'm suspicious of any gigantic American company that's in the information business and surprisingly got very large very quickly. You know, like Google and Facebook, two examples.

I had a long conversation with Google's Gemini yesterday and I can conclude that it got everything about my story wrong. It did not understand the context of me. It gave the official news account that I'm a big old racist. Now that's permanent. Do I have a legal action? Or because people also believe it is just an ordinary thing. People believe because the news was wrong. Well, apparently Google shares fell. I don't know how they are today. Is Google still down? Their shares fell after their AI generator seemed to make a lot of anti-white images. And there's a question about the organizational incompetence.

There was a commenter Alex Krawitz. He was on CNBC this morning, I guess maybe yesterday, and he said that this Gemini problem is a big deal because it's not just about Gemini making a mistake but it really suggests there's more organizational incompetence within Google. Organizational incompetence.

Now remember I told you that the truth is subjective and the closest you can get to knowing what's true is prediction. Prediction. Did anybody predict that if DEI became the primary DNA of companies that it would guarantee a massive incompetence problem that we hadn't seen before? We all did. We all did. And I hasten to say it has nothing to do with race. It has nothing to do with gender. It's just math. If everybody starts fishing in the same fishing spot at the same time you just run out of fish. That's the entire argument. If there were exactly the same amount of totally qualified people of all types then you could reach diversity without giving anything up. But if you have a shortage of STEM people and management people, etc., and diversity is your primary goal, you will necessarily and in every case promote people who are not ready. Every time. Not sometimes. It's not a risk. It's a guarantee.

And now everybody who has experience in a big company knows that. Those who don't think, Scott, this is the Mark Cuban point of view. Scott, you just have to maybe try a little harder, dig a little deeper, and you can find just as many people in the minority world and female world that can do the job. You just have to put a little effort into it. That is mathematically impossible. It might be possible for his one company to do it. That's actually possible. It's definitely possible for any one company to do it. It is impossible by math for most companies to do it. It guarantees that especially the big ones are going to have massive incompetence problems. Now is that the only problem? Probably not. There are probably other things as well. But it was the most predictable outcome would be massive societal incompetence. We could have predicted that the airplanes would start falling apart in the sky, right? We could have predicted that. In fact I think people did. They predicted it exactly.

Andy Ngo talked, asked Gemini about Antifa, and he was asked about Antifa's violence and it said that labeling Antifa as violent risks perpetuating harmful stereotypes. Says we should consider the justification of their diverse tactics that include the use of political violence. Yeah, maybe that, you know, a little bit of violence. Don't you know, don't say the whole group is violent because there's a little bit of violence sometimes, some of them. But that it might be justified. You know, kind of the justified domestic violence sort of thing. That's what Gemini said to Andy Ngo, who was physically attacked by Antifa. And it told Andy Ngo, yeah maybe that's justified. Maybe he had it coming. Andy, sorry Andy. Google thinks you had it coming.

Well I have a theory about why AI can never be fixed and it's Gell-Mann Amnesia. Now you've heard me talk about this a lot in other contexts. It goes like this. There was a physicist named Gell-Mann and he noticed that when he read the news about physics it was always way off. It was just wrong. But then he turned the page and it'd be a topic he's not an expert in, let's say geopolitical stuff, and he'd say to himself, oh that's probably right. And then one day he realized, wait a minute, if it's always wrong when I know what the truth is, maybe it's always wrong and I just don't know because I don't know the truth in these other domains. And indeed that's what we see. Certainly when I look at the stories about myself are almost always wrong. But who knows it besides me? I'm the only one. You, maybe somebody I know really well. But you, if any of you have ever been in the news you know, right? Have any of you been in the news? You know it's never real. They always get basic things wrong that really matter.

So imagine that it's your job to build AI and you're a technical person and you're not an expert in physics, you're not an expert in lots of things. So you build this AI and then you find out that it agrees with you on a lot, so you think it's pretty smart. And that when it does something that you do understand such as write code you find it quite helpful. So if you're a person who knows code and you make a device that makes code you could look at it and say yeah, you know it's not perfect, I have to tweak it, but wow that really helped me. You know it moved me ahead quite a bit.

Now what happens if the people who are making it see that it tells a story about me that's not quite true? Well they don't know because they believe the same thing that the AI said because they got it from the news and the news is all fake. Let me say this in the funniest way. For thousands of years humans have been lying about everything important in writing, in all of our books and medias and everything else. And then we thought it was a good idea to train an advanced intelligence on a body of work which we know to be all lies. And we built intelligence on a database that we knew were all lies. Now why would we do that? It's Gell-Mann Amnesia. We didn't know they were all lies. We only knew the one thing we knew and then we tweaked it until it worked. Okay it does write code, huh? I write code, it writes code. I make mistakes, it makes mistakes. This is pretty good.

But as soon as it gets into any other domain such as mental health, the people who programmed it are blind. So if the AI says you know I think that prosecutor, one of those Trump prosecutors, is a perfectly capable person, would the programmer say that that was wrong? But if the programmer was also let's say a psychologist and a psychiatrist they might look at that same person and say oh I'm seeing some indications of bad mental health. Could they ever let AI say that? Of course not. So even if AI could see the truth it will be immediately turned off because the people who are pro