Jacob Ward (3:19)

Leo, thank you for having me. I was saying to you before we came on three hours, brother, I don't know how you do it, but I also think, I'm not sure we're going to get through a whole week. This week in particular, in three hours.

Leo Laporte (3:29)

There's a ton of stuff.

Jacob Ward (3:30)

Yeah, this is my first appearance. This particular. Why couldn't I have picked a whole week where we.

Leo Laporte (3:37)

It's funny because I thought this was going to be. I thought this was going to be a light week and then you reminded me, oh, the signal thing happened this week and it was like, oh, gosh, this has been a long week. Ian Thompson is also here from thereegister.com hi Ian, great to see you.

Ian Thompson (3:51)

Good to see you as ever. How's it going?

Leo Laporte (3:54)

Yeah, it's going, it's going well.

Ian Thompson (3:57)

You say old friend. I was at a, I was judging a virtual reality conference at MIT a couple of months ago and someone came up to me. It's like, you're Ian Thompson. I've been watching you on Twitch since I was a teenager and you were.

Leo Laporte (4:08)

Just like, isn't that great?

Ian Thompson (4:09)

That's both great and also, God, I'm old, you know.

Leo Laporte (4:13)

Ian, in two weeks is our 20th anniversary show.

Ian Thompson (4:16)

Fantastic.

Leo Laporte (4:17)

We started April 14, 2005 and we're going to do something special on that show. In fact, I've already started Receiv's videos and somebody wrote a poem. Your memories of the first time you watched Twitter, where you watch it, things like that. We want to kind of honor our long suffering audience. Next two weeks on April 13th. Also with us, great to see him, Alex Kantrowitz, the big technology podcast and newsletter. It's always great to have you on, Alex.

Alex Kantrowitz (4:47)

Oh, it's good to be here, Leo. Thanks for having me.

Leo Laporte (4:49)

Such smart stuff you're doing at big technology. You know, it's kind of, it's kind of cool when two of you are former mainstream media guys who moved to, you know, be creators and the third is at the Register. You're the guy still in mainstream media. Ian, when are you going to start your newsletter?

Ian Thompson (5:12)

Honestly, I just enjoy being a writer again, you know, being a journalist. Yeah, you going after each story again. You know, on a. You wake up, there's always something new, there's always something you can get out there and it's great.

Alex Kantrowitz (5:23)

Hey, you know that independent, by the way, so you should join us. Hopefully we'll Convince you by the end of the show to go independent, start your own thing.

Jacob Ward (5:29)

I will say I. I miss teamwork, though. I miss, you know, I miss being able to call a Washington, D.C. bureau and saying, hey, I need a senator to talk to. I miss, you know, I miss the teamwork way, I think. You know, I love. I love so many things about this model. But. But putting people together to, you know, especially in the face of this maelstrom of news, have the idea that all of us can be writing about what, whatever we saw on the Internet. And that's going to, like, save democracy. I don't know, you guys. It is.

Alex Kantrowitz (5:56)

Three of us have your back, Jake.

Jacob Ward (5:58)

Yeah. Good. Well, great. Good to be with you then.

Leo Laporte (6:00)

Yeah. This is the new. This is your new team. Although we're all frenemies because we're all kind of in competition with one another. Limited number of eyeballs, and you can't call me and ask to meet a senator, so that's right out. Alex knows a lot of big names. You're. Alex is amazing. Getting the big names on his podcast.

Alex Kantrowitz (6:21)

I'll hook it up. I'll h up. Just hit me on the dms. We got you.

Leo Laporte (6:26)

So this was the week that Studio Ghibli burst onto the scene. Miyazaki's famous movies are beloved by fans, but now he's being turned into an AI meme. OpenAI added the ability for ChatGPT4O to create images. Actually been playing around with this a little bit, and for some reason, people really wanted to do Studio Ghibli stuff. This is the disaster girl. You know, the girl in front of the house on fire in the Disney style. She kind of looks like a Disney princess. Here's a Studio Ghibli grumpy cat. Very nicely done. This was less good Mona Lisa as Vincent van Gogh. Because you don't have. I think people might think, oh, it's a Studio Ghibli filter from Chat GPT. No, no. It can do anything. I even had it do Ian Thompson with big eyes.

Ian Thompson (7:28)

Yeah, aren't you very big?

Leo Laporte (7:30)

They did get the shirt right, though. Nice plaid shirt. And then finally Rembrandt and the twit logo. Nice brushstrokes. And very dark.

Alex Kantrowitz (7:38)

Very dark.

Ian Thompson (7:39)

Very dark indeed.

Leo Laporte (7:40)

But now what I want. Benito, can you give to a screenshot of the four of us? It would be great to have a screenshot of the four of us and before this show is over, have a Studio Ghibli or Ghibli. We've decided it's Ghibli version of this.

Ian Thompson (7:56)

All right, everybody smile.

Jacob Ward (7:57)

All right, everybody smile.

Leo Laporte (7:58)

Okay, Smile. Open your eyes wide.

Alex Kantrowitz (8:02)

All right.

Leo Laporte (8:02)

Now one of the things that happened immediately to me is that I got rate limited and actually Sam Altman says you guys are burning down the servers. But there is a larger question. If you're Miyazaki, you might be a little offended by this. In fact, Benito pointed out, and he's probably right, that it's only a matter of time before somebody does full length movies this way.

Jacob Ward (8:27)

I mean, I think that's. I've had a conversation. I'm sure all you guys have spoken to developers in that world. I was talking to somebody the other day at UC Berkeley who's working on. On something in this world. And, and you know, the, the. I said, how. How long before we do a full 90 minute movie? Because right now the trouble that they're having is they can't keep the same look of the characters scene after scene of scene. It sort of migrates over the course of it. And he was like, within a year we'll have 90 minutes. And that was, you know, I mean, and to me, I mean, like you say, Leo, like I. This company just can't stop walking headfirst into the. The most, you know, the, the. The. The weirdest interpretation of how art is supposed to be consumed and, and grab. And. And how tone deaf the grabbing up of it is. Like the, the Starlink or sorry, not Scarlet Starlink. Help me out, you guys. Oh my God, I'm blanking on it.

Alex Kantrowitz (9:19)

Scarlett Johansson.

Jacob Ward (9:21)

Well, there's Scarlett Johansson.

Leo Laporte (9:22)

I often confuse Starlink with Scarjo. Yeah, no, no, not.

Jacob Ward (9:26)

What's the one.

Alex Kantrowitz (9:27)

They're both in the stars house.

Jacob Ward (9:30)

What's the Stargate? Stargate.

Leo Laporte (9:32)

Oh, you know what? I don't blame you for being confused. That's the half billion. Sorry, half trillion dollar AI effort with Oracle and Sun and they name it for that.

Jacob Ward (9:42)

For the show. Right. And. And the conceit of that show is that an amazing alien technology that transports people between worlds is in fact being used by a super annoying race of enslavers grabbing slave labor from other worlds. And it's just that thing where you're like why would you. Why would you call it that? And in this case you're grabbing Miyazaki, the most beloved. You know, if you're a young person who likes to be alone in your room and live in a calm world of beauty and think about, you know, non technology things to have this as the, the way OpenAI shows off this Latest twist in this technology. To me, it's just the most tone deaf version again. Or, you know, or maybe that's video game art. Don't do Ghibli, you know, or maybe.

Alex Kantrowitz (10:30)

That'S the point though, because, you know, I mentioned Scarlett Johansson. She's on my mind because, well, she's.

Leo Laporte (10:36)

On all our minds, but like the Roman Empire, it's just.

Alex Kantrowitz (10:39)

No.

Leo Laporte (10:39)

From a condition of being a male.

Alex Kantrowitz (10:41)

Okay, no, but I'm talking about it from the tech standpoint that when OpenAI released the voice chat and open it was.

Leo Laporte (10:48)

Oh, yeah, she was pissed, she was upset.

Alex Kantrowitz (10:51)

They tried to work with her, she said no. They basically copied her voice. Anyway, Sam Altman tweets her, this is in May 2023 or 24.

Leo Laporte (11:01)

Tweets the word her running right into the spikes.

Alex Kantrowitz (11:05)

Exactly.

Leo Laporte (11:06)

Yeah.

Alex Kantrowitz (11:06)

So she says, all right, I'm going to sue you. Guess what happens. Everybody starts paying attention to the fact that OpenAI now has voice that's so realistic and sounds so much like Scarlett Johansson that it must be good. And then you go from a moment where OpenAI stagnated. Basically they hit 100 million users in February 2023. They don't crack that until about April or May 2024. So basically a year of stagnation. And once you found out that they had made this voice that was as good as Scarlett Johansson in her, the numbers boomed. OpenAI went from 100 million users to 400 million. I think that this, what they're doing now, the fact that they are amplifying, because they are amplifying the Studio Ghibli Miyazaki stuff is the same exact thing. I think they lean into the controversy as a marketing tool and it's absolutely working for them.

Leo Laporte (11:56)

I mean, in their defense, I don't. Did they mention Studio Ghibli? I. I think some, somebody in our chat is saying that was emergent behavior that just, that that went viral independent of what they.

Alex Kantrowitz (12:09)

Sam Altman's profile picture right now, look is an anime Sam Altman that looks a heck of a lot like a Studio Ghibli.

Leo Laporte (12:15)

Well, he's leaned into it, right?

Alex Kantrowitz (12:17)

So it's not like they're running away from it. And in fact, I would say that they're encouraging this. And, and, and it's. Again, it's working. Now. This is a tweet that Sam Altman posted just before we came, we came on the air. He said there's biblical demand. I have never seen anything like it. And he's begging people to stop creating images because they can't do anything. And in fact, someone was like, well, when is AGI coming? And he goes, as soon as you stop creating images, it will come a little bit sooner than you are right now. So I think that they understand slowing.

Leo Laporte (12:46)

Us down, says Sam.

Alex Kantrowitz (12:48)

Pretty much. I think they understand quite well what they're doing, which is that when they make these VI moments, they cement ChatGPT as the number one slash, maybe the only AI tool with mass adoption right now. And they need that to happen because we all know what happened with Deep Seek a few months ago, right? The entire foundational model layer has been commoditized. So actually, this image creation and the voice stuff, that's all they have, because it brings people into ChatGPT and, you know, say what you will about whether it's tasteless or not, you know, I think there's an argument to be made either way. The fact is that this is working and it's a big moment for them.

Leo Laporte (13:27)

Actually, there is.

Jacob Ward (13:28)

This is. As usual, this is me walking headlong into being the naive guy who thinks that art should be protected against these people. Because I totally.

Leo Laporte (13:37)

I don't know if you're naive.

Jacob Ward (13:38)

I think you're absolutely correct. I think it makes perfect sense, right? From an outcome perspective, from a strategy perspective. It makes perfect sense, right? Get the attention you want. But, you know, meanwhile you see all these, you know, these like, animators on social media literally weeping into the camera, saying, like, this is the end of it. This is it.

Alex Kantrowitz (13:54)

But, I mean, that's also a great way for them to get airtime. And let me ask you this, Jacob. What do you think about fandoms? Because we had a story in the prep doc that Leo and Benito sent around right beforehand, and the headline is, heyo, Miyazaki's AI Nightmare. It's from the Atlantic. And I'm reading this story and I'm prepared to read an op ed about how terrible it is that they're ripping this guy off. But we all know that films and studios like Ghibli have benefited from fandoms, people that are obsessed with what they make, and they make derivative work, and that continues to contribute to the popularity. Now, I definitely hear the artists, and I know Miyazaki is one who's come out against AI and I hear the objections, and I'm not going to diminish them. Like, I think that's a legitimate objection, but there are two sides of it. Because when I'm reading the story about Miyazaki's AI Nightmare, I'm Seeing the case being made that fandoms are propelling this type of work, and maybe that's going to be the case here. I mean, right before we came on air, we're trying to get the right pronunciation of Studio Ghibli or Ghibli, whatever it is. The four of us would never spend a Sunday afternoon trying to figure out how to pronounce the name. It just isn't relevant in most conversations. And now it is, I guess I would say.

Jacob Ward (15:07)

I mean, I take your point. And I think that the idea that in this particular historical moment, as we are seeing companies like OpenAI get attention by exploiting the creative work of other people, it is going to, I think, have a kind of flywheel effect on the fandom of that, of that group. I agree with you on that. But I also think the longer term concern, and when I think longer term, I'm talking 12 months, like after this historical moment. I think it's an open question whether we're going to have anything like a Studio Ghibli or an Aardman Entertainment. The guy who does Wallace Gromit. Yeah, Wallace and Gromit, exactly. You know, like. Like the market won't be there for people to make those kinds of original, handcrafted, beautiful kind of IP because it'll be so much cheaper and easier for. For kids to just pull up whatever they want. I had this conversation with Justine Bateman, who, you know, who from Family Ties, who went on after her career in that to get a computer science degree. And her big argument during the saga after strike was, if we let these studio execs have their way with AI, they're just going to make sort of DIY movies such that you sit down, you say, I want something with Harrison Ford and make it a Viking thing, a little bit of sex, but not too much, and no one will ever hire another new actor again. So I think you're right that in this moment existing IP and existing actors and Studio Ghibli and the rest of it will benefit to some extent. But I just worry that the long term effect. I'm curious what anybody else thinks about this is maybe a world in which we're not. We don't have anything new to sample. We're just rehashing the same stuff.

Ian Thompson (16:53)

I do feel in a way that you raise a very valid point indeed. And it's something I've noticed over the last 10 years ago. I mean, if you look at sort of like the top 10 films throughout the last decade, there's very little New stuff and a decreasing amount of it. It's all rebound boots or using an existing feature. And I think AI is just going to turbocharge that completely. And I'm sure a lot of the studios looked at it just like, yes, we can get rid of those pesky actors that cost us so much money, but the fact is they also turn in great performances, which AI not so sure it will.

Alex Kantrowitz (17:27)

Yeah. Then the AI replicating human performances is going to take a while. But I would also say if we're trying to think about the total creative output that we're going to have on this planet, just think about what it takes to create an animated film today. So this is from Derek Thompson, Mr. Abundance himself. You know, he cited in a tweet last week that Disney can't make an animated feature for less than $200 million. And right now we're seeing this stuff actually work in video as well. Now, is that a threat to studios like Disney? Sure. But is it also an opportunity that you can start making stuff for less than that 200 million and I don't know, may. Maybe they. Maybe the best creative output will rise. Like if you're Disney and your entire moat was that the everyday person couldn't create what you can create and it didn't matter what the story was. I would say that you're failing. Disney's great at story. So potentially they have an advantage because now there can be more creative proliferation at a lower cost and they're in the driver's seat to lead this change. But they're now able to do their specialty and at a cost maybe at a fraction of what they were paying previously.

Leo Laporte (18:37)

So 30 years ago, Steven Spielberg was making a dinosaur movie called Jurassic park. And he had Stan Winston, the great model maker, making a giant T. Rex, a physically giant 8 foot tall, or, I don't know, maybe even taller T. Rex. But at the same time, as they were preparing for this movie, Spielberg happened to walk through Industrial Light and Magic where two programmers had created in the computer a T. Rex. Spielberg saw it and stopped Stam Winston. And that was the first example of CGI used 30 years ago in a film. I'm sure at the time the model makers were pretty upset. Star wars was entirely practical models, but a new way of making movies was created. And I, while I'm not a huge fan of CGI in movies, it's in every movie, whether you know it or not. And it has changed filmmaking without destroying filmmaking. I don't think that an AI created Studio Ghibli movie, or even a real life action movie featuring Indiana Jones would be as good as one created by humans, acted by humans. With all due respect to Justine Bateman, I think that's giving a little too much credit to, to the AI I don't think it's, it's the same. Humans have something they do that's very different. Coincidentally, Studio Ghibli. This is from Business Insider. Probably cannot stop what OpenAI is doing. Business Insider interviewed IP attorneys who said, you know, the juries or the courts are still out on this one, but generally copyright laws allow artists to mimic a visual style. It's not considered stealing. And in fact, when you talk about Disney, Disney was copying the Brothers Grimm, probably. Studio Ghibli was heavily influenced by the early animation art of Disney. And there will be another generation that will be heavily influenced by the animation art of Studio Ghibli. This is the process. We copy each other. No human does it all in a vacuum.

Ian Thompson (20:46)

I do find it kind of amusing though that we've just had Napster being bought by somebody and you remember Napster when all the studios were just like, intellectual property is everything. We must protect it at all costs. And now they're ripping it off for this. So, you know, strange how the way these things turn around.

Leo Laporte (21:00)

I'm more sanguine than you are, Jake, about the future of art. I do think artists and creators and humans will still do what they do. Although Miyazaki says AI is anti human, he is not a fan.

Alex Kantrowitz (21:12)

Yeah, he calls it an insult to life itself.

Leo Laporte (21:15)

Yeah, an insult to life itself. Yeah.

Jacob Ward (21:17)

I think that this, like I, I, I just think I, I always try whenever anyone says it's not good enough yet to do, you know, a true simulacrum of the, of the human experience of watching Harrison Ford or whatever it is. I just always say give them two financial quarters and they'll figure that stuff out. Like the quality is going to go up and up and up and up and up. As we've already seen, it's exponential. You know, I wrote this book, the loop, predicting the AI mania. I thought I was like 10 years ahead at the time that came out.

Leo Laporte (21:45)

When did you write that book?

Jacob Ward (21:46)

It came out nine months before ChatGPT did.

Leo Laporte (21:48)

Yeah, even OpenAI was shocked. They did not expect Chat GPT to be successful. So you're right, we're moving at a rapid pace. You know, the only real question is that last bit of it, right? You look at self driving cars, they do really great 90% of the situations. But then something Weird happens and they. They can't handle it. They drive through a paper wall.

Alex Kantrowitz (22:12)

This is exactly the point, though, which is that we're going to see humans and the machines work side by side to produce better outputs. It's not AI Necessarily disintermediating. Disney is Disney with AI. That's going to be great. I mean, just going back to your Jurassic park example, was it. Did people walk away from that movie saying, oh, wow, that CGI dinosaur was really special. I'm glad we watched that movie because we saw the CGI dinosaur.

Leo Laporte (22:40)

I don't think they even knew.

Alex Kantrowitz (22:41)

No, what they said was, it was a great story.

Leo Laporte (22:43)

Right.

