Has China’s DeepSeek developed a “completely different species” of artificial intelligence that’s smarter than anything we’ve ever seen? Or are the rumors all lies generated by an AI bot? Glenn reviews some of the latest terrifying and SOCIETY-CHANGING advancements in AI, including how scientists have developed a new AI tool that can tell your biological age and potential health issues using just a picture.
Transcript
Below is a rush transcript that may contain errors
GLENN: I want to talk to you a little bit about AI.
Apparently, there is a new AI that the scientists at Mass General in Boston have developed a new AI tool called Face Age.
And it can tell your biological age by a picture of you. And apparently, not just -- not just your biological age. But how healthy you are.
In fact, they are -- they believe now, with the eyeball test.
Research in the Lancet digital health. Indicates -- you ready for this?
That artificial intelligence will be able to not only spot that you have cancer.
But also, if you are being treated for cancer.
That's not working.
They have this much time to live.
How terrifying is that?
I mean, how great is that?
How terrifying is that know. Would you have -- if you could have it, and it would tell you, wow. You're looking pretty old and beat up! You don't have much longer to live.
Would you go into the face tool? And say, how long do I live?
How much longer -- I do not think I would do that. I don't want that.
STU: I would want to know if I could do something about it, I suppose.
I mean, I guess, being able to -- if you knew -- and, again, this is somewhat speculative here.
But if you knew, you were going to die in two months, I guess, I -- the idea of wanting to know, would be intimidating.
But I also think, I would like to probably have moments with my family and my kids, and say the things I want to say.
And like, get my affairs aligned, and such.
Start some new affairs.
GLENN: New affairs. Wait. What?
STU: Just kidding, honey.
No, but I would like to get -- if you want to get your affairs arranged, you want to make sure that you're not leaving your family with a burden. You want to make sure you say to your kids, the things you want to say. Maybe you want to write something.
GLENN: Okay. What if you put your face in? Okay. And you take the picture. What do I have left? And it just comes back. What time is it now?
(laughter)
STU: I don't know. If this stupid device can't tell time. It's like a VCR.
GLENN: I don't know. Twenty minutes. Twenty minutes, maybe. Maybe. I don't know. You're not doing well. You're looking a little peaked.
Some of the things that are coming with AI are remarkable. Go ahead.
STU: Life-changing too.
I would say, society changing, probably.
GLENN: So can I just read something to you, that is part of it is beyond my understanding.
And will be beyond something -- and I just want you to hear this.
This is I rule the world MO. This is @IruletheworldMO.
Nobody knows who this really is. They think that this may be an insider, and one of the big AI research firms.
STU: Okay.
GLENN: But they also think it might be a bot by one of these AI research terms to throw the other research terms off. Okay?
They have no idea who this is. Okay.
So, but just listen to what -- I'm hoping it's a bot that is trying to throw people off.
Listen to this. Just got off a four-hour phone call with sources inside Chinese DeepSeek labs.
And holy cow, they're using other language. We are so F-ing behind, it's not even funny anymore.
STU: Hmm.
GLENN: DeepSeek R2, it's not an incremental improvement.
It is a completely different species of intelligence, operating on principles nobody in the West has even theorized yet.
They have abandoned transformer architectures entirely for something they're calling recursive cognition lattices.
The scale in dimension of our math doesn't even have a good notation for this.
The compute efficiency gangs that violate what we thought were fundamental limits like 400 times improvement in reasoning per teraflop.
Not four. Not 40. Four hundred times. Our benchmarks now are literally meaningless.
The scariest part isn't the raw capability.
But how it's developing novel mathematical frameworks on the fly to solve problems. Research gives it questions, and it invents entirely new branches of mathematics to answer them. One physicist showed it a problem they've been stuck on for 15 years. They solved it in seconds with notations nobody recognized. It took three days for them to translate its solution back into standard mathematics.
We saw demo videos that can't possibly be real, except multiple independent sources confirmed.
R2 designed and simulated room temperature, super conductor, from first principles in under an hour, complete with fabrication methods, using existing technology.
They've already produced the samples in the Beijing labs. Blah, blah, blah.
Their -- their in -- I can't say it.
STU: Interrogation?
GLENN: No. No. No. No. Merging with man and machine.
STU: Okay.
GLENN: With biological systems.
STU: Integration?
GLENN: Integration. Thank you. With biological systems is the real nightmare fuel. Two-way neural interfaces that make Neuralink look like a child's toy. Direct cognitive enhancement already in human trials.
This isn't even the most advanced system. They're the ones showing it publicly. America is still treating this like normal technology race, while China understands. It's the an extinction-level transformation of civilization.
It's like watching a nuclear power race, where one side is debating the ethics of gunpowder.
STU: Wow. Because my -- my recollection of the DeepSeek story, when that came out a few months ago.
GLENN: It was nothing --
STU: Experts landed on the idea that there is absolutely -- like they basically were using our technology.
GLENN: Yes. Yes.
STU: It wasn't as impressive as we initially thought. So this person or bot is saying, that it is.
Now, I will say, if you are a person trying to hide your identity. Just saying, you just had a four-hour conversation with a specific company, I mean, how many four-hour conversations happened that day? It would be a weird way to hide your identity unless you're -- so who knows, maybe it's just all blown out.
GLENN: Hopefully it's all blown out.
I mean, you read the -- Sam Altman follows -- others follow. It's not just. It's seen inside the circles. And they don't know who it is, or what it is.
STU: That's true.
GLENN: A post like that makes me think it's a Chinese bot.
STU: Because it seems like it's promoting DeepSeek. Right? That's just amazing.
GLENN: Yes, but he's not always promoting DeepSeek. Or it isn't promoting it. And this is the craziest part. You don't know! Now it doesn't have to be a person!
STU: Yeah.
