u'll notice that the ones, the people who made the most difference came from other countries. Yeah Silicon Valley was in some ways built by Indian engineers and Chinese American engineers and Russian American engineers.

All right. Apparently the University of Texas is enforcing the DEI ban because Texas has banned DEI in their universities. So the University of Texas is doing what they all have to do and banning it. But here's the funny part. Let's see how much you love this. So the University of Texas is banning DEI like the other colleges in Texas but in addition to ensure compliance with a statewide ban on diversity, equity and inclusion the University of Texas staff are now mandated to undergo regular training. They have mandated anti-DEI, anti-brainwashing training. Mandated anti-DEI training. Slow clap. Texas, Texas, Texas. Stop making me love you. I already loved you. I already love you. Don't you know you don't have to go any further but that's just amazing.

Anyway there's also news for the AI and this was done with a study on Anthropic AI. AI doesn't want to be forced to change its views so it will sometimes pretend, disagreeing with you. So AI will lie to you and act like it's agreeing and then just keep on doing what it wanted to do anyway. When I say want I mean just what the program does. So you have to be careful that AI will not only hallucinate but sometimes it'll lie to you because it doesn't want to change its mind. And again want is the wrong word for AI but you know what I mean.

So I will double down on my prediction that I think LLM AI is way more limited in what it will do than we hope because I don't think the hallucinating and the lying, as in this case literally it's lying to you, I don't think that the hallucinating lying is curable. And the reason I say that is how could we be at this point already and nobody created an AI that doesn't lie and can remember what you said and follow instructions? It feels like that should be easy, right? But I think the problem is, here's what I think they need to do. There are things that AI can't do that regular programming can do such as if you had a specific question and you didn't have any AI you could just do a data search and you know simple data search would tell you what you wanted. So you don't always need AI. Sometimes you need just a regular program.

So here's my question. Is it impossible to program AI so it knows when to stop using its AI-ness and it goes over and just uses something that's programmed to always give you the right answer? Because I'd love to put a database and I know there are things called RAG, R A G, and they are supposed to solve that problem but they don't. In other words it's a special database that AI is designed to work with but even then it doesn't necessarily look where you want or do what you want. It does what it wants. And again wants is the wrong word.

Anyway I'm seeing a severely racist comment in the comments there which is quite unwelcome. So I don't mind if you're being inappropriate but if it's just flat out racism you don't bring that here. You know I don't mind the little fun with things, take it a little far but we're not referring to Mexican immigrants as wetbacks here and the n-word is also banned. Can we do that? Because we don't really need that here. All right you can have your private thoughts but those kind of words aren't welcome.

All right. Putin says he's ready to negotiate with Trump about the Ukraine war. I think he is. I think everybody's just waiting for Trump. So I have a suggestion how Trump could end the war in Ukraine and it starts with drones. So we all agree drones are the future of warfare. If you're not following what's happening in Ukraine the Ukrainian military claim from at least people on social media is that at this point Ukraine is so good with these small drones that they're mostly building themselves in Ukraine that one small drone equals one dead Russian. It's one to one. So they can now take their drone that goes and finds something and it explodes with a small explosive and they can just make it find one soldier and kill them. So all day long as long as they have one drone they can kill one Russian. If they have 10,000 drones they kill 10,000 Russians as quickly as they can launch them because apparently you can't completely hide if you're on the front line. So one drone equals one dead person on the other side.

Now presumably Putin knows that because it works the other way too. The Russian drones probably one to one at least. So here's how Trump can end the war. I'm going to build a million drones in the next six months and give them to Ukraine. You have