Alex Kantrowitz (22:44)

And ultimately, it doesn't matter what technology you have. If the story is great, if it touches people's hearts, people will watch it animated or not. And honestly, it's a challenge for us humans. We have to be able to keep our ability to reach other human beings in a way that's better than the robots can. And if the robots can only get 90, 95% of the way there, then I think our creators are in great shape because we can get 100% of the way there. We know we can do it. Ghibli does it. Miyazaki does it. So many people do it. And if you give the people that are able to convey that message and tell those stories and touch the hearts, this technology, then you are looking at a great democratization of what we do and an opportunity for a creative explosion in the best case.

Jacob Ward (23:34)

I certainly don't want to be like, I spend my entire first appearance on this great show being like, Mr. Anti Tech. I promise you, I have lots of positive opinions about technology, but this is one that is definitely.

Leo Laporte (23:47)

You're allowed. Jake, this is a debate that we have been having for the last two years. Great, okay. And it's unresolved. It's really an unresolved debate.

Jacob Ward (23:55)

Even it's right. And we're not going to figure it out in these three hours. But I also think that it is a. It is a. It is a really fundamental thing we got to keep talking about. And one of the things that came up for me recently is, Alex, to your point of, you know, the, the, you know, and Leo, you made this point too, that, that, you know, the technology will take us 90% of the way, and then that last 10% maybe will be where the humans get involved. And, and that's where the sort of the human magic can happen. And, you know, the dream that, that these makers always talk about with, you know, humans and robots working Hand in hand. And with that up to a point, there was a study that came out from a big pharmaceutical company when ChatGPT first came out. And they, and they began using a proprietary AI package to try to synthesize new chemicals and figure out, you know, new potential products. And it was incredibly effective at making people more productive. And it, but the distribution of productivity was very, was very much like this. This top tier of very skilled scientists used it well and, and everybody else got kind of dropped out the bottom because they couldn't really adapt to the technology and they couldn't really figure it out. So that, so the projection was that that top 10% was kind of the group that would survive. But here's the part that really brings me up short on, on this hand in hand quality is that everybody, and especially the top 10% who are using it the most, intensely, reported deep dissatisfaction with the job. They hated being in the role of catching the AI's work and getting it ready for prime time. And I worry that there's something in. And I'm not here to defend Disney like that's not it. But Studio Gibliet, I think, is worth defending. And I'm not sure that the people that, you know, suffer and suffer and suffer to be in that world are going to be okay with or are going to be willing to get into that world to just take a computer's work across the finish line.

Alex Kantrowitz (25:48)

I just build more computers to do that job also. So.

Jacob Ward (25:51)

Yeah, that's right.

Leo Laporte (25:52)

Well, I love it that Sam Altman is begging people to stop.

Alex Kantrowitz (25:59)

On these images.

Leo Laporte (26:00)

It underscores the problem that they have, is that the more successful they are, the more it costs them.

Jacob Ward (26:06)

Yeah.

Leo Laporte (26:06)

And it's a very weird situation they're in. Incidentally, it is not widely agreed necessarily that Studio Ghibli's art is the great, greatest art. In fact, at that Atlantic article, Ian Bogost says when asked, are the Studio Ghibli images evil? He said, I don't think they're evil. They might be stupid. You could construe them as ugly. They're also beautiful. He's not a fan of the aesthetics of Studio Ghibli, to be honest.

Jacob Ward (26:40)

I got a quick, quick comment about that. Is that, like, people are more, more hijacking the idea of Ghibli. It's not like the actual style. It's like what you think when you.

Leo Laporte (26:49)

See a spirit, big eyes.

Ian Thompson (26:51)

Exactly.

Leo Laporte (26:51)

So they're hijacking what you pull out.

Jacob Ward (26:53)

Of a Ghibli film. They're Hijacking that onto there.

Leo Laporte (26:56)

Yeah, but if you. If you watch, you know, Spirited Away, it's. That's an experience. At some point, you don't even think about the illustration.

Jacob Ward (27:04)

Exactly, Exactly. It's not so much about the illustration itself. It's more about other stuff that comes with it.

Ian Thompson (27:09)

Yeah.

Leo Laporte (27:09)

Story. And it was the same thing with Jurassic Park. It isn't about the cgi. It's got to be good enough that you believe it. Those dinosaurs are running after those guys. But that's all. And it was. And you believed it, and the story worked. But I will.

Jacob Ward (27:22)

I mean, I will point out to Ian's point of everything being a reboot, you know, an Avengers reboot these days. Like, no one's going to those movies for the story. You're going to the movies to see the explosions and.

Leo Laporte (27:32)

Yeah, that's true.

Jacob Ward (27:33)

And the spaceships.

Alex Kantrowitz (27:33)

But that. But. Yes, but you trust in the brand because the brand has told you an overarching story. Story over time. Like, if it was just the explosions without any kind of story arc, you're not watching that. And by the way, these. These. These images are so ugly that they've just taken over the Internet. I don't. Yeah, I don't accept that at all from me.

Ian Thompson (27:54)

I don't know. The fact that Michael Bay still has a career suggests that people will just go to see the explosions.

Jacob Ward (27:59)

Hey, Michael Bay's an artist.

Alex Kantrowitz (28:01)

Point. I'll take Leo. I mean, just for the next two hours, make things explode and test the hypothesis.

Leo Laporte (28:09)

We could. I'd probably make more money, to be honest with you. But, yeah, this is the kind of. This show is the antithesis of what the market's demanding. People thinking and talking about things with no explosions at all. It's probably the wrong way to go anyway. Is this a breakthrough in image generation? No. Right. In fact, I had just seen some Homer Simpson AI art where the Simpsons characters as real people, and I thought, oh, that's cool. And I was looking to pull it up, and then I saw that they were doing the same thing three years ago. I don't know if this is really a breakthrough in any sense. Midjourney's been doing it. Stable Diffusion has been doing it for a few years.

Alex Kantrowitz (28:55)

But it is just better than any of the other stuff. If you ask Mid Journey to draw you an image of Tim Cook and Mark Zuckerberg, it will blend both CEOs and give you two people who are.

Leo Laporte (29:07)

Like, part of that is a safety thing, I think. Alex. I think. Yeah.

Alex Kantrowitz (29:10)

So go ahead.

Leo Laporte (29:12)

I Just think that the real names often are rejected by these tools.

Alex Kantrowitz (29:17)

Well, I think it's both. So I think this is an advance from OpenAI. And you have this story from Ethan Mollick that I dropped in the document on screen, just talking about how this is something that's being created by the model itself, not just sent out an interpretation, then dropped back in. So it's native to the large language model, and that's what's been able to do things.

Leo Laporte (29:39)

The title of it is no Elephants, because if you ask an AI to create an image with no elephants in it, there almost always are elephants in it.

Alex Kantrowitz (29:49)

Exactly. So it's listening now because it's in the model. But the other side of this is.

Leo Laporte (29:55)

It may also be that humans have tuned it that way. Right, Right.

Alex Kantrowitz (29:58)

Well, they've been dumb, they've been pretty stupid, these models, but they're getting smarter. And you're seeing that Ethan is also saying, create an infographic. And he's giving it the exact text and it's pulling it off ye. And he's telling it to have an otter hold that infographic. And it can do it. And he's saying, have the make the otter an action figure. And it's doing that. So it's. It's definitely.

Leo Laporte (30:19)

It's pretty impressive. Yeah.

Alex Kantrowitz (30:21)

But the fact that it's smarter and combined with a pullback on safety or refusals, if you're, if you want to take that perspective, that, I think is what's made this launch so dramatic, is that it's both a better model and it will refuse you less. Like, for instance, we've shown it imitate famous studios. We've shown it imitate the Simpsons. We've shown it put celebrities and public.

Leo Laporte (30:43)

Figures stuff that it was barred from doing in the past. Right. Correct.

Alex Kantrowitz (30:48)

It was not doing it before. So what OpenAI has done here is it's shown a real advance in its artificial intelligence by making this native to the large language model. And it's taken down those safety precautions and basically let the Internet go buck wild. And that's what we have as a result. And that's why we're melting OpenAI servers right now.

Leo Laporte (31:07)

It is kind of a her moment. You're right. It's OpenAI saying, screw it, let's make it sound like Scarlett Johansson. They did hire an actress, not Scarlett Johansson, to create that voice, but she sounded just like Scarlett Johansson.

Alex Kantrowitz (31:20)

Yeah. At the end of the day, this.

Jacob Ward (31:24)

Is the very moment that Google and OpenAI are going to Washington and asking the Trump administration to Basically take all legal restrictions off them when it comes to copyrighted work. They basically are saying we should have fair use on absolutely everything.

Leo Laporte (31:38)

And I think the administration will be open to that, don't you think? I think this is a sea change in Washington about all of this. The Biden administration really wanted to stop AI a little bit and really focus on safety with David Sachs in as the AI and crypto czar in the White House. By the way, David Sachs, compadre, and I may regret this. Jason Calacanis will be on the show next week, and I will ask him. I will actually ask him. Well, I'm gonna ask him a lot of things, but among other things, I will ask him what the administration's plans are for AI because I think that what's happened is Trump has been convinced by Sachs, Marc Andreessen, Sam Altman, and others that if we don't do it, the Chinese will. This is so analogous to what happened in 1939 when the physicists Albert Einstein and Leo Szilard wrote a letter to Roosevelt saying, you got to develop an atomic bomb because the Nazis are gonna. And Roosevelt took it to heart, created the Manhattan Project. We have now created the equivalent of 10 or 20 Manhattan projects in terms of spend and effort put into it to create AGI. And I think that Trump's been convinced that if we don't do this, that the Chinese will, and if they do, we're in trouble. Do you agree?

Jacob Ward (33:05)

I'll be really curious to see. I mean, when you have Jason Calican, I would really want to know, like, did the gamble pay off? Right. Because you saw Andreessen and all these.

Leo Laporte (33:14)

Guys, they bet on the word. They bet on the new administration. They did.

Jacob Ward (33:17)

That's right. They bet hard on this new administration. And did it pay off? Because you're seeing other people who bet hard on it not have it pay off. So Zuckerberg, Right. Who put a million dollars of his own money. Yeah, all that stuff, you know, he's still gonna face the music with the ftc, supposedly the new ftc.

Leo Laporte (33:33)

Yeah, we'll get to that. Yes.

Jacob Ward (33:34)

Head says, you know, he's gonna. He's gonna go after anyway.

Leo Laporte (33:37)

Yeah.

Ian Thompson (33:38)

I mean, Google as well. Yeah. I mean, they. They bow down and kiss the ring. And now they're in legal trouble.

Leo Laporte (33:44)

So it's not enough to kiss the ring. You've got to give the ring a quarter of a trillion dollars. I think we've learned that.

Jacob Ward (33:50)

That well and even. And we just don't know, moment to moment, what, what how the mood is going to change from the very top. So we don't know.

Leo Laporte (33:57)

I don't know. Yeah, I don't. You know, I would. I don't think Jason is a full maga, so it's going to be very interesting. He was for a long time very close to Elon Musk. And when Musk bought Twitter, he was one of the cadre of people brought in to. To whisper in Elon's ear. Now Elon's whispering grande for him. Yeah. So I, you know. Yes, we'll have lots of questions for Jason and he knows.

Alex Kantrowitz (34:23)

I was in favor of letting these AI generators run wild and take the safety controls off, but somebody in the Discord turned the four of us into Muppets. And now I'm in favor of.

Leo Laporte (34:35)

I knew that this is what I wanted the screenshot for. I knew the creativity of our club Twit members would take over. And do we. Are we good looking Muppets?

Jacob Ward (34:46)

Oh, I bet.

Alex Kantrowitz (34:46)

No.

Jacob Ward (34:47)

10 to 1.

Leo Laporte (34:51)

Why am I blue? And why is he in a dog? That's what I want to know.

Ian Thompson (34:55)

That's a little concerning. Yeah. Okay.

Alex Kantrowitz (34:57)

I mean, I look like Beaker. I'm not very happy about this.

Leo Laporte (35:02)

I think it's. I like it.

Ian Thompson (35:04)

Woof. Right?

Leo Laporte (35:05)

Thank you. Pretty fly for a Sci Fi guy or assist guy. Appreciate that. That was creation of our, of our club Twit. Very interesting. You know, the question that we are continually asking and it's getting boring now, is will AI at some point cross that final threshold and become as creative, as talented, as intelligent as actual human beings? Is the singularity near? We just had Ray Kurzweil on two weeks ago on intelligent machines. He says that 2029, five years, four years from now, AI will be the equivalent of human imagination and human intelligence. And that by 2045, the singularity will happen where it actually surpasses us. You guys, just quickly, before we move to a break.

Ian Thompson (35:54)

I don't know, I'm. I kind of look on it in the same way that, you know, I look on sort of advanced fusion techniques and that sort of thing. It's always 10 years away, but it, you know, we're moving slowly towards it. 2045, that's reasonable.

Leo Laporte (36:08)

Quantum cryptography, quantum computing, fusion and artificial superintelligence, all are just around the corner. It often feels like it may never happen. It's like Zeno's Paradox. It will always be just a little bit away or something could snap and it could happen and the world would change. In all three of those Circumstances dramatically in ways unknown to us at this point.

Alex Kantrowitz (36:33)

Point, yes. It doesn't have to get there to have a profound impact on us.

Leo Laporte (36:36)

Oh, it already has. That we agree.

Alex Kantrowitz (36:38)

We're seeing already is going to change the way that we work. And yeah, you know, it's popular in some areas to just be so anti AI and say the whole thing is going to collapse or which it might because there's so much money invested and we need to see an ROI on that investment or at least the investors do. But the other side of this is it's already so useful. There is a path now that we're seeing with some of these reasoning techniques. Everybody's introducing different forms of agents that it's going to be able to do a tremendous amount of the work that we do already. So whether or not we, you know, officially like OpenAI declares. That's right, AGI and emancipates itself from micro from Microsoft, I think that this is just going to be like such a world. It already is and it will continue to be a world changing technology. Just think about how what we saw over the past five days is going to change advertising. Every ad mockup is now going to go through this ChatGPT image generation.

Leo Laporte (37:37)

And how much of that do you think is because of safety being ignored? Like they decided, you know, the safety thing was a mistake, let's just let it all hang out.

Alex Kantrowitz (37:48)

Well, I don't think they would say that safety was a mistake, but certainly taking the guardrails off is taking off the guardrails. It seems different. That will create more, you know, many more applications. And I think it's a good thing that in the beginning of this, the, the research houses aired on the side of refusal because they didn't want everybody to go around and do what they had done previously with chatbots, which is the Internet would always turn these bots into Nazis. I remember reporting on Teya, which was one of Microsoft's early chatbots. I wrote the story, I pinned to my profile on Twitter. I went to sleep, I woke up and I had a hundred like messages saying, can you please take down the Nazi story from here? And in that amount of time, in like 12 hours, the bot had become, you know, a Hitler supporter. And so I do think that it's smart for them to have decided, okay, we don't want them, the bots, to go that route. This is a new technology. It's not like taking answers out of a database. It's learning from its responses. But I think they've gotten to A point where they're pretty comfortable and now they're going to start to loosen it a bit to make the products more useful. So that's what we're getting to. And when they're more useful, to go to your point, Leo, they are going to really change the way that we work and the way that we live.

Leo Laporte (38:55)

There you have a piece.

Alex Kantrowitz (38:56)

Coding thing is amazing, just being able to code up a piece of software. So I think this is pretty wild.

Leo Laporte (39:01)

You have a piece that you just wrote that says, yeah, I'm starting to think AI can do my job after all.

Alex Kantrowitz (39:09)

That's right.

Leo Laporte (39:11)

Not your job though, right?

Alex Kantrowitz (39:14)

Well, I wouldn't say it can do my full job, but can do parts of my job. And, you know, this is basically the history of this piece is last year for the Boston Globe, I wrote a story that said, wait, ChatGPT didn't take my job. And it was this story basically kind of mocking the bots and mocking this narrative that AI could talk and therefore we're all going to be out of work. And this story was the follow up. And basically I said, listen, last year that was true, but this year I'm starting to see things from the bots that are making me a little bit nervous. And the story that, I have a couple stories in there, but the one that really inspired the piece was I spoke with this journalist, Evan Ratliff, who has a great podcast called Shell Game, where he clones his voice and attaches it to GPT models and then puts like 8, 000 biographical words in there to let it learn to be him and go out and be him. And. And he came on my show, we talked about this. He told me that the bot went out and did an interview with a tech CEO and got better stuff than he did because the tech CEO felt less pressured that the bot was listening. He knew the bot was going to be open to what he's saying, and he knew he basically had unlimited time to get his point across.

Leo Laporte (40:23)

It's kind of like Errol Morris is in Terratron, you know, where he interviews people in the. From the other room so they feel more comfortable. That's interesting. So.

Alex Kantrowitz (40:31)

Exactly. And so I think that experience isn't going to be unique to journalism in customer service. For instance, people are going to feel more heard by customer service bots that will listen to them.

Leo Laporte (40:41)

Really? Because people right now hate bots and.

Alex Kantrowitz (40:43)

They hate voicemail, but they're getting better. Remember the Scarlett Johansson thing? I know I kind of said it as a joke in the beginning, but if people believe that this voice can approximate what Scarlett Johansson did in her.

Leo Laporte (40:55)

It flirted with. It flirted with him, and probably we're.

Alex Kantrowitz (40:59)

Gonna see a lot of people fall in love with these bots. I'm not kidding.

Leo Laporte (41:02)

I spoke Kevin Bruce. He was the first.

Alex Kantrowitz (41:04)

Yeah, he was the one that one tried to break his marriage up. But since then, we haven't heard many stories about this. I spoke with the Replica CEO not long ago, and she said she's been invited to multiple weddings between people and their AI bots.

Leo Laporte (41:18)

No.

Alex Kantrowitz (41:20)

This is the technology. Is multiple weddings way better? So think about this. If the technology has the ability to make you fall in love with it, is the technology going to have the ability to handle your customer service request when it doesn't have a metric of getting you off the phone as quick as possible, which is how human customer service has, because they need to handle volume to justify the costs? Like, that's where we're headed.

Jacob Ward (41:43)

I mean, this is the thing, right, is like. I mean, I think Alex is exactly right that this thing does not have to be, you know, in any kind of, like, philosophical or even scientific sense, AGI for it to fulfill all the. The, you know, the necessity to check all the boxes for a human being to feel, you know, heard on the phone or, you know, heard by a virtual girlfriend or whatever like that, you know, the human. The. The human, you know, attack surface is very pliable. We are really workable when it comes to. At the same time, we've learned, like. But we're at this place where, like Alex mentioned, we're suckers. We're suckers. And this stuff is right up against, you know, Anthropic has this new model context protocol that you can just turn loose to write code and it'll bang out anything you want. I just think every one of these, I'm like, there goes my job, There goes your job, there goes that job. And we are not, you know, as much as these people who build these systems say, oh, we're going to now just live a life of leisure. There's no moment in the history of capitalism where that's ever been true. You know, so I just think, like, all of this can be. Is going to be. It's, you know, as soon as this thing can do a reasonable simulation of a human thing, it's going to be allowed to do that. And humans who do that for a.