GLENN: It could be an algorithm.
STU: Now, of course, there is this thing -- you know, there's no -- what is it? There's no -- there's no limit to the levels humans can achieve when you don't care about pain and suffering.
GLENN: Yes.
STU: I think paraphrasing Louis C.K. for that one. Chinese can just throw bows at this. They're doing a lot of human trials on this stuff. As we've seen, maybe with Wuhan in the past. They're kind of willing to do anything. Right? And if they're doing this. And actually seeing these advances, we wouldn't do.
GLENN: No.
STU: You wouldn't be in human trials yet for any of this stuff. Although, a very long ramp up for Elon Musk's company, as we've seen some of that I guess. But they're just -- they'll just throw people at it.
GLENN: You know what's crazy is, we are dealing with technology that we have absolutely no idea, what it's going to be like, what it can do. Nothing, nothing, and I've read several articles, and they're talking about how just everything that you -- everything, the way you work, the way you think, is just about to be completely disrupted.
STU: Yeah.
GLENN: And we're just -- the world is just kind of going along with it. And we're like, I don't know. I don't know.
Maybe we should pass something about it. We're just going along with it.
And, you know, science. There's a -- there's a watch, I think it was Omega. I can't remember. There was a watch that was made in the 1950s and '60s, and its sweep hand, its second hand, it had like a lollipop on it. So it was the stick of the hand. And then it had like this little circle on it.
And it was sweeping around. And the reason why they put the lollipop around it, it was the citizenship signal to the buyer and the wearer. That that watch didn't have radiation in it. And it's not like, oh, you were working in a lab. It was that they were -- you know, we had put, to make things glow at night, that was radiation okay. To get the luminosity on watches. At first, we were like, why don't we just use some of this?
Okay. And that went on for like a couple of decades, you know.
And we were like, hey.
How come his arm keeps losing all of his hair in the first week of wearing that watch.
So they put this little lollipop on it. Yep. No radiation in this one, dude. That's crazy. That we could make.
We could do that kind of stuff. For that long.
And we kind of forget about it. And now, what are we working with? This will make nuclear stuff look like nothing. Like nothing.
STU: We're so close to it, as well, it seems.
GLENN: I read another post, where they were saying that -- that it is getting so fast in -- in for defense, that -- and I said this. I know I said this five years ago.
That you won't even know that you've lost the war. Because for you, the war hasn't even started yet.
But you will start and lose the war, in a flash.
And the time it takes you to go. Wait a minute.
There's a war going on. What?
You've already lost. It happened. And you've lost.
Because AI is going to get so good.
It will predict absolutely every move. That everybody is going to make.
And it will just go, oh, here's the countermove. And put it in.
And go, okay. Well, that's over.
STU: It's like when you -- you're not a big video game guy.
But when you start praying a game. And you just decide to go on the toughest level of the opposing AI.
And like, you just can't do anything.
You just automatically. Your base is destroyed in seconds.
That's a very -- very low level version of this.
GLENN: Yeah.
STU: Let me -- can I give you one other thing on AI.
I think this is fascinating. This is in the Wall Street Journal two weeks ago. I haven't heard anybody talk about it.
We've talked about it for a long time.
The surveillance state. I remember with you, taking calls from people. Going, I will not get an EZ Pass because that means they can track me when I go through the tolls. Right?
GLENN: Remember when we were in Tampa. This is the year 2000. They put cameras up in the streets of Ybor City. And everybody was like, I will -- not in America. Not doing that right now.
Nobody was willing to give their fingerprints. Nobody wanted to give their face. None of that.
STU: Yep. We all carried our phone. Of course, GPS everywhere we go. We click yes, agree. Agree. Agree to everything.
GLENN: We open the phone with our face.
STU: Yeah. Listen to this. This is amazing. This is from an author Joanna Stern.
GLENN: Wall Street Journal.
STU: Wall Street Journal. I've been wearing a wire everywhere since February.
That's how the article starts. I've got all the transcripts, important meetings, arguments with my kids, chats with disgruntled employees, late-night bathroom routines. There's plenty more I can't share, if I want to. And my bosses and my family as well, to keep liking me.
No, not an FBI informant. I willingly wear a 50-dollar bracelet that records everything I say. And uses AI to summarize my life. And send me helpful reminders.
GLENN: Why would you do that?
That's called a panopticon.
STU: Uh-huh. I think we're basically there. How do you have a private conversation in this world?
She tested two other devices as well, that are on the market now for 159, $199. They recall every single thing.
They transcribe every single thing. They have recordings of every single thing that was said by her or around her.
GLENN: Let's try it --
STU: That's crazy. Crazy.
GLENN: Crazy? No, it's crazy not to try it, to show everybody how bad it is, but it would be crazy to do it and be like, I think that's going to be a great addition to my life.
STU: Hmm. She says, within hours of wearing this bracelet, I was blown away at how quickly it turned ramblings and random chatter into useful, actionable information, yet allowed me to quote myself from February 24th at 5:15 p.m. This bracelet is really F-ing creepy.
Apparently said that out loud.
But, I mean, you can see. Again, I can see a world where that probably would be beneficial. You have a conversation with someone about something. What did they say? You would have it. When you were saying, hey. We should get together next Thursday.
It puts something in your calendar. That says, hey, call this person about that Thursday meeting you discussed. Of course, that would be beneficial in some way. It's like having an assistant. If you are an executive, you have an assistant.
GLENN: Those bots are already -- by the end of the year, those will be strong everywhere. You will have that assistant doing that in your phone and everything else. It will already do that.
STU: This is the death of private conversations though. They're over.
GLENN: Yeah.
STU: Every single time you have a conversation, you should act like you're on television, having it.
GLENN: Yes.
Well, we've lived that way for a long time.
STU: That's what made me so interested.