Leo Laporte (43:00)

Living, that's probably this year.

Jacob Ward (43:02)

That's right. That's right.

Leo Laporte (43:03)

I mean, it's just around the corner.

Alex Kantrowitz (43:06)

My weird, hot take on this is that there's this theory that like, okay, once they get smart enough, the robots will enslave us and we'll just work all day on behalf of robots. And if you think about like just the US economy, I mean we're spending $300 billion this year. Exactly. The tech giants are going to spend $300 billion this year on CapEx. Most of that's going to go to AI servers. That's a huge part of like the profits in our economy going towards feeding the AIs.

Leo Laporte (43:35)

We're already at the scary argument that what's happening really is that these super rich billionaires and these giant companies, companies would like to replace human beings. We're a problem for them, we cost them money, we go on strike, we don't want to do defense work. Just get rid of all of them. And the problem, here's the problem with that, there's a big problem. The supply side problem or actually the demand side problem. If you get rid of all the people who's going to buy your stupid product. Right. You need people on the other side of what you're making to buy it. And if you fired them all, they won't have any money to buy it.

Jacob Ward (44:15)

I mean, Sam Altman has this famous quote that he has a, there's a bet going in the text group that he's in between a bunch of these CEOs as to when the first billion dollar single person company will come along.

Ian Thompson (44:30)

Yeah.

Jacob Ward (44:31)

Already you're seeing these companies come any day now.

Leo Laporte (44:33)

Yeah, yeah.

Jacob Ward (44:33)

These companies are already out there that are pulling down hundreds of millions of dollars in recurring revenue and have like 20 employees.

Leo Laporte (44:40)

I don't know if that's a good thing.

Jacob Ward (44:41)

I mean, well, this is the thing. Like, like who's gonna, you know, I don't care how big a house you build, you're not going to be able to employ that many plumbers, you're not going to be able to employ that many true driveway graders. Like who's gonna live? How are we going to live if, if only if each of these people have a single billion dollar company, there's only a certain number of them that are going to be available.

Leo Laporte (45:03)

You know, how many billionaires can there be?

Jacob Ward (45:05)

Yeah, exactly.

Leo Laporte (45:06)

Using these products. Is it billionaires to four billionaires by billionaires.

Jacob Ward (45:12)

Maybe, you know, this is why let's.

Leo Laporte (45:14)

Start a new company. Fbbb. Yeah.

Jacob Ward (45:16)

This is why I fully imagine like building a little studio in the backyard of our home in California so that my kids can live there. Because I Just don't know where they.

Leo Laporte (45:24)

Might want to grow tomatoes, green beans. Just, you know, get, you know, generate your own electricity.

Jacob Ward (45:29)

How?

Leo Laporte (45:30)

I'm sure, dig a well.

Jacob Ward (45:31)

Loose rules on cows in the backyard.

Leo Laporte (45:34)

Move up here to Petaluma. There's a law that says anybody can have a chicken farm if they want it. Everybody has chickens here.

Ian Thompson (45:42)

That's right. Chickens and butter, right?

Leo Laporte (45:44)

Chickens and butter, yeah. We had Anthony Aguirre on intelligent Machines a couple of weeks ago. He founded, with Elon Musk and Nick Bostrom and others, the Keep the Future Human Society or institute. And he is very. He was very concerned that artificial superintelligence and the key to it from him, from his point of view, was autonomous artificial superintelligence would just say we don't need the humans. In fact, the humans are consuming resources we need. So screw them. We don't need them.

Alex Kantrowitz (46:17)

They need us.

Ian Thompson (46:19)

We'll see.

Leo Laporte (46:20)

We shall see.

Ian Thompson (46:21)

By the way, I like Ian Banks's idea where they sort of treat us like pets. And it will be very amusing. Occasionally.

Leo Laporte (46:27)

We have something Ray Kurzweil told me 30 years ago when I asked him about this. He says, don't worry, they'll think of us as their mother and father and they'll revere us and honor us.

Ian Thompson (46:35)

Well, it is Mothering Sunday in the UK at the moment.

Leo Laporte (46:38)

Happy Mothering Sunday. We had Gary Rivlin on last week. His book AI Valley just came out. Fascinating history of AI this coming Wednesday, Keech Hagee will join us. He just published an excerpt from his new book on AI which comes out this week. There's a lot of competition for your book, Jake. His book on AI has the kind of explosive inside story of how Sam Altman got fired and rehired.

Jacob Ward (47:04)

Somebody needs to do that story.

Leo Laporte (47:06)

Well, he's got it. It's got. He did a lot of research. He got a great story. There is a good excerpt in the Wall Street Journal. Anyway, we'll be talking about that on Wednesday. Cory Doctorow joins us the following week. Harper Reed is going to be on. He's going to talk about vibe coding. Actually, he does more like pair programming with AI So that'll be interesting. He's a coder, so we've got some interesting people coming up.

Ian Thompson (47:28)

Corey, Cory Doctorow is a real pain to interview because everything he says sounds like a great quote. And it's just like, I can't just transcribe.

Leo Laporte (47:37)

Basically, when we have him on. Have you been on with him on the show? Because when we have him on, we try to get only one other person and some. And we just shut up. Yeah.

Ian Thompson (47:45)

I've been on with him and it's. I, I get exactly what you say. Just ask him about BBC Sounds. They'll just go on for 10 minutes about that.

Leo Laporte (47:53)

Yeah. They did the, didn't they do the original Hitchhiker's Guide? I think they did.

Ian Thompson (47:58)

BBC. Yes. Presents. Yeah.

Leo Laporte (48:01)

Hitchhiker's Guide to the Galaxy.

Ian Thompson (48:03)

Unfortunately, they're about to cut off overseas users. So I'm going to be.

Leo Laporte (48:07)

That's okay because we're about to cut off NPR and pbs. So, you know, it's just. We'll be living in this desert, this, this cultural desert.

Ian Thompson (48:16)

Well, Radio Free Europe might actually survive. There's a bunch of European governments talking about keeping it fun.

Leo Laporte (48:20)

Oh, that's interesting.

Ian Thompson (48:21)

So. Yeah.

Leo Laporte (48:22)

Yeah. Because they defunded Voice of America. Yeah.

Ian Thompson (48:24)

Yeah.

Leo Laporte (48:25)

Interesting.

Jacob Ward (48:26)

All right.

Leo Laporte (48:26)

We're going to take a little break. We come back. We should talk about the signal story. We will. The Signal saga. I don't know how much there is to say about it, but we will. And if you, if you have a minute or two, watch the Saturday Night Live cold open from last night because it was hysterical and it was a group chat with a bunch of teenagers and Pete Hegseth. It was great. I think Marco Rubio joined it a little bit later on. We have a great panel. This is. I'm sorry, Jake. It's going to be a long one. I tried. I really did. We have lots to talk about. It's great to have one new person and we're really thrilled to have Jacob Ward with us. Is his newsletter and podcast the rip current dot com. It feels like you belong instantly. So we're.

Jacob Ward (49:15)

So. I feel like I'm home. I really want to spend time.

Ian Thompson (49:18)

Yeah.

Leo Laporte (49:18)

These are your people. Yeah. Alex Kanterwitz, who I love, who is doing such great work at the big technology podcast and newsletter bigtechnology.com incredible interviews. Of course. Ian Thompson, who is the last ink stained rich in the. Actually there's no ink involved, is there in your.

Ian Thompson (49:35)

No, no. We've been online all for our entire 27 year history, but.

Leo Laporte (49:40)

Wow.

Ian Thompson (49:41)

Yes. That's good. Yeah, it's. I don't. I do like Internet journalism. I do. The problem with dead tree publishing was if you made a mistake, it was locked in there forever. At least with the Internet then it's kind of like if you make a spelling mistake. Yeah, you can fix that. But yeah, still battling away on a.

Leo Laporte (50:00)

Day to day basis covering the news, doing great stories. One of Your stories was food for conversation on intelligent machines this week. I don't remember which one, but we did talk a lot about you, just so you know.

Ian Thompson (50:13)

Okay, fair enough.

Leo Laporte (50:14)

Good to have all three of you. Our show today brought to you by NetSuite. If you listen to this show, you probably understand that a lot of what we talk about is trying to understand what's ahead. What does the future hold for business? The problem is if you ask nine experts, you're going to get 10 answers. Rates are going to go up, rates are going to go down, inflation's going up, inflation's going down. It would be nice if somebody would just invent a crystal ball, but until then, over 41,000 businesses have future proofed their business with NetSuite by Oracle. The number one Cloud ERP bringing accounting, financial management, inventory and HR all together into one fluid platform with one unified business management suite. There's one source of truth giving you the visibility and control you need to make quick decisions. With real time insights and forecasting, you're peering into the future with actionable data. You know when you're closing the books in days, not weeks, you're spending less time looking backwards and more time looking at what's next. If I had needed this product, it's what I'd use. I'll tell you what, whether your company's earning millions or even hundreds of millions, NetSuite helps you respond to immediate challenges and seize your biggest opportunities. This is the time. Speaking of opportunity, download the CFO's Guide to AI and Machine Learning. It's free right now at netsuite.com twit absolutely free. It'll be a good introduction to NetSuite, what they can do for you and also a really interesting explainer on AI and how it's going to apply to your business going forward. A little crystal ball for you@netsuite.com TWIT we thank him so much for supporting this week in tech. And you support us when you use that address. Netsuite. Netsuite.com TWIT thank you Netsuite for the support. So I almost forgot this story until Jake said, oh, are you going to cover the signal thing? And I went, what? Oh, that was this week. Wow. So I'm sure everybody by now knows that the Secretary of Defense, the Secretary of State, the dni, Tulsi Gabbard, the who was Mike Waltz? He's intelligence something. We're all in a National Security Advisor, National Security Advisor, NSA were all in a signal chat, which was a little weird. That Accidentally also included Jeffrey Goldberg, the editor in chief of the Atlantic. Goldberg published it, and at first he thought, this is somebody's pranking me. This can't be real. J.D. vance was in there, director of the CIA, John Ratcliffe, Scott Besant, the Secretary of Treasury, Stephen Miller. I mean, this was a heavy duty group. He thought, this is a joke. Until they put details of an attack on the Houthis, including time and date. And then an hour later, it happened. And Jeffrey Goldberg, who's sitting in his car looking at the news, is going, this wasn't a fake. There's no way this couldn't have been real. And wrote about it. Now, there's been a lot of noise ever since. Hegseth not only denied it, but. But called Goldberg names, at which point Goldberg said, I'm not a liar. In fact, here's more stuff that I held back.

Ian Thompson (53:56)

Well, they were saying it wasn't classified, so why not publish it? In that case, you know, it wasn't.

Leo Laporte (54:01)

War plans, it was attack plans. Please. Any event, there's been General deny you did it. You had a great podcast, actually, Jake, about. What do they call it? Dorvo. Darvo.

Jacob Ward (54:13)

Darvo, yeah. Oh, man, don't get me started.

Leo Laporte (54:16)

This was Darvo in a nutshell.

Jacob Ward (54:18)

Had the same thought. And you see it at every level of government in the response to this. So Darvo, for anybody who doesn't know, was coined by a researcher named Dr. Jennifer Fried, who was a guest on the Rip Current podcast. And she coined this term way back when Clarence Thomas was being confirmed, and Anita Hill came forward with her accusations, and she had come up with this acronym to describe the way that people accused of wrongdoing, public figures accused of wrongdoing, defend themselves. And the acronym stands for deny, attack, reverse victim, and offender. And it is an incredibly effective way of disorienting and deflecting accusations. And she documented it over time, you know, through life.

Leo Laporte (55:02)

I have to say, they've polished it considerably since.

Jacob Ward (55:05)

It is the language of politics now. It is the language language of politics. And the response to this is exactly that. You know, you see the. The press secretary being asked, you know, but aren't these classified plans the rest of it? And she says, well, it's, you know, this is at the hands of a. Of a Democrat, you know, journalist who's, you know, this on the other. I mean, you know, just the reversal mechanism. Yeah.

Leo Laporte (55:25)

Nobody was texting war plans. And that's all I have to say about it. They were attack plans. In any event. In any event, there is a tech story here, which is that they were using Signal, which is widely considered to be good, excellent, in fact, encrypted communications. Our own security guy, Steve Gibson on Tuesday did point out that signal is encrypted in the air, but the actual messages are decrypted so that you can read them. And so on your personal phone, they are decrypted. He also mentioned that, that we've had a hard time getting the Chinese out of our phone system. And so it's very likely that. Well, the truth is, signal is probably appropriate for you and me, but not for the department, the Secretary of Defense, the Secretary of State, the dni, the CIA director to communicate top secret information.

Ian Thompson (56:25)

Well, there were two things that struck me about this was. First off, there was a story went out that the fba, that the NSA had warned that there was a vulnerability in signal. Not quite true. There's a vulnerability in our handsets, as you point out. But also, why are they using Signal? You know, we have.

Leo Laporte (56:42)

It's illegal, as far as I know, for them to use Signal for this.

Ian Thompson (56:45)

Well, this is it. You have records laws. They have hardened communications.

Leo Laporte (56:49)

Well, that's why I think they were using it. They wanted to have deleting messages. In fact, many of those messages were set to delete. A court has now ruled that they may not delete them, that they must preserve them. They don't get the federal records. Yeah, yeah, well, and they'll go, he's got a copy. Just ask Jeffrey.

Alex Kantrowitz (57:08)

I mean, I do love. The administration was like, we've been very transparent about this whole situation. It's like. Yeah, you invited a journalist.

Jacob Ward (57:14)

Right.

Leo Laporte (57:17)

How does he get invited? I use Signal, as far as I know, in order to invite somebody to a group chat, you either have to have their signal handled or their phone number. Right.

Alex Kantrowitz (57:27)

I think there's little doubt that he was texting with administration members. And. Yeah, they just brought him in. I mean, it doesn't happen by accident.

Jacob Ward (57:34)

Or he's an incredibly well sourced guy. And of course, he'd be. He'd be in regular chat communication with some. A couple of those people, probably.

Leo Laporte (57:41)

Here's my suspicion, completely unfounded. Somebody was in on that chat, knew this was going on, maybe it had gone on before, knew it was illegal and covertly included Goldberg, because he, you know, is in effect, a whistleblower. Probably one of the names we know.

Jacob Ward (58:01)

But seems dumber than that.

Alex Kantrowitz (58:03)

No, it was added by Mike Waltz, who was.

Leo Laporte (58:07)

Well, we know he was added by Walt.

Alex Kantrowitz (58:09)

So that had to happen. I think from Waltz's phone. So it's not like Walt says he doesn't.

Leo Laporte (58:14)

He never has talked to Goldberg.

Ian Thompson (58:16)

Well, yeah, I mean, there's a picture that came up onto. It came up on the inset almost immediately after he said that, of them two, the two standing next to each other at an event. I think, honestly, it's idiocy is possibly the best, best solution because there is a senior US Government official with the initials JG and he's the trade representative. And I suspect that they just picked the wrong J.G.

Jacob Ward (58:39)

Yeah, that's right. That's right. That makes good sense. You know, I have. I have two things I've been thinking about with this. One is, you know, so. So your typical way of communicating in this way is through these skiff, these sensitive, compartmentalized. I don't remember what the security facility. Yeah, security, something, something, right. Where you got to go into a special room and if you. It's the ultimate return to office. You have to go to the office to use it. If you're a lower level person, only the very highest level people get a skiff built in their houses. And so if you've come up in the.

Leo Laporte (59:09)

By the way, most of these people would have a skiff in their house, right?

Jacob Ward (59:12)

Well, they would.

Alex Kantrowitz (59:12)

That's right.

Jacob Ward (59:13)

But you know, when traveling and everything else, it's a pain in the ass. You gotta like, go do. You know, you gotta get up from your, you know, get up and walk in your bathrobe over to the skiff. Right. It's a pain. And the. And this, the Twitter generation who aren't built for the patience for that. So that's one thing.

Leo Laporte (59:28)

Another thing is I remember when Obama became president and did not want to give up his BlackBerry, but was compelled to use. He was compelled by the NSA to use a Windows CE phone that he loathed, and he even talked about it on the Tonight Show.

Jacob Ward (59:45)

That's right.

Leo Laporte (59:46)

Those days are long gone. Right.

Jacob Ward (59:47)

So then the other thing, that one thing that's been jumping out at me about. About this is if you look at the screenshots and Bonito, I don't know if it's possible to actually put the real screenshots on.

Leo Laporte (59:55)

I can get them from the Atlantic.

Jacob Ward (59:56)

Yeah, you can pull them up here. One thing you'll notice in the. In the writing from Hegseth is there are EM dashes in what he writes. So. Right. If you're. If you're a magazine nerd like me came up through magazines, you'll like. I use em Dashes.

Leo Laporte (60:16)

I use EM dashes, but it's hard.

Jacob Ward (60:17)

To use them and you can only use them on a keyboard. You can't use them on a phone. I playing with signal just now. Signal doesn't have M. Even if you.

Leo Laporte (60:26)

Put two standard dashes in, it won't convert it to an M dash.

Jacob Ward (60:29)

Oh, maybe, I don't know, maybe. But there's an mp right here's exit. I understand your concerns and fully support you. And if you go down below. Right. The one. The one apostrophe or one. You know. Sorry, just where he's numbering things. Restoring freedom of navigation. Core national entry. Right.

Leo Laporte (60:46)

Yeah.

Jacob Ward (60:47)

So somebody was thrown at me. They were like, so either he's writing this off screen and cutting it, cutting and pasting it into signal, or. And here's the creepy part, did he use Chat GBT to write some of these messages? And if you did, how did you. How did you tell it what you wanted your mix messages to be about?

Leo Laporte (61:09)

You know, I hope he didn't use Chat gp.

Jacob Ward (61:14)

So I looked at that, I was.

Leo Laporte (61:14)

Like, maybe you should have had Studio Ghibli emojis instead of the standard flame emojis.

Alex Kantrowitz (61:19)

And, well, the White House did use the Studio Ghibli illustration in a real tasteless way.

Leo Laporte (61:24)

That was very tasteless, by the way.

Alex Kantrowitz (61:26)

Yeah, I'm glad he stuck with emojis on this one.

Jacob Ward (61:29)

Yeah.

Alex Kantrowitz (61:29)

By the way. Yeah. The idea that he's copying and pasting from Chat GPT just seems a little bit too far fetched for me.

Leo Laporte (61:36)

I mean, he's using a keyboard then that was automatically inserting M dashes. You think that's really interesting?

Alex Kantrowitz (61:42)

No, I just looked on signal. You can have an EM dash you. If you do two dashes. Yeah.

Leo Laporte (61:46)

Is that right?

Jacob Ward (61:47)

Can you do it?

Alex Kantrowitz (61:47)

Yes, you can do it.

Jacob Ward (61:48)

If you just go two dashes, it'll turn.

Leo Laporte (61:50)

Remember Pete was a journalist on Fox. He might have the same love of EM dashes that you do, Jake.

Jacob Ward (61:59)

Yeah, it could be.

Ian Thompson (61:59)

Could be our favorite DUI hire. But yes. Do you.

Leo Laporte (62:05)

Do you use EM dashes, Ian?

Ian Thompson (62:08)

Oh, yes, of course.

Leo Laporte (62:09)

I love EM dashes. Yeah.

Ian Thompson (62:11)

I mean, it's just one of those things. If you see it, then, you know. Yeah, that's proper and good.

Leo Laporte (62:16)

So is it a settings that you have to turn on, Alex, or is it. I'm looking.

Alex Kantrowitz (62:21)

You just open signal on your phone. I mean, I'm doing it right now. It might. Might also be operating system specific. But yeah, you go to. I have an iPhone. I hit 1, 2, 3 on the bottom of the keyboard, I hit dash twice, hit a space, it turns into nice.

Leo Laporte (62:35)

Oh, well, there goes the.

Jacob Ward (62:37)

I don't know.

Alex Kantrowitz (62:37)

Yeah, maybe, but maybe.

Jacob Ward (62:40)

I mean, we can agree it's egregious enough without using chat GPT, right?

Alex Kantrowitz (62:45)

Yeah, we don't have the scope.

Leo Laporte (62:46)

He's not using that.

Ian Thompson (62:47)

Well, I was, I was reading up on news over the weekend and there's a lot of very angry soldiers about talking about this because it's like if we did this, we'd be in Leavenworth serving time. Quite why the Secretary of State for Defense is allowed to get away with it. It's another question.

Leo Laporte (63:04)

Well, he's not going to get fired, according to Trump. But there should be a lesson learned here, right? Don't use signal. Lesson number one.

Jacob Ward (63:11)

Right.

Leo Laporte (63:12)

Will they be forced to use skiffs? And there is. I understand. I don't know. But I saw stories that said the DoD does have a secure messaging platform it can use.

Ian Thompson (63:21)

Yes, oh, definitely. Yeah. And they should be used in these kind of circumstances. I mean, one of the people in this group chat was actually in Russia when they signed onto the chat on a, you know, on a foreign visit, so.

Jacob Ward (63:32)

Which has been trying to break signal forever.

Ian Thompson (63:34)

Yes, exactly, exactly. But it's so much easier just to. It's. It's kind of like the old XKCD argument. Do you break through, you know, very high level encryption using hundreds of thousands of data sets or do you get a $5 wrench and beat the IT administrator up until he gives you the password? And it's the same thing with this. Signal is totally secure, but individual handsets aren't. And that's the key.

Leo Laporte (63:56)

Yeah. In order for you to read it, it must be decrypted and is decrypted on your handset. And if your handset has been cracked, which by the way, is still happening. Pegasus, there's still, you know, malware out there that can zero click hack a iPhone. You shouldn't be using your personal handset for this. But you know what? I don't know, and this would be a good story for somebody to do, is what does the government do with these people? Do they replace their personal phones? Do they have a business, you know, a government phone that's secured or guaranteed secured? One of the complaints that security researchers have about Apple's iPhone is they can't tell often if the phone has been compromised.

Ian Thompson (64:48)

Yeah.

Leo Laporte (64:48)

Because Apple locks everything down so.

Ian Thompson (64:50)

Well, I was going to say, I mean, if you're at that level of government, you really should Have a dedicated handset which is just used for critical information. Yeah. Keep your personal handset for personal stuff, but for really serious stuff like this, you've got to basically get. The NSA will do you a custom stripped down mobile operating system which is very secure and very hard to break into, and that should be you used.

Alex Kantrowitz (65:13)

You would imagine also that these people in that group chat are the ones that are like top of the list of other countries to hack. And if they do have those stripped down handsets, I don't think this is the ones that they were using, given that they have a journalist phone number or signal username in there.

Ian Thompson (65:30)

Absolutely.

Leo Laporte (65:31)

Oh, well, there's. I mean, so much ink has been spilled on this story.

Alex Kantrowitz (65:35)

I know. And all of us reporters are like, why can't that happen to us?

Ian Thompson (65:38)

I know.

Alex Kantrowitz (65:40)

That's the easiest scoop Goldberg ever got.

Leo Laporte (65:43)

No kidding. I thought he handled it quite well, though. He was very careful about what to reveal. He had another reporter from the Atlantic working with him side by side to validate information. I thought he handled it.

Ian Thompson (65:57)

Yeah.

Leo Laporte (65:57)

I mean, I also thought it was risk. It was a risky thing for him to do. Do you feel like he understood that what he was doing could end up putting him in prison?

Jacob Ward (66:08)

Oh, I think he absolutely. I mean, he absolutely did.

Ian Thompson (66:10)

The whole.

Jacob Ward (66:11)

The, the. This is the. My greatest complaint about journalism is that it doesn't. Doesn't. We don't do a good enough job of explaining what we do and how we do it. And in this case, I think he did a pretty good job of explaining. Here's how we made the decisions we did.

Leo Laporte (66:25)

Right.

Jacob Ward (66:25)

About the timing of the release of this. So that, and, and I, and you know, and every single thing he did went by a lawyer, no question, who I'm sure figures in this shows that there's no intent to inform a military adversary or any of that stuff, you know, but yeah, the, like, the, the, the rock steady process by which Goldberg and his team seems to have gone through this is really a testament to it. And it's part of why sticking around is really hard to beat him up on his, on his process. But yeah, your point? I totally woke up this morning with a. If, like this is the problem with being journalists is when you see great from other people, it fills you with self loathing.

Leo Laporte (67:07)

God damn it.

Jacob Ward (67:09)

Why aren't they texting me?

Leo Laporte (67:10)

How can I get in their phone book?

Jacob Ward (67:12)

Seriously.

Ian Thompson (67:13)

And to be honest, I mean, they also haven't released the full, you know, the full text string from the chat because in one case, there was a member of. A serving CIA member was mentioned by name. It is remarkable, though, that they let that this information leaked out, because when you go through the chats, they're really quite detailed in that one of Houthi's top bomb makers, for example, was observed going into his girlfriend's apartment. So they leveled the apartment building. Building. And having that level of detail, they must have had someone on the ground and have, like, burned that source now. So. Well, that's right.

Jacob Ward (67:46)

The Israelis are supposedly very upset about that, about that particular thing.

Alex Kantrowitz (67:50)

I just could not even imagine being Jeff Goldberg at that moment where you're like, he was in Austria, I believe, at the time, and saw these texts start to pour into his phone and go from. There's no way. These people who are like, using emojis are like, serious. And this has to be some sort of operation to. Then you see the bombs hit and be like, that was real.

Jacob Ward (68:11)

Real.

Leo Laporte (68:12)

It's crazy.

Alex Kantrowitz (68:13)

Just the shift is. Must have been totally nuts for him.

Ian Thompson (68:16)

Must have been a fantastic. You know, just. There would have been a fist pump going on, I think, to be honest, even though people have just died, this is like, yes, story's confirmed. We're going to run with this.

Leo Laporte (68:26)

There'll be a movie. Let's hope it's not created by Chat GPT A Ghibli FILM Ghibli JIB Studio Ghibli presents the Signal Group in recent weeks.

Alex Kantrowitz (68:41)

Who added Goldberg?

Leo Laporte (68:43)

That's good. Who added Goldberg? That'd be good. I like it. You know what? I might after the show have it written in the style of Neil Simon and Aaron Sorkin. And we can.

Ian Thompson (68:56)

Make some of this stuff up because the Senate and House were having intelligence briefings this week. And so they got all people like Tulsi, Gabbard and the rest of them in front of the Senate Intelligence Committee. And we're asking them about this. And she's like, I don't recall anything critical being done. And then he released the second tranche of messages as they were at the House of Representatives intelligence briefing. And it was like, so did you lie to us or what?

Leo Laporte (69:22)

You know, couldn't be my. And then, you know what, though? They're gonna escape with impunity. There's no.

Ian Thompson (69:31)

Can you imagine if a Democrat had done this? That would have been. That would have been.

Leo Laporte (69:34)

You know, they'd still be speaking for years. Hearings for years. In recent weeks, Customs and Border Protection agents, agents have been searching through the phones of people coming into the United States. A doctor coming in on an H1B visa was deported to Lebanon after CPB found, quote, sympathetic photos and videos of Hezbollah leaders. A French scientist was turned away after a device search unearthed messages criticizing Trump administration's cut to research programs. That's enough to get you sent back home. This is the story from the Verge. Courts have gone back and forth over whether the police and ICE and CPB can go through your personal electronic devices. They don't go through very many of them. Do you worry, Ian? You're not a. You're a green card holder.

Ian Thompson (70:27)

That's correct, yes.

Leo Laporte (70:28)

Do you come and go into the United States? Do you wipe your phone and devices before you cross the border? What do you generally?

Ian Thompson (70:36)

I haven't. Just simply for convenience. I mean, if I go to Black Hat or defcon, then I'll take a burner phone and use that.

Leo Laporte (70:43)

But that's not because you're worried about the government.

Ian Thompson (70:45)

No, no. But I mean, but also, you see, if you do take a burner phone, because I've been talking to a lot of people about this over the last couple of weeks because things, as you say, they've really cracked down on it. And I think it would be more suspicious to have a burner phone. And the way the law is at the moment, you could still be asked to sign into your online services.

Leo Laporte (71:05)

Right. They now search your social media. So unless you scrub your social media, it doesn't matter what you do on your phone, you.

Ian Thompson (71:12)

Well, Trump announced this week that basically, if you're applying for a visa, you are now going to have to hand over your social media details and they will be gone through.

Leo Laporte (71:20)

Supreme court ruled in 2014 that warrantless searches of cell phones violated the Fourth Amendment with the exception of the border.

Ian Thompson (71:29)

Yeah.

Leo Laporte (71:29)

Courts have held the border searches, quote, are reasonable simply because they occur at.

Ian Thompson (71:33)

The border, although not just at the border, because the border technically is, I think it's either 100 miles or 150 miles from my border.

Leo Laporte (71:41)

So if you live in New Hampshire, Vermont.

Ian Thompson (71:43)

Yep.

Leo Laporte (71:44)

Anyway. Anywhere you want to go. Anywhere you go. Or Maine. Anywhere you go. I do know people, I think. Well, I won't say any names, but people we know do wipe their stuff or carry Chromebooks and power wash them before they cross the border. Some people have international travel phones, but that was just a few years ago. Now, if you're a U.S. citizen, you have the right to say no to a search.

Jacob Ward (72:14)

But they can hang onto your phone.

Ian Thompson (72:15)

Yeah.

Leo Laporte (72:16)

They are not allowed to bar your entry, but they can keep your phone, laptop or other devices.

Ian Thompson (72:21)

Yeah.

Leo Laporte (72:21)

And Go through them. Permanent residents. That's what you are, right, Ian?

Ian Thompson (72:25)

Yeah, I'm permanent resident.

Leo Laporte (72:26)

You can research, but you may end up in airport jail, which, as I understand it, is exactly.

Alex Kantrowitz (72:31)

Yes.

Leo Laporte (72:31)

No fun.

Ian Thompson (72:32)

No.

Leo Laporte (72:33)

If someone with a green card leaves the US for more than 180 days, they're screened for inadmissibility upon returning to the country. Green card holders who have certain offenses on their records may also be deemed inadmissible.

Ian Thompson (72:47)

Yes, it is a worry because my mum's of a certain age and it's just like, if I had to rush over there, could I rush back? The way. The way things are going at the moment, you've just got to be incredibly careful.

Leo Laporte (72:59)

The thing that scares me is, as a US Citizen traveling abroad, I'm worried about retaliation.

Ian Thompson (73:08)

Well, British customs officers are. Can't say the word are. Are notoriously grumpy and lacking a sense.

Leo Laporte (73:14)

Of Mounties, by the way, people think Mounties wear big hats and save people from waterfalls. But, no, Mounties are scary. And I have faced customs and border patrol going into Canada, and it's. It's been difficult.

Alex Kantrowitz (73:31)

Even though that border is no joke.

Leo Laporte (73:33)

It isn't a joke.

Alex Kantrowitz (73:33)

That border is intense and.

Leo Laporte (73:35)

And given. And this breaks my heart because I love Canada. I worked in Canada. I love Canadian. Some of my best friends and many of our best artists hail from north of the 49th parallel. But I. They can. Canadians do not like us right now. And I can see retaliation. I. A lot of Canadians have changed their summer travel plans into the States.

Ian Thompson (74:00)

Yep. Yeah.

Alex Kantrowitz (74:01)

Over antagonizing them for, like, no reason. I mean, this idea that if you want to go after trade deals, that's one the of one thing. Right. I don't think some of the trade deals that we've had and with NAFTA have worked out in the US Best interest. And probably, we probably should have been smarter before running into free trade the way we did, because we.

Leo Laporte (74:18)

By the way, Trump made that deal in his last presidency.

Ian Thompson (74:22)

Exactly.

Alex Kantrowitz (74:22)

But this is the thing. The. The antagonizing for no reason and calling Trudeau, Governor Trudeau, and calling Canada the 51st state. I. I just don't see how that serves the United States interest. Interests.

Ian Thompson (74:35)

It's also backfired massively because they're going through an election campaign at the moment, and the Conservative Party had been easily due to defeat the Liberals, and now that's switched completely because Canada is in the throes of massive nationalism. I was having dinner with a Canadian friend a couple of weeks ago, and they're taking it seriously if America invades. And he was like, yeah, if they invade, then they'll beat us initially. But it's not just taking the territory, it's holding it. And Canadians are armed and really peeved about this. This.

Leo Laporte (75:07)

I don't.

Alex Kantrowitz (75:08)

Well, the US Is not invading, but it's not exactly being a very neighborly. It's not a neighborly moment, shall we say, between us.

Leo Laporte (75:17)

Wednesday is Liberation Day, which is the worst name for the day that these tariffs go into effect, including 25% tariffs on Canada and Mexico. We know cars are going to cost more. People don't talk a lot about technology, but I imagine iPhones and laptops and computers of all kinds will also cost considerably more.

Ian Thompson (75:41)

Yeah, we've upgraded our systems. Just before the tariffs came in, it was just like, right now is the time to buy a decent new computer while it's. Well, you still can at the price.

Leo Laporte (75:49)

That'S good for the PC industry. Thank you.

Ian Thompson (75:54)

Well, no, my wife bought MacBooks.

Leo Laporte (75:57)

Well, and the thing is, I mean, as much as Apple is trying to. To make iPhones or at least parts for iPhones in the US it's not doable in the near future. TSMC is building a plant in Arizona. That's the company that makes the chips in Taiwan for the iPhone. But they will be legacy nodes. They won't be the A processors that they go in the iPhone or the M processors that go. The Apple Silicon M processors that go in the laptops. Those are going to come from Taiwan. And even if the phones are assembled in Vietnam, I imagine they'll still be a consequence.

Ian Thompson (76:33)

Yeah, I mean, you can't just spin up a fab just like that. You know, it takes two, three years. And TSMC has run into real problems. They've had to import people from Taiwan over here because the skills base just isn't there to get these things built quickly.

Leo Laporte (76:44)

Right.

Ian Thompson (76:45)

You know, intel is doing the same. They're trying to build, spin up fabs and ask Pat Gelsinger how well that went. But, you know, I mean, these things take time and tariffs are a very blunt instrument to try and force it.

Leo Laporte (76:58)

Well, we'll see what happens. I feel like if you're going to buy a car, now's the time, right?

Ian Thompson (77:04)

Yeah.

Leo Laporte (77:05)

You have three days.

Ian Thompson (77:07)

Tesla's going very cheap now, as I understand it.

Leo Laporte (77:12)

Actually, yesterday was the big national protest at your Tesla dealership day. I don't know how many people were out, but there were hundreds at our local Tesla dealership here in Northern California.

Ian Thompson (77:23)

Well, Elon Was Rage tweeting this morning that calling for the arrest of somebody who gave a finger to a cybertruck owner. And it was just like, didn't he.

Leo Laporte (77:31)

Sue the advertisers who were boycotting Twitter? Is this different?

Ian Thompson (77:36)

Well, there's a report in the FT yesterday saying that, in fact, major advertisers are spending on Twitter only a token amount just to show that, you know, that everything's okay between them, but they're not spending what they were.

Leo Laporte (77:49)

Let's take a little break. We will come back because Elon sold Twitter to himself. And somebody I hope will understand what that means means and explain it to us in just a little bit. You're watching this week in Tech. Great to have our panel here. Alex Kanterowitz from the big technology podcast Always Day One, the book about Amazon.

Jacob Ward (78:09)

Just.

Leo Laporte (78:10)

Do you feel like you made the.

Alex Kantrowitz (78:11)

Right move going independent? Absolutely. Yeah. I mean, I went independent in May 2020. That's when I gave my notice to Buzzfeed. A lot of people said it was really stupid and crazy. I thought, thought it would be worse to stick around. I just had this book come out. I had these ideas that I wanted to pursue, and I wasn't spending a lot of money because I was in lockdown. So to me, it was exactly the right moment. And it's been. Been quite cool. And honestly, just love. I love doing the podcast. I mean, Leo, you know this, right? It's just a lot of fun to be able to bring people on and talk about what's going on in the news, talk about things that you might be curious about and hear perspectives from people that are in it. So I'm loving every minute of it.

Leo Laporte (78:56)

You do a great job and you get the best guests. You just had Dan Hendricks on from the center for AI Safety. Amira Frade, who does amazing work at the Information. Jan Lecun, the head of AI at Meta. That was incredible. Reid Albergatti from Semaphore. I know you had. Look, Panos. Panay. Amazing.

Alex Kantrowitz (79:18)

Yeah, we've had a good run and we have a very special show coming up on Friday. Friday. So everyone should stay tuned for that.

Leo Laporte (79:24)

Is it a secret who you're going to have on?

Alex Kantrowitz (79:27)

I can't say who it is because it's tied to some news being broken in the AI world, but it'll be a good show.

Leo Laporte (79:33)

Oh, my God. I'm gonna. If you're not subscribed, you gotta subscribe. And, you know, the good news is you can subscribe to the podcast for free. You can read the newsletter for free, but you should probably pay for it because there's great content behind the paywall too. Alex Cantrowitz, deservedly successful as in your industry independent career. Well done. Do you read Japanese? Is that why you have a Japanese book behind you?

Alex Kantrowitz (79:57)

So my book got translated into something like 11 languages. So that's just a translation of the book.

Leo Laporte (80:03)

Wow.

Alex Kantrowitz (80:04)

It says always day one just in.

Leo Laporte (80:05)

The language, but that's the question. Actually, it's Chinese. It might be Japanese with Chinese characters, but I do see the character for one one. So there you go. I do see day one.

Alex Kantrowitz (80:20)

These are the blurbs, so clearly.

Leo Laporte (80:22)

Oh, so the best blurbs. Awesome. That's fantastic. Congratulations.

Alex Kantrowitz (80:27)

Thank you, Leo.

Leo Laporte (80:28)

Ian Thompson, also here from Thereegister.com. when's your book coming out, Ian?

Ian Thompson (80:32)

Oh, well, you must have one in.

Leo Laporte (80:34)

The drawer behind that whiskey bottle.

Ian Thompson (80:36)

I've got three. I've got three on the go. So, you know, it's just, it's getting them finished. That's the thing. You just need the discipline.

Leo Laporte (80:44)

Novels or non fiction fiction?

Ian Thompson (80:45)

Novels.

Leo Laporte (80:46)

That's what I figured. I think you should be writing novels. Yeah, yeah. You have a literary style and you do use the EM dash, so that's good.

Ian Thompson (80:56)

Very well.

Leo Laporte (80:57)

Are they sci fi?

Ian Thompson (80:58)

Yes, yes. The problem is I've read so much science fiction over the years that whenever you write it, then you go back to it the next day and you think, no, that's too derivative of Gibson. That's two derivative of Clark here. Know, what can you do?

Leo Laporte (81:12)

So I don't know how musicians write new music. There's. There's only a limited number of chords. It must, it must be so difficult, you know, but that's what AI see, that's the thing. I, you know, AIs are doing the same thing, just ingesting other people's work to create new works, just like humans do.

Ian Thompson (81:28)

Well, I was talking to a journalist about this when ChatGPT first came out and they're just like, that's it for journalism. It's like, no, not quite. Because AI can't do breaking news, Right? It can do features and stuff like that, but it can't actually truffle out interesting facts. It can't cover a breaking news instance.

Leo Laporte (81:45)

Yourself to a pig. Is that what you're doing?

Ian Thompson (81:48)

Well, you know.

Leo Laporte (81:50)

Oh, Truffle dog. We saw the. We saw the Chat GPT.

Ian Thompson (81:54)

Oh, yes, the Muppet dog. Yes.

Leo Laporte (81:58)

Great to have you. And of course, brand new. He's a newbie, but he's fitting right in. Jake Ward Jacob ward.com and the rip current dot com. His book the link Loop is out now. And of course, the Rip Current, brand new podcast since 2024.

Jacob Ward (82:13)

That's right. That's right.

Leo Laporte (82:14)

It's good to put that on the wall because 20 years from now, you'll look back and go see since 2024.

Jacob Ward (82:18)

I know, but since back when humans.

Leo Laporte (82:20)

Were doing jobs, back when we were working.

Jacob Ward (82:24)

That's right. Just want to say it's a pleasure to be with you, not least because this is like, I feel like for so many years. I'm sure both Alex and Ian will agree with this. You know, the, this, this beat, you know, thinking about tech was considered this thing just for kind of like upbeat nerds to just kind of be off in our corner sort of talking about it. And now, I mean, the reason that I'm doing the ripper.com is that, you know, the, the, the Venn diagram of money.

Leo Laporte (82:52)

Yeah.

Jacob Ward (82:53)

And tech and billionaires and politics is now collapsing together in this crazy way. And to have been in this game as long as all of us have. Right. And have watched these people, people who also themselves were once considered kind of off in some corner of society, now be at the very center of it all makes, you know, that's why I'm trying to do what I'm doing independently, you know, and it's, and it's why I think this beat. You know, once my speech to when I worked at NBC and before that at Al Jazeera, I just kept trying to say to people like, this is the sort of lingua franca. This is the universal language. Now, if you, if you're, if you're being literate in this stuff is how you are going to understand why all of this other stuff happens.

Leo Laporte (83:35)

Yeah, it's kind of crept into the mainstream.

Ian Thompson (83:39)

It's interesting that some, you remember we used to have tech editions of some newspapers, right? That's gone. Because tech is mainly ones.

Leo Laporte (83:47)

Yes.

Ian Thompson (83:48)

Right.

Jacob Ward (83:48)

Yeah.

Leo Laporte (83:49)

I mean, I've been doing this for a long time. Back in the day, when they did tech on tv, they had a picture of a, a hand moving a mouse and a guy looking at the screen and the screen scrolling up, and they just roll that over and over again.

Alex Kantrowitz (84:03)

What is Internet, Allison? What is Internet?

Leo Laporte (84:07)

Yeah.

Jacob Ward (84:08)

At NBC News, I used to, you know, the, the format was that you would, you would write your, you know, you write your piece, you get it done, and then you're handing it off, typically to an editor who, who edits it, and you find it because it's breaking news. You're, you're seeing it for the first time along with the rest of the United States and the world. And so you don't get much input on the edit and, and I, and my, I always had like this little speech I would give to any new editor I was working with to be like no dark typists. No like hoodie.

Leo Laporte (84:35)

No hoodies.

Jacob Ward (84:36)

Hooded hackers.

Leo Laporte (84:37)

Yeah.

Jacob Ward (84:38)

And no zeros and ones streaming.

Leo Laporte (84:40)

There's such cliches.

Jacob Ward (84:42)

Yeah.

Ian Thompson (84:43)

It's interesting that a friend had got in contact with me who I used to work with and she sent me a copy of my very first tech article from 1991 and it was about this thing called multimedia. You could embed pictures and text and someday video into a document and it was a hilarious read.

Leo Laporte (85:00)

Looking back, it really is amazing. Our show today. Anyway, great to have all three of you. We have more to talk about of course, because Jake is expecting a three hour show and gosh darn it, I'm going to deliver our show today. Brought to you by Melissa, the trusted Data expert since 1985. And what I love about Melissa. Yeah, they've been doing, doing this for what is that, 40 years now. Amazing. But they don't rest on their laurels because they have been acquiring companies, they've been adding AI. Melissa's AI enabled data quality solutions go far beyond simply address verification. They're leveraging those four decades of accumulated data knowledge, advanced machine reasoning and now with cutting edge AI to transform raw data into actionable, reliable insights sites for your business. Melissa's ability to enrich and cleanse data spans multiple industries from fintech to healthcare to government, education, real estate. The list goes on and on. Melissa's suite of verification and cleansing services will benefit any business that doesn't want to act as its own data scientist. Whatever rules your business operates within, Melissa is there to support you. I had a great conversation with the Melissa data scientist a few weeks ago ago talking about how they can incorporate business rules into AI to give you actionable insights. It's really impressive. It's so much more than just address verification. Now imagine having a data expert that never sleeps. Malicious intelligence system verifies identity that's valuable in so many businesses. Know your customer rules in gaming operations. In medicine, Melissa ensures valid patient and medicine identification in healthcare systems. I mean really, this is much broader than just did we get the right address? Melissa securely updates and verifies constituent data across government databases. And Know youw Business enables verification and monitoring for financial institutions. Often Melissa solves a Compliance problem with Know youw Customer and those kinds of rules that couldn't be solved any other way way and they help you eliminate fraud. Melissa will guide you through all that complex data management with ease making advanced data quality accessible to everyone from small businesses to enterprises, yes even to giant businesses and governments with real time data validation, comprehensive enrichment, cross reference verification with gold standard reference data and intelligent anomaly detection. It's no wonder why Melissa is the trusted data quality expert worldwide. I was talking to Melissa about how the they match a patient's prescription with pictures of the medication to make sure that they're getting the proper medication and the proper doses, things like that. Now you don't ever have to worry about your data. With Melissa they securely encrypt all file transfers and they have built an information security ecosystem based on the ISO 27001 framework. They adhere to GDPR policies, SOC2 compliance. I want you to contact Melissa's team, find out what they can do to elevate your business, evolve your data quality. It's having a data scientist on your team. Get started today with 1000 records cleaned for free at melissa.com TWiT M E L I S S A melissa.com TWiT really, really great company. Very impressive team. Thank you Melissa for supporting this week in Texas.

Jacob Ward (88:36)

Check.

Leo Laporte (88:37)

So Elon Musk, you actually. Here's a picture of Jacob standing in front of the Twitter headquarters for NBC News the day that Elon bought Twitter November 2023 for $44 billion. He didn't want to. The Delaware courts made him. Actually, it's 2022, wasn't it? It's. Gosh, I can't believe it's. It's been two and a half years. Well now he has somehow managed to.

Ian Thompson (89:06)

Move.

Leo Laporte (89:08)

Twitter $33 billion worth into Xai. He says Xai bought Twitter but that's nonsense. I mean, what is it? Does anybody understand what happened?

Ian Thompson (89:24)

Well, it's. No, I mean silent.

Jacob Ward (89:28)

I mean. Yeah, go ahead, you take a stab. I'll try.

Ian Thompson (89:31)

Well, I was going to say, I mean it struck me as basic financial shenanigans. He manages to offload. Yeah. Basically by himself. And that gets him off the hook for an awful lot of debt that was used to buy Twitter.

Leo Laporte (89:45)

He gave those borrowers X AI stock. Right. So now they're all. Or. I don't know exactly. I think some of it goes to.

Alex Kantrowitz (89:56)

To.

Leo Laporte (89:57)

I don't know.

Ian Thompson (89:58)

Yeah, well, I mean you sort of.

Leo Laporte (89:59)

Remember is there Twitter stock at all? It's. But neither one is publicly held. Right. They're all privately held, so. But they're, but they're investors. The, the, the investors get XAI stock. Is that what it is?

Ian Thompson (90:10)

That's my understanding. Right.

Leo Laporte (90:11)

Yeah, that seems to be the case.

Jacob Ward (90:13)

And I think he's betting that somehow XAI will ride the same kind of valuations that OpenAI has, and that will create this incredible windfall for these investors who otherwise are taking this extraordinary bath on where they're at with Twitter. I love this enough about. Maybe everybody else can help me with this. Like, isn't there something about the Twitter financing through Tesla shares is coming due or something? And that maybe.

Leo Laporte (90:39)

I think there was for a while a concern that as Tesla stock falls and it has been in freefall that at some point the, the loans will be called. Called. But I think that's over. I think the banks have reevaluated. I think that's right and I think that that's over. So I don't, I.

Jacob Ward (90:59)

But I think the main thing here is this thing of, of this is, you know, must try to keep his investors in the black and. Trying to, trying to.

Leo Laporte (91:06)

Right.

Jacob Ward (91:07)

You know, it says it's, it's, it's just moving the shells around such that the, the, you know, I mean. Right. All these companies are, are, are, are ways of leveraging debt essentially, and, and, and betting on a plan. And if he can make the case that X AI is going to somehow go the same way as open AI, then they're all fine with it, you know, And I think at this point, like the Nor, like, for a lot, for so many years, it was my job to like, come on the air every day and be like, why would he. Why is Elon Musk doing this again?

Leo Laporte (91:40)

What the hell's going on?

Jacob Ward (91:42)

He's gone investor cost investors so much money and blah, blah, you know. But now as the second in command of the United States, who's going to cross him? Yeah, who's going to cross him? And what banker who doesn't want to be associated?

Leo Laporte (91:54)

Well, I think Twitter's worth a lot more now because of that. Right.

Jacob Ward (91:57)

Yeah. So it just feels like the normal rules don't apply to this particular guy anymore if they ever did.

Leo Laporte (92:02)

And so, so there's an interesting phrase for the way these people manage their tax debt called buy, borrow, die. And this is from the Bud Budget Lab at Yale University. Elon doesn't take money out of these investments, he borrows against them.

Ian Thompson (92:24)

Yeah.

Leo Laporte (92:25)

And this is really important because you neutralize your tax. If you took money out, you'd pay taxes on it. But borrowing is not a taxable event. The way. In the same way. And then the, the. So the, so the buy Twitter with borrowed money, not real money. If you need cash, you borrow against your Twitter. And that's kind of. He borrowed against Tesla stock. The DAI part is kind of interesting because you get to pass this on to your heirs. Elon admittedly has many heirs, but you get to pass it on without taxation because when the assets are inherited, their cost basis is stepped up to fair market value you at the time of death. So it's basically a tax avoidance mechanism. And I have to think that what's happened here with Xai and Twitter solves a number of problems, but chiefly one of them is tax avoidance. I love Axios's story. Musk shuffles $80 billion in assets, reshapes government. Just another week at the office office. What did you do this week? Did you restructure an entire country? Sit for a national TV interview and casually reorganize about 80 billion of your own assets on the side? If you did, you're Elon Musk. Last night's deal to merge his AI business xai with his social media company. X is normally something that might consume all of a CEO's time, but Musk did it while running Doge, reshaping the federal government. Government incidentally. Doge has now announced that they are going to rewrite the Social Security software written in Cobalt. So the Social Security Administration estimated It would take 15 years to rewrite this ancient code base. Musk says we're going to do it in six months. And good news, we're going to do it with the help of AI.

Ian Thompson (94:28)

Yes.

Jacob Ward (94:28)

Reminds me of the, this is the, the, you know that picture you put up of me standing in front of the Twitter headquarters back in 2022. You know the, the, that was that phase when he was like want like going in with pliers to just cut out servers and, and scoot.

Leo Laporte (94:45)

He feels like it worked at Twitter. He could do it now with the.

Jacob Ward (94:47)

Government, you know, but it's okay if Twitter goes down and you, you know, or you lose your account, like no.

Leo Laporte (94:52)

One'S dying or Nazis takeover or whatever it is, right.

Jacob Ward (94:56)

But in so with Social Security, right. The thing that all of you know there was a, I was reading a quote from a, you know, an in house government tech person who was saying, you know, you, if you can do it the way he's going to do it, but you're inevitably going to leave somebody out. They're not going to get paid. They're not going to get paid in time. They're going to get paid too much or too little. You know, something's going to break. And you know, I, my, my parents are just entering the Social Security collection age and you know, it would be a disaster.

Leo Laporte (95:26)

I am on Medicare and Social Security and I know many people who are because we're all in that same baby boomer demographic. It is a little chilling to think that they're going to take that millions of lines of cobol. By the way, it's very clear the Doge team does not understand cobol. That's where all that date stuff came from. COBOL doesn't have a native date format. So, so that's how people can be 185 years. They don't get deleted. It all started at the same time. There's no date for. So they didn't even understand that. That's why they said, oh, we must be paying people who are 185 years old. No, that's how COBOL works. Nevertheless, they want to completely replace this language, this code, in a matter of months. 60 million lines of COBOL code originally released in 1960.

Ian Thompson (96:23)

Good Lord.

Leo Laporte (96:24)

It's mainframe software.

Alex Kantrowitz (96:26)

Shouldn't we want, I, I, I don't think it makes sense to go through this recklessly and destroy the Social Security system. But how are we going to accept a government that's going to tell us that it's going to take 15 years to write a computer, rewrite a computer program? Why would we accept to have our Social Security maintained in a, you know, antiquated code? I mean, from the very little visibility that I've had into government software, we should all be outraged at the fact that the government is working on stuff that's decades behind what everybody else is working on.

Leo Laporte (97:00)

And by the way, I disagree with you on this.

Alex Kantrowitz (97:02)

If you're Just one last thing. If you're going to rewrite the code, of course you're going to want to use like an AI copilot. If every developer is going to be using AI copilots, then you're going to just say, I'm going to hand code it for like, what reason? Okay, but I'm willing to hear the counter argument.

Leo Laporte (97:17)

But there, well, that's exactly the counter argument you just gave, David. If it's working, you don't need to rewrite it. You know, just by because it's in Cobalt doesn't mean it's bad.

Alex Kantrowitz (97:29)

But don't you think that eventually that's not going to be sustainable, Leo, because like how, how long? Let's say we have a coding language that's completely antiquated. You know, how many, how long are we going to have engineers that are going to be able to maintain these systems? Like don't we want to have systems that are, are effectively up with the time? So not only can we disperse funds, but we could also do analysis of these government programs, which at this point are opaque.

Ian Thompson (97:58)

One thing I would say is that, yeah, I agree with you that having a system that old is less than desirable, but it does work. My worry about this is they're going to do a half assed job on it and after six months go right there's going be to. To be, you know, maybe a year or so where things don't quite work, don't work very well. And it was interesting this week. They had one of the Doge guys on giving a press conference and he was just like, look, my mother takes Social Security. I'm sure she wouldn't mind if a check came a little late once in a while.

Alex Kantrowitz (98:29)

Yeah, I don't trust them to rewrite. I'm not saying. I think they are pointing to a real problem. I don't trust them to solve it.

Leo Laporte (98:37)

You should also point out the air traffic control system is equally old and antiquated. Yeah, ideally you'd rewrite all of this stuff, but you would do it in parallel. You would do it with extensive testing and yes, you would take your time to do it. Right. Because an error in an air traffic control system, I guess an error in Social Security could really be a life or death situation.

Alex Kantrowitz (99:05)

You have my vote, Leo. I agree with that 100%.

Leo Laporte (99:07)

And I don't disagree that we need to modernize. Yeah, and we could do it. I mean Bill Clinton downsized government considerably. People forget this by letting people retire in an orderly fashion. The Social Security administration was going to rewrite its code. Covid forced them to rethink that. But it's been in the works. I don't think it's a good idea to rush. This is. I guess I agree, I agree.

Jacob Ward (99:40)

Bothers me about this is that I feel like we're in a country that doesn't that only has these two categories of reformer or I guess two categories of people whose attitude with an attitude about reform. One is the position that people have been put into of somehow defending Cobol right. We that like that, you know, that. That we want to defend the old way the bureaucracy worked. And then the other one is people who want to Bridge, break it and improve on it. And those people tend to overlap with people who make money off of the improve, you know, the upgrading of technology. And we don't have like that, that category in the middle. There's a. There's a fascinating case study of Estonia, which, if you want to think about, if you want to nerd out on, on government efficiency, that is an amazing test case. So when Estonia, which was a Soviet holding, passed over into being a Western democracy, you had this explosion of innovation and creativity and all this stuff. Now, this is a country of less than 10 million people. So it's not, you know, it's apples and oranges in some ways, but. But they have a system today. And I've talked to the president of the. Who was the guy who was president at the time, Thomas Elvis is the name. And he, he created a system in which, like today, you can pay your taxes in four minutes on your phone. You know, you can, you know, anything, any interaction you're going to have with the government, you can do, do on your phone, no problem. But the reason that it worked so well is that you had these deeply patriotic people who'd been held back under Soviet rule for so long, they wanted to kind of do this right as a matter of principle almost. And to me, when I think, you know, when I see people from DOGE wanting to change the Social Security Administration, I think, okay, you know, I can imagine a version of that in which I trust those instincts and I trust that process. But then I look at them taking apart the Consumer Finance Protection Bureau.

Leo Laporte (101:31)

Yeah.

Jacob Ward (101:32)

Which was arguably the most efficient government agency ever. That place, you know, returned, you know, billions in revenue on only hundreds of millions in costs, and yet they wiped that place out because clearly it was a threat to the people in, you know, in power in Silicon Valley. To me, that's where I. Where I lose it. But, I mean, I agree with you guys that there should be a way to be more efficient.

Alex Kantrowitz (101:58)

These can all be true. These can all be true, that we can get more efficient. We can be outraged at the antiquated software that our government is using, because when we use better software, we can do more things as a government. But we could also be upset at the fact that DOGE is not doing this in a sort of way that is logical.

Leo Laporte (102:19)

After the collapse of the ACA website, the Obamacare website, a number of Silicon Valley people got together, went to Washington, D.C. and the United States Digital Services was formed. Matt Cutts, who was the director until fairly recently, as a regular on this show Very smart guy, came from Google, left his job at Google in order to run this agency, brought in really smart people and worked responsibly to update technology. But what happened I think is now with Doge you've got basically the same philosophy as Facebook, move fast and break things. Is Silicon Valley idea that hey, we're smart, we're quick, we're fast, we're going to, you know, some things may go through the, but we're going to make it work. And I just don't think that's the way to handle these critical, mission critical systems. The Social Security database is a custom database called Madam the Master. It's not SQL, which by the way, Elon Musk said it was. It's the master data access method. It was written in COBOL and assembler for mainframe computers. In the 80s the Social Security Administration again had a modernization plan which was put on the back burner in 2017. I think you could, obviously you should update it. But anybody who's written massive data code bases, look at Microsoft Windows, let me give you an example which has roughly 60 million lines of code. Not in code ball in C. It is non trivial, you know. Yeah. Microsoft might say, you know, the best thing we could do with this is to rewrite it from scratch. They're not. Every time they've tried that, by the way, they've abandoned that. I don't think if Microsoft can't do it, I don't know if these 25 guys in Doge can do it.

Ian Thompson (104:20)

I'm not sure I'm willing to trust my Social Security. Someone calling themselves big balls, to be honest.

Leo Laporte (104:26)

That's who it is, by the way.

Ian Thompson (104:27)

Yeah, yeah. I mean, and I don't think big.

Leo Laporte (104:30)

Balls can read COBOL code and understand what it does.

Ian Thompson (104:33)

No, no, this is it. I mean this was one of the benefits of for COBOL programmers was Y2K because that did the retirement earnings for an awful lot of people who write cobol.

Leo Laporte (104:44)

I think the nice thing about cobol by the way, it's good for business logic. It you can read it and you can see what it's doing and it isn't as abstruse or weird as C, for instance. Instance. And just because you didn't grow up writing COBOL doesn't mean you can't look at that code base and make some sense of it. Anyway.

Ian Thompson (105:07)

Yes, it's going to be a weird one.

Leo Laporte (105:09)

You know, it'd be wonderful if they could magically create a whole new system. But anybody who's ever worked at an enterprise and tried to replace an existing code base with a more modern code co base knows how challenging and risk prone that is and, and how things break.

Alex Kantrowitz (105:28)

I mean, if you do break Social Security, that's a massive political liability and there will be consequences to it.

Leo Laporte (105:34)

So a lot of old people, we vote. Yeah, we vote.

Jacob Ward (105:39)

Last one to do well.

Ian Thompson (105:40)

I mean particularly with the other changes that they're announcing there where they're shifting to a more. You've got to. You can't verify your identity on the phone anymore is the plan. So you either have to go down to a field office which are being shut at a rate of not or you have to go online. Now a lot of Americans either don't feel comfortable online. They've got really lousy Internet connections. It's, it's one of those things that seems penny smart, pound foolish to use a British expression. You know, it's just you're really saving very, very little with a potential cost for an awful lot of people.

Leo Laporte (106:13)

Yeah, I'm, I'm sorry, I got a little bit head up. I apologize. Calm down. All right, another quick break and then we will talk about your DNA for sale. Congratulations. No, you don't get to participate in the sale. Sorry.

Alex Kantrowitz (106:32)

Just put it in a COBOL system.

Leo Laporte (106:36)

Nobody will be able to read it. So Cobol's dates begin in 1875.

Alex Kantrowitz (106:43)

That's when it was originated.

Leo Laporte (106:44)

Yeah. No, I'm kidding. They just didn't. They didn't, you know, I mean that's true In Unix it's 20. Like we're going to read reach this 32 bit limit in 20. 38. Y2K was the same problem because they were only using 2 digits for the year.

Ian Thompson (106:59)

Yeah.

Leo Laporte (106:59)

So these things happen. But I mean that's gotta love the.

Alex Kantrowitz (107:02)

Optimism that they say, all right, we're just going to code it and we'll.

Leo Laporte (107:09)

Now Darren Okey, who is a very accomplished programmer in our club, says I disagree. I've done lots of migration projects over the years. It's quite feasible, especially with AI. Darren is a big AI, AI partner coding guy. But it would be about understanding how to break off little bits of it, put in comprehensive regression tests, then pour to that bit of it. And that's the problem is, you know, especially if it was written in cobol, I doubt it's really very modular. I think it's. That's the problem. It's a big hairball of a program. I don't know if you can break little bits of it off.

Alex Kantrowitz (107:43)

I love how the discord is now discussing vibe coding. Social Security Security. I mean, just put a prompt into Claude about a new social.

Leo Laporte (107:50)

That's what scares me when they say we're going to use it.

Alex Kantrowitz (107:52)

Six months to do it. I could do it in an afternoon. Let me have at it with Anthropic's latest coding tools.

Ian Thompson (107:59)

Well, bear in mind also this is an Elon Musk deadline we're talking about. You know, if we'd actually taken his deadline seriously, we'd been on Mars about five years ago. So, you know, it's, it's one of those things. I don't understand quite why they'd set themselves up with this must be done by six months. Because that just seems to encourage people to make mistakes in a rush to get it done. But yeah, by all means, recode it, but test it. Test it. Test, yeah.

Leo Laporte (108:22)

Don't put it in production. I mean that's the problem is they're going to be putting it in production. Right. As quickly as they possibly can.

Alex Kantrowitz (108:28)

We don't know that. I mean they haven't yet. I don't want to. You know, not that I'm coming to their defense, but we haven't seen them mess this up yet. So I think.

Leo Laporte (108:35)

Yeah, but by the time they do, it's too late.

Alex Kantrowitz (108:38)

It's too late.

Jacob Ward (108:38)

But we've seen them mess up quite a lot of other things. The track record is not great so far.

Alex Kantrowitz (108:43)

Correct. But if they mess this up, the political cost, like I said before.

Leo Laporte (108:47)

Right.

Alex Kantrowitz (108:47)

Is heavy to pay. Yeah, let's just keep that in mind before they, we, you know, say they're definitely going to roll out a broken social.

Leo Laporte (108:55)

No, you're absolutely right. And I also agree with you. There is a lot a great case to be made to modernize government, to shrink government. There's a great case to be made. I think that's why there is what support there is for what's going on with Doge is because, yeah, these things did need to be done. I don't think anybody denies that.

Alex Kantrowitz (109:11)

I spent a very short amount of time in government and I can tell you that the software that government runs on is awful. It's miserable. It makes government employees less efficient.

Leo Laporte (109:21)

Yeah.

Alex Kantrowitz (109:22)

And there should be a prioritization for the government to be able to do this. And the reason why there's bad technology in the government is because the procurement process with the government sort of favors old school software companies that will not build the best in breed technology for the government. We definitely need to fix that. And there's this statement, if you Work in government long enough, you're going to hear, year. It's called good enough for government work. Where people are like, it's just the government. I'm going to get, you know, 80, 90% of the way there and we'll just kind of wash our hands of it. And we should be disgusted by that phrase and that that attitude exists within government and with the contractors and the procurement vendors that are getting these, these contracts to run to build these systems. It's just, it really does not. We should demand more than we're getting there.

Leo Laporte (110:09)

Right?

Ian Thompson (110:10)

It is astonishing. I went down to NASA JPL for the Curiosity landing on Mars. And, you know, this is. We all went down, we had briefings, very high technology project, that sort of thing. And then when they showed the presentation, it was on a system running Windows xp. And you're like, my goodness, how out of date is this? So, you know, you're right. Government software tends to be pretty poor.

Leo Laporte (110:33)

But solving it is difficult.

Jacob Ward (110:35)

I just wanted to be someone who doesn't make money off the product, you know, has no. Has no.

Leo Laporte (110:38)

There you go. There you go. Yeah. And please don't use Grok to do it. Okay? I'm just, I'm just begging.

Alex Kantrowitz (110:45)

Grok now, now is under the same umbrella as Twitter, so it's.

Leo Laporte (110:48)

Oh, that's. What could possibly go wrong. All right, little break, Come back. We do have many more stories to go through and only a little bit of time left. So let's move this, hustle this along with a great team. I could spend all day talking to you, but Jake says three hours. Really? So we won't make Jake do that. More to come in just a bit. You're watching this Week in Tech, brought to you this week by Oracle. There is a growing expense eating into your company's profits. It's your cloud computing bill. Now, you may have gotten a deal to start it, but now the spend is sky high, increasing every year. What if you could cut your cloud bill in half and improve performance at the same time? Well, if you act by May 31, Oracle Cloud Infrastructure structure can help you do just that. Oci. It's the next generation cloud designed for every workload, where you can run any application, including any AI projects, faster and more securely for less. In fact, Oracle has a special promotion where you can cut your cloud bill in half. When you switch to OCI, the savings are real. On average, OCI costs 50% less for compute, 70% less for storage, and 80% less for networking. Join Modal, Skydance, Zoom, and today's innovative AI tech companies who upgraded to OCI and saved this offer only for new US Customers with a minimum financial commitment. See if you qualify for half off@oracle.com TWIT that's oracle.com TWIT thank you, Oracle, for support supporting the show we're doing and thank you, dear listeners, for supporting us by going to that website, oracle.com twit so they know you. You saw it here. Let's see. 23andMe. I'm trying to decide. I sent my spit to Susan Wojcicki's 23andMe Many, many moons ago. They have it still. In fact, as they've upgraded their hardware, they use a chip to analyze it. I have spent more money to Upgrade it. Now 23andMe, unaccountably, is filing for bankruptcy. The CEO, Anne Wojcicki, the founder, has stepped down. And a number of organizations, including I think the Commerce Department, are saying, you know, you might want to delete your data. Data. If it's sold, whoever buys it will get all that data. It's interesting because Wojcicki, according to Engadget, is planning to bid for it to buy it. Maybe that's why she stepped down, so she could. She could buy it. The board of directors special committee has previously rejected a proposal from her. Should I be deleting my date data?

Ian Thompson (113:55)

Yes.

Jacob Ward (113:55)

Hell, yes.

Leo Laporte (113:58)

Alex. You want to make it unanimous?

Alex Kantrowitz (114:00)

Roll the dice, Leo. And give it to McDonald's.

Leo Laporte (114:02)

Who cares?

Alex Kantrowitz (114:03)

Let's have a good time with it.

Leo Laporte (114:05)

Yeah, who cares? Once valued at $6 billion, last month down to 50 million, they laid off over 200 employees late last year. Probably the nail in the Coffin was in 2023 when hackers accessed the information of almost 7 million customers. 23andMe didn't notice that until five months after the breach and didn't announce it until six months after the breach. A month after they discovered it, they did settle a class action lawsuit for $30 million, which, you know, if you're making 50 and you settle a lawsuit for 30, you get some idea of why they're in financial trouble. The California Attorney General. I misspoke. It wasn't the Commerce Department. The California Attorney General has suggested that you delete your data.

Ian Thompson (114:58)

Yeah, I mean, it's. It boggles the mind on one level because, I mean, I understood, you know, you can understand the attraction of it. It was a very cute idea. But, you know, I speak to the EFF probably once a week and, you know, they would have a chat with them once over, over a Pint or something. And they're like, we're trying to protect everyone's genetic data and they're actually paying a read it and store it. You know, it's just kind of, it's incredibly frustrating from a privacy perspective. But yes, it's. What could they do?

Leo Laporte (115:27)

What could they do? What is the, what is the worst case scenario? Let's say I don't delete my data and that somebody buys it. Who would buy it? I don't know.

Jacob Ward (115:35)

Okay, I'll throw one out here. So I'm back in. God, I don't even remember 2018 maybe. I wrote a piece for the New York Times Magazine about a field that is still moving but for some reason has not caught a lot of attention called genoeconomics. And this is people who are, it's economists and geneticists coming together and they use. This is going to sound like I'm making this up or like it's a science, like it's one of Ian's novels. But this is, this is people using technology, genetic technology, to make a prediction about social outcomes in their life. I was standing in the, in a conference room at Stanford when this guy was first trying to explain to me what this was. And I was like, wait a minute, you're doing what? And he's like, yeah, no, we're using Genesis to predict things like whether you are more or less likely to graduate from a four year university. And I was like, what are you talking about? Why would you, how would you, like, first of all, why would you do that? And second of all, how do you do that? And he said, well, we, we. I'll do the second one first. The, how we do it is using a huge pool of people and, and they draw that, the data mostly from 23andMe along with the UK Bio bank in your home country, Ian, to profile a sort of pool of people. And they can create these predictive scores that account for things like your socioeconomic background, your, you know, all these various sort of factors in your life and then, oh, put that together with your genetic data, create a predictive score that then can then be ported over to other people. And already, you know, you, you already have the opportunity if you're at a, you know, if you're, if you're a couple trying to get pregnant and having trouble and you want to go to a fertility clinic, which is a totally unregulated industry. The fertility clinic is allowed to offer things like you can choose whether you choose to go forward through IVF with the kid who's got the lighter skin or the darker skin or blue eyes or green eyes or whatever it is. And soon, if this technology keeps going, you'll be able to do things like which one? Which kid's got a greater chance of going to college or not. So then when I asked this guy, hey, what is the other, you know, why would you do this? He's like, well, it's because we want to be able to help people make up for difficulties in someone's genetics such that we can, you know, give them the better education, though.

Leo Laporte (117:51)

What's wrong with this?

Jacob Ward (117:52)

Well, this was the same idea behind the IQ test back in the day, which went on to be used to explore people and do eugenics. Right. I mean, it's a thing. This is the pattern is that once, once the market gets a hold of this kind of stuff, it's not about compensating for people's.

Leo Laporte (118:07)

I'm not worried, though. My Medicare and my Social Security check, I don't ever have to worry about that. I'm going to be fine.

Ian Thompson (118:15)

Yeah, I'm afraid not. No. I mean, this kind of thing is a gold mine for the health insurance industry, particularly when you tie it to information from data brokers. So, for example, does this, you know, you've got the genetic material, you know, how much they earn, you know, what they're spending it on, you know, what life choices.

Leo Laporte (118:33)

Isn't it more valuable to know if I go to Dunkin every month or every week or every day?

Ian Thompson (118:37)

Oh, yeah, but you get both of them.

Jacob Ward (118:39)

More the better.

Ian Thompson (118:39)

Yeah, the more information, the better.

Leo Laporte (118:42)

You know, I think insurance companies don't worry about that because what they do, in effect, before they insure you, and I'm sure you've all been through this, is they, they ask you a lot of questions like how often do you drink, how often do you smoke, do you go to Dunkin every day? And if you lie on that, then they can deny you. They hope you lie on it because then they can deny you, you, your benefits when you need insurance, when you need to pay off. I don't think, I think that they don't. I don't know. Is there evidence that they're using this information to deny, for instance, insurance to people?

Jacob Ward (119:18)

No, but there's no law against it. There's no rules against that thing. And once that becomes a powerful enough tool to be as predictive as asking you about smoking, there's no reason that the market wouldn't do it. I mean, the other one, Right, So there's the insurance insurance one. The other one is that 23andMe has complied with all kinds of subpoena requests for all kinds of data to try and lock down, you know, to, to compare the relatives of criminals to criminals to try.

Leo Laporte (119:42)

Right.

Jacob Ward (119:43)

You know, isolate stuff and, and you know, the, the, that people have been caught.

Leo Laporte (119:49)

In fact, didn't they, didn't they finally figure out how Zodiac was from, from the genetic. Was it Zodiac or was another.

Ian Thompson (119:55)

I think Zodiac's still in question on that.

Leo Laporte (119:57)

We don't know Zodiac, but I mean.

Ian Thompson (119:59)

There have been many, many cases where people have me and found out that guess what, one of their parents isn't their parent. You know, I mean that information is, is findable and doable. But I think coming back, but there's.

Leo Laporte (120:13)

Some benefit to that. We have a family member who I think was ancestry, not 23andMe, but who was a long lost son of, of somebody in, you know, somebody's grandparent who was reunited with his family, said fat, you know, and they actually, my ex wife just had a big dinner, you know, Thanksgiving with all of the reunited families. Thanks to the genetic information@ancestry.com I've myself have found second and third cousins on 23.

Jacob Ward (120:41)

You know, I, I think it works really well for people who have never experienced, I mean, you know, people like me and Julio who, who maybe have not experienced discrimination or wouldn't experience any kind of discrimination on the basis of our genetic backgrounds, you know, but you have, have this hacker group used that 23andMe leaked data way back when to try to assemble a list of people with Ashkenazi Jewish heritage.

Leo Laporte (121:04)

Right.

Jacob Ward (121:05)

Who knows what creepy ass stuff they were going to do with that?

Leo Laporte (121:07)

You know, that's an interesting question. Yeah.

Jacob Ward (121:09)

You know, like the, the ways in which it gets used against.

Leo Laporte (121:12)

I was on that list, by the way. I.

Jacob Ward (121:13)

Were you? Yeah. I mean, this is the thing, you know, like so, so I guess I'm, maybe I'm the only one in this group who, who wouldn't suffer from this. But you know, the discriminatory possibilities are incredible. And you can't change your DNA the way you can change your password. You know, when, when they tell you, oh your, you know, your DNA has been, your password's been leaked. Change all your passwords. Great.

Alex Kantrowitz (121:34)

It's absolutely.

Leo Laporte (121:35)

If I were 25, I would do it right now. But I'm an old man, what are they going to do to me? Right? It's, for me there's probably less hazard. It's probably the case if you're younger, you should Probably.

Ian Thompson (121:48)

I mean, honestly, the less data out there, the better. In these kind of circumstances. I take the view very much that you minimize your data footprint just on basic privacy grounds.

Leo Laporte (121:58)

But I am an experiment in the opposite direction. You understand that, right?

Ian Thompson (122:02)

Yeah.

Leo Laporte (122:02)

I'm wearing an AI pin that's recording everything that happens on the show. Everything I say, everybody I talk to then gives me notes. At the end of the day, it sends it off the recordings off to some unknown AI in the cloud. I've given up, I guess. But I also have. You're right. Like you, Jake, I'm at low risk. So, you know, I'm a CIS white male and so that's probably no risk to me. If you are at risk, delete your 23andMe data. Actually, Steve Gibson made a shortcut GRC.sc Byebye. Bye bye. You have to log into 23andMe and then click the link and it'll take you right to the page where you can download your data and delete it. It's not the only genetic company I've given my spit to, however. So my spit is the Discord is.

Alex Kantrowitz (122:59)

Saying that we're going to get a Leo clone. How would you feel about that?

Leo Laporte (123:01)

I would love that idea.

Jacob Ward (123:04)

Then it's make six hour shows. I'm not doing that.

Leo Laporte (123:09)

Okay. In fact, now's the time for Leo number two to take over.

Jacob Ward (123:12)

I'm tired. I gotta go.

Leo Laporte (123:14)

Good time.

Alex Kantrowitz (123:15)

My AI My G. AI in for the the second half here.

Leo Laporte (123:19)

Your eyes are getting wider and wider as we speak.

Alex Kantrowitz (123:24)

AI Leo in the Discord just said he's ready to take over the show, so.

Leo Laporte (123:27)

Crap. There is an AI Leo in the show. He is.

Alex Kantrowitz (123:31)

He is.

Leo Laporte (123:33)

Anthony Nielsen made him some time ago. I feel like we could update him maybe, I don't know, asked, are you ready to take over? Oh, absolutely. I'm like a cat waiting to pounce on a laser pointer. Just itching for action. What's on your mind? Tech questions? Gadget gossip? Maybe you just want to hear my latest dad jokes. Either way, I'm ready to roll. Says AI Leo. In my voice, I might add.

Alex Kantrowitz (123:54)

Welcome to hour four.

Leo Laporte (124:01)

All right. By the way, speaking of AI, the CEO of Replit, which is a great online coding tool I've used myself, says, hey, don't study coding now. AI is about to take over. One should not waste time on studying coding now. Oh my, look at all the pop ups on this.

Alex Kantrowitz (124:25)

I would rewrite this headline. It would be coding. AI CEO says AI coding will be big.

Leo Laporte (124:33)

Good point.

Ian Thompson (124:34)

Very much.

Leo Laporte (124:35)

Although for A long time. You know, he started Codecademy and Replit for a long time. He really supported people on learning how to code. Of course Replit now has, you know, what they call vibe coding built in. You write a prompt and it writes the code for you.

Ian Thompson (124:52)

I certainly wouldn't suggest as a career for an up and coming, you know, young student, it's getting a coding. Maybe learn how to use AI to get into coding. That would be something to study, but just coding itself. There's always going to be jobs because AI code has to be checked and improved. But longer term, there's going to be a lot of coders out of work, I suspect.

Jacob Ward (125:13)

I had a conversation with a guy the other day who I was at a party and this guy was lamenting that his daughter wants to go into art rather than being a lawyer or an engineer is what he was saying he wanted her to do. And I was like, I don't know, man. Are you seeing what's going on with lawyers and engineers right now? Like lawyers are getting fired left and right.

Leo Laporte (125:32)

By the way, law school. Law school's reporting a 20% increase in the number of people applying to become lawyers.

Jacob Ward (125:40)

No, I bet that's this little bump of people my age who have fallen out of all the other. I don't know, man.

Leo Laporte (125:45)

Well, the, the article I read said that people want to make a difference in the world, that people are becoming lawyers because they want to fight. So that's, I prefer to think that's the case. I don't.

Jacob Ward (125:58)

Well, in this case, like with this, with this coder question, I had this sort of been talking to this guy, I was like, you know, maybe your daughter, if she can really become a disciplined, prolific artist, maybe there is some sort of new. You know, I'm curious what everybody thinks about this. Like, what, what is. Like I've been kicking around this book idea of like, like the sort of the rebellion against AI in terms of resistance. Yeah, yeah. What would that look like? Like you are literate in the tools, but you're trying to push against this trend of every single thing is to going, going to look like a AI generated thing. What's the like, I don't know. Like, what is the, what is the career path of. You know, my 13 year old should be thinking about if it's, it's not coding, right? We. It's not, it's not being a lawyer, I don't think. But I'm not sure it's being a, you know, sculptor either.

Ian Thompson (126:48)

I think it's possible that There could be.

Alex Kantrowitz (126:50)

Sorry, I was just going to say stay away from Studio G. You're, you're.

Jacob Ward (126:55)

Not going to work at Studio G. That's exactly right. That's right.

Leo Laporte (126:57)

That's a really interesting question. What should a 13 year old think about their future? You know, my kids are now 30 and 32 and they look at the people coming out of college right now and go, oh, I feel bad for them. These are the poor kids who went to high school during COVID They're getting out of college now and they're looking at a job market that is in complete disarray way and they're being daily. And by the way, housing costs are so high that they can't afford to rent without three roommates and two jobs.

Jacob Ward (127:34)

Well, the housing form of entertainment is to, is to think that they should be famous all the time.

Leo Laporte (127:39)

That's the answer, right? Oh, we should.

Alex Kantrowitz (127:40)

Maybe they should.

Leo Laporte (127:46)

My son is doing quite well. He had, he's a tik tock chef who is turning transferred his knowledge and viewer base to Instagram and he has a cookbook out that was a bestseller and now he's opening a restaurant in New York. So it's too late. You can't do that anymore.

Alex Kantrowitz (128:03)

Well, why not though? I mean, isn't the answer here in some ways to do entrepreneurship like we talked earlier today about the $1 billion one person startup? Well, why does it have to be just that with the tools that are available today, thinking about the, even the coding examples that we use where you could code up programs for yourself, you can create art for yourself. I mean all these things that it would take an entrepreneur, you know, or would make being an entrepreneur or one person entrepreneur cost prohibitive. Those barriers are coming down.

Leo Laporte (128:35)

Well, it used to be knowing coding was the way to do that, right. You were going to be the next Bill Gates. So now what is it? What skill should you learn?

Ian Thompson (128:43)

I was tossing this idea around with a friend a couple of, a couple of, couple of weeks ago and I think there might be a market for hand human produced art documents, that sort of thing where you know that an AI hasn't been involved. And I don't know, maybe there'll be a premium for that kind of thing, but it'll, it'll take a while to come.

Jacob Ward (129:03)

I think that's like the, the fanciest possible thing is a, is a thing made by human hands.

Leo Laporte (129:10)

So be a potter is what you're saying, Right?

Alex Kantrowitz (129:13)

Hospitality also seems like a area that's going to consistently be like making people feel welcome. Making people feel appreciated. Seems like that's the AIs are going.

Leo Laporte (129:23)

To take over customer service.

Alex Kantrowitz (129:25)

Customer service is not all of hospitality. Like there's still, there will still be you, you might have somebody that will help you deal with your problem, that will be a robot. But the person that's going to welcome you into a restaurant, a hotel, that's the one that's going to put those experiences together. Your son for instance, has the exact, I mean basically it's the business of the future. Right. It's content but also hospitality.

Leo Laporte (129:45)

He's putting a studio in the front window of the restaurant. Restaurant.

Alex Kantrowitz (129:49)

There you go.

Leo Laporte (129:49)

So people see him making the sandwiches, making the videos and then they go and they buy the sandwiches. Yeah.

Alex Kantrowitz (129:55)

What about the care business? I mean, you know, people joke we're.

Leo Laporte (129:59)

All going to need nurses soon. Yes.

Alex Kantrowitz (130:01)

I think nursing is going to be a profession that already is in high demand. There's already huge shortages. I think that is going to be an area where nursing doctors. What about there, there's going to be room for like basically nurse practitioners. I think people who have, have you know, a little bit more training to really be able to have somebody come in the office and be like listen, I just read about what I'm dealing with with chat GPT. Do you think this is the right medicine? And they prescribe it to you? So these are, there are areas where we're going to be able to see your human flourishing and the need for humans. It's not all going to go to the robots.

Leo Laporte (130:38)

One thing you should not do, create a website of history. A history website.

Jacob Ward (130:46)

Website.

Leo Laporte (130:47)

You talked to the CEO of the World History Encyclopedia. As AI takes his readers, Alex writes a leading history publisher wonders what's next. His traffic to his site has dropped 25%.

Alex Kantrowitz (131:02)

That's right. So this is Jan van der Kraaben. He's the publish CEO and publisher of World History Encyclopedia. It is the number two history site in the world I think outside of archives. Right. So I guess the government's beating him there. That must be some good technology.

Leo Laporte (131:19)

Don't worry there. The government's shutting down all the museums. So I think he's good. That's going to help, right?

Alex Kantrowitz (131:24)

Yeah. But basically what he has, he has this site that covers world history. He has nine full time staffers, they write about 500 articles a year. And what he is faced with right now is that ChatGPT and AI overviews have ingested, tested his content and they're, they're presenting it summarized alongside other history websites. To users. And so he saw, initially he goes, oh yeah, that's our stuff. And with Google, at least there's links. Then a big Update came in November 2024, and then all of a sudden, basically overnight traffic dropped 25% and this.

Leo Laporte (132:02)

Understand though, that he made his success based on, on the demolishing of Encyclopedia Britannica and the World Book. Right?

Alex Kantrowitz (132:12)

I mean, listen, yeah, I think there's also, there's like, there is a mutual relationship with him and Google too. Right? Like 80% of his traffic still comes from Google. And he built his site based off of Google. But here's what's different. I was like, all right, so like, what's different now? He says, right now it's just Google that benefits. And the same is true, let's be honest, for ChatGPT, for Anthropic, for many of the AI chatbots that gobbled up.

Leo Laporte (132:36)

Couple up the content, I use perplexity. And I very often, even for things like recipes don't end up at the website because perplexity gives me the content.

Ian Thompson (132:45)

That's right, yeah.

Alex Kantrowitz (132:46)

Look, at least he has historians and volunteers writing the articles. And what happens like there, there's a value exchange there, right. And I think that deal that publishers have always had with the platforms that they produce, and the platforms can basically highlight the stuff and exchange for being featured at some point that is changing and now the platforms are just taking it.

Ian Thompson (133:09)

But I mean, we're all journalists here. We've seen this happen before. When it came to the collapse of advertising revenues going to publishers rather than to Google and Facebook. We've been down this route before. We will go down it again in other industries. I can't help feeling this is what.

Leo Laporte (133:24)

Happens with technology, right? Disintermediation. And.

Alex Kantrowitz (133:31)

So the question is, what comes next? And the answer that Jan had was pretty interesting. He thinks there's a possibility in the future that instead of them publishing to the websites, they'll just be paid by the platforms. They already have a revenue share deal with perplexity to just write for them.

Leo Laporte (133:47)

So I can use perplexity. It's okay.

Alex Kantrowitz (133:49)

Yes, I think perplexity is the best of them all.

Leo Laporte (133:53)

But basically, what a release. I was feeling guilty.

Alex Kantrowitz (133:56)

This quote really got me. He said, we'll no longer be publishing to readers, we'll be publishing. We'll be feeding the machine. And the readers read what the machine gives them. And he says, who. Who controls the machine? What biases do the does the machine have? How does this affect the world? He's quite Concerned about it. And I don't think that's off base.

Ian Thompson (134:14)

No, no, I mean this is the problem with AI generated comment is you've got to, you've got to check it, you know, I mean, when things like Chat GPT first came out, that sort of thing, you do what you, everyone does. You ego surf and you push in, you know, you push in your details and see what it has. And according to ChatGPT, I am an award winning journalist with the New York Times and the Guardian. And it's like never written for either of those, I'm afraid. Yeah, well, this is it. But I mean it's one of the greatest PR tricks the AI industry has ever done is dubbing these things hallucinations rather than mistakes. Because they are mistakes. But if you give it a go.

Leo Laporte (134:57)

But humans make mistakes too, right? I mean you have to, you have to check your sources and, and get another opinion and.

Jacob Ward (135:06)

But the relationship we have with computers is not like that. We expect a computer to give us accurate information.

Leo Laporte (135:11)

Benito, our producer, the, the voice of God talking. Yeah, that's what the, that's what the point of this is.

Jacob Ward (135:17)

Yeah, sorry. I was gonna say this is the thesis of my book, is that we are, we as humans are terrible at telling the difference between we can't judge a system we can, we can trust and one that we can't. Our, our brains are shortcut machines. Teens that don't like to do work. That's how we survived, that's how evolution trained us, is to be as quick, as decisive as possible, using as little energy as possible. And AI is the perfect kryptonite for that. Because we are going to have no ability to tell, to, to take a second to see whether Ian writes for the Register of the New York Times. Like we don't do that. And, and, and yet the business model is going to be built on the idea that, that, that you can trust these systems. Yeah, and to me there was this, you know, there's a, I interviewed a guy once who's a federal judge who was talking about how you can make systems more efficient. And the market is always going to try and do that, but there are certain frictions you want to save, you want to keep. And in law they call it's weak performance. Affection is, is the legal term for this. And it is basically like he was saying that you could, for instance, make it much easier to enter a guilty plea or a not guilty plea. And that doing so, you know, could, you could swipe right or swipe left and be Done. Currently, the law, the legal system makes you come in. In person to, you know, say it in front of a judge. And. And the judge, you know, has to go through this very complicated process because you only get one shot at this. Once you enter a guilty or not guilty plea, you're gonna. It's gonna change your life forever. And you never. And he was like, we have to. You know, we could make it better, but we instinctively. We do it in on purpose, make it hard on purpose so that you have to engage your thinking about it. And for me, there's something about the way that. That the flow of. Of information and what we're supposed to assume is objective information is being made more and more efficient, and that is creating, I think, real trouble.

Leo Laporte (137:19)

Efficiency is our enemy, is what you're saying.

Jacob Ward (137:22)

Not enemy, but it. But. But it can't be our own trap. It can't be the only thing we measure this stuff.

Leo Laporte (137:28)

Does this relate to Daniel Kahneman's Thinking Fast, Slow?

Jacob Ward (137:32)

That's what it is. So my book basically is pitched as Thinking Fast and Slow meets her is sort of the. Is the is the book is trying to say, like, I spend like, three chapters summarizing Kahneman and all of his disciples work to say, based on what we found over the last 40 years, here's how the brain works. And then, oh, my God, here's all of these companies about. About to, you know, drive your autopilot for you.

Ian Thompson (137:56)

A friend of mine's a teacher, and it drives her nuts when kids bring in stuff and they, you know, essays that are wrong. It's just like, well, that's incorrect. He goes, well, no, the AI told me it was correct. And it's like, yeah, now she's got to teach them to check those sources. So, yeah, I'm looking forward to that book, actually. It's. It sounds like a very interesting read.

Leo Laporte (138:16)

Kathy o', Neill, who's going to be on our show in a few weeks, says it's a fascinating and unsettling search survey of the known spectrum of human biases. I am ordering the book right now.

Jacob Ward (138:25)

Oh, I appreciate that. Thank you.

Leo Laporte (138:27)

That sounds really interesting. Should I order it from your site? Is that a best.

Jacob Ward (138:32)

Yeah, can I. I don't know. I can't get away from the machine. So.

Leo Laporte (138:37)

Amazon, huh? Oh, yeah. Order your copy here. I always like to. That probably gets you a little kickback.

Jacob Ward (138:44)

I would hope smart enough to do that, but, yeah, maybe. So feel free. Please do.

Leo Laporte (138:49)

Oh, and there's another audiobook version.

Jacob Ward (138:51)

Yeah, that's my. It's my voice.

Leo Laporte (138:53)

Yeah, you have a great voice. I will, that's the one I'll be ordering. I think that's a really interesting point. I hadn't really thought about the whole. Kahneman's whole thesis is we have two kinds of processes in our brains. The fast one that is quick and intuitive and the slower one that is a little bit more methodical and that's interesting. So that's right.

Jacob Ward (139:17)

And the, and the, and the new, the fast thinking brain is our ancient most brain that's we have it in common with primates like it's been around since long before we were even standing on our hind legs. And, and the new fat slow thinking brain, the creative rational one that we think is who we are, you know the Atticus finch, the, the thoughtful creative one, that's who we, that's our self image as humans. My experience, when you talk to people at the top of Silicon Valley companies, they think that's who humans are. Are, you know, we are going to be thoughtful and reasonable. Turns out that system barely gets involved in your day to day life and the rest of the time it is just the ancient system making decisions. And the problem with AI and these ways it's being deployed in the market is that it plays on the fast thinking brain. It plays on the really ancient, ancient system that identifies berries and outsiders and all that.

Leo Laporte (140:11)

And that's what you mean by the loop is that we are, we are in this kind of vicious circle that is tightening around us.

Jacob Ward (140:17)

We're going to feed in more and more of these kinds of ancient choices, our bad instincts into these systems which are going to deploy them as products and deploy them as a smaller list of choices. And I think a lot of people talk about AI as this kind of flywheel that's going to create this endless opportunity and endless creativity. It's going to be this sort of open wheeling universe of fun. And my thesis is, is if it, it could go that way. I hope it does. But I think it's given what I've, you know, given what I've reported in the book, if it goes the wrong way, it's going to be a tightening and tightening spiral and at the bottom of it we're all going to be like wearing beige and drinking Soylent and sense of humor.

Leo Laporte (140:59)

I just don't want the high water pants they wear in her. That's the only thing I don't want to do that. I just ordered it on audible I think. I had no idea. It sounds fascinating. All right, we're going to take a little break. A couple of final stories. We're almost done. Alex Cantrowitz. Jacob Ward. Ian Thompson. So nice to have all three of you this week in tech. Was a good week to have you on, actually, as it turns out, this week in tech brought to you this week by Shopify. We were just talking about my son, Salt. Hank.com when you think about businesses whose sales are rocketing like Allbirds. Wearing them right now. Untuck it. Wearing them right now. Salthank.com you think about an innovative product, a progressive brand and button down marketing what is often overlooked the unseen secret that's actually the businesses behind the business. Making sense selling. And for shoppers, buying simple. For millions of businesses, including those three, that business is Shopify. Love that sound. Salt. Hank got started, you know, he was just a guy, you know, just a kid and he said, dad, I've got this TikTok following. I want to sell salt. He created a site with Shopify. Thank you. And, and it's grown from there. He's now selling pickles. And I have to say a little disclaimer. I am an investor in the pickle business and I'm thrilled to see him using Shopify. Nobody does sales better than Shopify. It is the home of the number one checkout on the planet. The not so secret, secret shop. Pay boosts conversions up to 50%. Now that's huge. You know one of the problems, you go to buy something a, let's say a case of fine flaky salt or his fabulous truffle salt and then you get distracted, you change your mind, you walk away and the shop cart goes abandoned.

Alex Kantrowitz (143:06)

Right?

Leo Laporte (143:07)

Not with Shopify. Way fewer carts go abandoned. Way more sales are being made. Can you hear that sale? That's how I know. That's how I know that's a sale right there. So if you're into growing your business, your commerce platform better be ready to sell what you whatever and wherever your customers are scrolling or strolling on the web in your store, in their feed and everywhere in between. And I'm sure when the restaurant opens in New York City on Bleecker street right next to John's Pizza. Salt Hank will be using Shopify. Businesses that sell more sell on Shopify. Upgrade your business. Get the same checkout that Hank uses at Allbirds uses. Sign up for your $1 per month trial period at shopify.com TWIT. That's all lowercase case TWIT shopify.com TWIT to upgrade your selling today. Shopify.com TWIT. Oh, we just sold another jar of pickles. I don't think Shopify knew that I would incorporate my son's website into this ad. Thank you. Shopify really does. It really does empower the little guy and little galaxy to create a business. It's kind of cool. Thank you, Shopify. Oh, here's a business. Madison Square Garden. You hear what the Dolan's are up to? Madison Square Gardens has face recognition. As you know, they've banned people that Dolan doesn't like. The latest. A fan wore a T shirt that says Ban Dolan. He was concert on Monday night. Radio City Music hall, owned by msg, Owned by the Dolan's. Frank Miller was there with his parents for their wedding anniversary. He didn't see the show because the minute he scanned his ticket, he was pulled aside by security. Told by staff, you are barred from MSG properties for life. For life. For incident at the Garden, Madison Square Garden in 2021, Miller said, wait a minute. I haven't been to the garden in 20 years. They hand me a piece of paper letting me. This is from the Verge. Letting me know that I've been added to a ban list. There's a trespass notice if I ever show up on any MSG property ever again. I mean, that's not just Radio City Music hall. It's the Sphere in Las Vegas, the Beacon Theater in Boston, the Chicago Theater Theater. He was baffled at first, the Verge writes. Then it dawned on him. This is probably about a T shirt he designed years ago. MSG Entertainment won't say how they figured out it was him or whatever, but this goes back to 2017, when Charles Oakley was removed from his seat near Nick's owner and Madison Square CEO James D. Dolan. It became a legal battle for Miller. Oakley was an integral part of the 90s Knicks, and with his background in graphics design, he made a shirt in the style of the old team logo that read Band Dolan. Oh, there's your mistake. This is the world we're entering in now. There's no anonymity because of face recognition. And the Dolans are quick to use this. They've used it before.

Ian Thompson (146:33)

Well, they banned. I think it was a lawyer went to one of their properties.

Leo Laporte (146:37)

She worked at a law firm that had a lawsuit with Dolan. She wasn't involved in the lawsuit, but didn't matter. And of course they have the right. It's their venue. They can ban people, but this is problematic.

Ian Thompson (146:51)

It's the wave of the future, to be honest.

Jacob Ward (146:54)

Sorry, sorry, sorry. Go ahead. Sorry, sorry.

Ian Thompson (146:57)

No, I'm just thinking it comes back to what you were saying, Jake, about discrimination and the increasing ways that technology can, can be used in that way. And this is a prime example, you know, facial recognition technology does something which, or finds something which every, this guy obviously thought everyone had forgotten about and boom, instaban, you know, and this is the way it's going to go in the future.

Jacob Ward (147:21)

I was going to say that the thing that is crazy to me is how often. So, you know, the, the China is sort of held up as, as this, you know, and it is, I mean it's an authoritarian regime and, and it's constantly held up by the heads of big tech companies as the, as sort of the ultimate evil and we have to compete with them. You know, we don't want to become like China. China, meanwhile, just, I don't think it was last week, but the week before just passed these new rules basically limiting the amount of fish. This is wild pace.

Leo Laporte (147:55)

And they are ahead of us.

Jacob Ward (147:57)

Yeah, they're pulling back because they think it, it's not right. This is from, you know, I think this from your colleague, right. Simon Sherwood at the Register. You know, you know that it's, it's the, the fact that they have these rules that seem to be trying to pull back on it out of interest for individual privacy. And that we meanwhile are living in the Dolan's world is a, is such.

Ian Thompson (148:21)

A weird thing you've got to say when even China is saying, hang on a bit, let's just, let's just slow this down little bit. Even China, you know, and of course.

Leo Laporte (148:31)

Some of the issue is false positives. Right. You know, I, I have mixed feelings about this. Taylor Swift was, uses facial recognition at her concerts to keep harassers and stalkers out. And I think that seems like a good use as long as it doesn't.

Jacob Ward (148:51)

Mistakenly identifies harassment and stalking. Right. But there's no law, not in the, in a first Amendment regime against having a T shirt that criticizes the owner.

Leo Laporte (149:00)

No kidding.

Alex Kantrowitz (149:00)

Well, he's within his rights. It's just that this is such a thin skinned move to do something like this. And as a Knicks fan, as someone who likes to go to the sphere, I will say no more.

Leo Laporte (149:14)

And that was Joey Cantorowitz, our, our Brooklyn columnist.

Ian Thompson (149:22)

Also, is the sphere that good? Because I've, I've seen footage and it does look like a bit of a money pit.

Leo Laporte (149:26)

I'm dying to go. Are you?

Alex Kantrowitz (149:28)

You better use my Muppets face for the Thumbnail.

Leo Laporte (149:32)

It's too late. I'm dead. I've, I've given Dolan way too much trouble in the past. I, I'll never get into anything like that. Wouldn't that be wild though? I almost want to be stopped because then I could, you know, like make something out of it. There'd be a story there.

Ian Thompson (149:46)

Well, I mean there's, there's already anti paparazzi clothing, you know, scarves you can wear, which, you know, muck around with cameras. I'm sure something will come along like that.

Leo Laporte (149:55)

Yeah. All right, we got more coming up in just a bit. You're watching this week in tech with a great panel. Alex Cantrowitz, Jacob Ward, Ian Thompson, so glad you're here this week. Moving on. Gemini. You know, we've been talking about AI and ChatGPT and perplexity. We're not talking a lot about Google's AI Gemini 2.4 Pro, which is currently at the top of the leaderboard. And now Google's done in something interesting because people aren't talking about it as much as it's competition. They're making that available for free to its users. I have a paid Gemini Pro account, but anybody can use Gemini 2.5 Pro. It says experimental. There are rate limits on it. It's funny, I'm now going to have to think about using it because I never do. It's, I mean they have a perception about it, right?

Jacob Ward (150:53)

Is the, the, the, it's sort of, it, it reveals the degree to which these companies are subsidizing what is incredibly expensive processing compute it takes to put together something is amazing. I've just been recently thinking, talking to a bunch of technologists about the potential of building AI tools for journalists. And I've been playing around with one of the fundamentals, fundamentals of, of what would be necessary to do that. One of them is confidentiality. You got to be able to, you know, feed stuff in and not let it go anywhere else. So you can't use a subsidized model like this. The reason Google would subsidize the use of this and make it free to non paying users is because then they get that data back to do more training. That's one of the benefits of doing so. And if it turns out if you're not going to do that, if you're not going to let them have the data, it's expensive. You know, a tool token is about a like two thirds of a word. So I was looking at it, I'm like, wow, if you put Together like a. You know, the complete financial documents of a company for a year into that and get two pages of summary out of it. That's like five bucks if you're really paying the market.

Leo Laporte (151:57)

But how useful is that? I mean, that's incredibly useful.

Jacob Ward (151:59)

Super useful. But how many times can your average substacker afford to do that before they start wiping themselves out, you know? And so to see them handing over over five bucks, you know, per. At least to. To people, is really amazing. To me, that just suggests just what a bath they're willing to take on this stuff.

Leo Laporte (152:19)

Yeah, well, and it calls us back to the beginning of the show where Sam Altman said, please stop using our AI generation, image generation, because it's costing us. The servers are burning up. Um, all right, well, I think. I think I've used up all the stories in my pile, so there's only one thing to do, is call it a day. I am so glad to have all three of you on here, Alex. Keep up the great work. Alex Cantrowitz. You know, you all should subscribe to the big technology substack and podcast. It's awesome. What great guests. And some mystery guests. This week. There'll be big news.

Ian Thompson (153:02)

News.

Alex Kantrowitz (153:04)

Is that for a tease? Yeah, it's gonna be fun. It'll be a fun episode. I promise that.

Leo Laporte (153:08)

Can we do 20 questions? Is it. Is it a man or a woman?

Alex Kantrowitz (153:12)

I mean, I. I will add, in service of taking extreme caution here and not exposing and losing the interview questions, but I will promise you will enjoy it. So tune in on Friday.

Leo Laporte (153:26)

Okay, here's a secret that I will reveal. Reveal next week. Jason Calacanis is going to be on the show. So.

Ian Thompson (153:34)

And Alex as well.

Leo Laporte (153:36)

So if you have any opinion about Jason, good or bad, I want you to tune into the show next week. I will tell you he's going to be on, which means he's definitely not going to be on the big technology podcast.

Alex Kantrowitz (153:51)

It's not Jason that's the clue I can quote. But Jason, if you're listening, come on some time. We'll talk.

Leo Laporte (153:56)

He's fun. He's an interesting. You know what? As we get closer to our 20th anniversary in two weeks, I think more and more about the people who made this show, you know, succeed over the years. And Jason was on a lot, kind of in the middle period. Great character. He used to do the best audible commercials for us, so.

Alex Kantrowitz (154:16)

And congrats, Leo, by the way, on 20 years. That's not 20 years. It's pretty impressive.

Ian Thompson (154:21)

That really is Congrats.

Alex Kantrowitz (154:22)

To you.

Ian Thompson (154:22)

Wow.

Leo Laporte (154:23)

We are asking people send in your videos. You can either put them on social and@twit or send them to me leoville.com and we are going to get as many of those videos how you watch twit when you discovered Twit your twit story. So we can run those because I think really the story of twit is absolutely the story of its community. An amazing group. We love all of you and we thank all of you for this long time support. Jacob, are you now part of the family? Jacob Ward. We kept it under three hours, barely.

Jacob Ward (154:55)

I'm amazed you guys, I don't know how you like. I was snacking the whole time. I always took like a sip of laqua. Like what the hell. I don't understand.

Alex Kantrowitz (155:05)

This is done.

Jacob Ward (155:06)

It's incredible.

Leo Laporte (155:06)

I don't know how many pod do it.

Jacob Ward (155:09)

As long as chops.

Leo Laporte (155:10)

You know you have to work your way up to it. It's actually I think sprinter.

Jacob Ward (155:14)

I don't.

Leo Laporte (155:15)

I'm going to look actually, but I think the first twit 20 years ago was I think a half an hour.

Ian Thompson (155:22)

Wow.

Leo Laporte (155:23)

Okay.

Ian Thompson (155:23)

That's Mission creep has come in.

Leo Laporte (155:26)

That's my guess. But let me see. I think I can go there If I go twit TV twit one, the first episode was 34 minutes and 44 seconds. Wow.

Jacob Ward (155:41)

That's a pretty reasonable amount of time, Leo.

Leo Laporte (155:44)

I'm just saying that's how long your shows are.

Jacob Ward (155:46)

Yeah, exactly.

Leo Laporte (155:47)

But 20 years from now, now you'll be five hours, six hours. You'll be all day. Trust me.

Jacob Ward (155:51)

You're just, you're, you're, you're reaching that super commuter audience.

Leo Laporte (155:57)

You know, that was, I think the secret was as people's commutes got longer in the past 20 years, it's really been beneficial. Even with the times Joe Rogan and me. That's it. Jacob Ward get his book the Loop. I did. I just ordered it. Follow him@theripcurrent.com. that's where his newsletter and podcast, our brand new podcast podcast. Great guests already. Thrilled for you, Jake. It's great.

Jacob Ward (156:22)

Appreciate it.

Leo Laporte (156:23)

Leo. Welcome to the independent creator verse. It's been good to me, it's been good to Alex. I'm sure it'll be great for you too.

Jacob Ward (156:32)

Well, I appreciate it.

Leo Laporte (156:32)

Thank you. Jacob. We'll have you back soon. I hope I. I say I hope because I hope you will come back.

Jacob Ward (156:39)

Oh, dude, I'm in.

Leo Laporte (156:40)

I'm in. Okay. I promise you we will invite you then. Just Hope you'll say yes. Ian Thompson. How many you've been here for? 20 years. Years, almost. It's been a long trip. 15.

Ian Thompson (156:51)

Yeah, 15 years now. It's.

Leo Laporte (156:53)

I know because the picture of you is very old.

Ian Thompson (156:58)

Well, to be fair, you know, I was younger and pretty in the prettier in those days, but yeah, I would.

Leo Laporte (157:03)

Stick with that picture. I think it's a good picture of you.

Ian Thompson (157:05)

I came over.

Alex Kantrowitz (157:06)

Yeah.

Ian Thompson (157:06)

On a one year contract and then just kind of kept on extending, you know.

Leo Laporte (157:11)

Were you. Did you start with a register in the. In the uk? I feel like you did.

Ian Thompson (157:14)

No, no, I was with. The original plan was I was with a site called vnu. Net, which is now being expunged from the Internet. So, yeah, I was going to come over, hire an American lead and then go back and run the site, but then there was a re. Organization, so I stayed with them for three years and then moved to the register.

Leo Laporte (157:32)

Nice. Yeah.

Ian Thompson (157:33)

Well.

Leo Laporte (157:34)

So you've been registered just as long as you been on our show? Yes, pretty much.

Ian Thompson (157:38)

Yeah. Yeah, yeah. In fact, that's when I started doing it. I think maybe it's back when Megan Maroney was doing look how cute.

Leo Laporte (157:45)

Look how cute he was 15 years ago. Just a lad. Just a lad. He's taught us so much about the English language. And by English language, I mean the language of England.

Ian Thompson (157:59)

Yes.

Leo Laporte (158:00)

Yeah.

Ian Thompson (158:00)

Some phrases have got through grips. My muffin is always a popular one. But yeah.

Leo Laporte (158:05)

So many shows that you've been on titled with some of those anglicisms.

Ian Thompson (158:11)

In one case, quite rude.

Jacob Ward (158:13)

Yeah.

Leo Laporte (158:13)

I didn't know what it was. How was I supposed to know? Yes, but two countries separated by one language.

Ian Thompson (158:20)

Indeed. Churchill was right.

Leo Laporte (158:21)

Yeah. Wow. There are nine pages of appearances.

Ian Thompson (158:25)

Good Lord.

Leo Laporte (158:26)

Let me go back to the earliest appearance. Tech news tonight, number 34 in 2014.

Ian Thompson (158:33)

Wow.

Leo Laporte (158:34)

You appeared a few times in Tech News Tonight before we brought you up to the big leagues.

Alex Kantrowitz (158:39)

Yeah.

Ian Thompson (158:40)

And it's only a half hour show, so this isn't three hours.

Leo Laporte (158:42)

I really appreciate it. Thank you.

Ian Thompson (158:44)

No, it's always great fun. Yeah.

Leo Laporte (158:46)

You were on this week in tech 500.

Ian Thompson (158:49)

Wow.

Leo Laporte (158:49)

525 episodes ago.

Ian Thompson (158:51)

Good grief.

Leo Laporte (158:53)

Wow.

Ian Thompson (158:54)

Thank you. When you're having fun.

Leo Laporte (158:57)

Thank you. Alex, Ian and Jake. We really appreciate your time and your brilliant thoughts. Thanks to all of you who listen as well. It's. What a long, strange trip. It has been. 20 years and two weeks. Do send us your videos. I'd love to hear how you found us and how you watch. Of course, if you're a member of Club Twit. You can also just post it into our Club Twit Discord. That's a great place to hang. Seven bucks a month gets you ad. Free versions of all of our shows, gets you access to the Club Twit Discord. Lots of benefits, special events, special shows. I think, come to think of it, don't I have a show coming up this week? Yeah. Photo Time with Chris Marquardt is Thursday, April 3rd. So he's our favorite photographer. We get him on every month. Brilliant, by the way, is the theme of the photo assignment. So if you haven't yet submitted a brilliant photo to Flickr to our tech guy group on Flickr, do that. We'll also be doing Micah's crafting corner on the 16th. 16th he's doing Lego plants. And on the 18th, Friday, coffee time is back. Mark Prince will bring Liz Happy beans together. We will talk about coffee and then the AI users group at the end of the month. And that's just the month of April. Seven bucks a month. You get so much benefit. And I hope you will join us because it makes a huge difference to us in terms of our bottom line and it will help me towards my goal of buying the Starship Enterprise because I've always wanted to be a starship captain. That's pretty good. I like it. Thanks for being here, everybody. We'll see you next week. Jason Calacanis and Alex Wilhelm. It's kind of a this weekend, startups takeover. The following week, our 20th anniversary. But as I have been saying for 19 years and 50 weeks, thanks for joining us. We'll see you next time. Another twit is in the can.

Ian Thompson (160:58)

Bye.

Leo Laporte (160:58)

Bye. He's amazing.