Ian Kreitzberg (2:47)

Yes, I am, yeah. Good to be here.

Leo Laporte (2:49)

It's nice to have you. What did. Is this something you've been interested in for a long time or did this suddenly come up?

Ian Kreitzberg (2:57)

Yeah, I've been doing this for a couple of years. It probably tracks the whole Rise of ChatGPT, right? I was doing business and tech journalism at the street and my focus was on, you know, Elon and the Mag seven guys.

Leo Laporte (3:14)

Oh, that must have been fun.

Ian Kreitzberg (3:16)

Oh, yeah. Every day was just a fresh.

Leo Laporte (3:19)

You know, it's gotten weirder and weirder.

Ian Kreitzberg (3:21)

Actually, it has gotten. I am very glad that now my sole focus is AI, because as weird as that gets, I don't have to go through Elon's Twitter profile every minute of every day. And so that's a major improvement.

Leo Laporte (3:37)

Well, and AI, you know, look, I've been covering technology since, believe it or not, since the 70s, and AI is still steam powered.

Jeff Jarvis (3:46)

Back then, I was going to say.

Paris Martineau (3:47)

People are saying there wasn't even technology back then.

Leo Laporte (3:51)

There was, but it wasn't great. But the point being, people, I've seen a lot of things come and go, and I've even seen a few AI winners in my time, but this is perhaps the most interesting technology breakthrough getting, by the way, more interesting by the minute. So you've got a great beat.

Ian Kreitzberg (4:10)

Yeah, there's definitely no shortage of stuff to do, stuff to cover. I'm always. I never have the problem of, you know, trying to make something out of nothing, trying to find the next story that there's always just so much going on that you kind of got to just be really picky and choosy and drill down on what you actually want to focus on.

Leo Laporte (4:32)

That's been my experience, too. We have more stories than we can cover here. This is Puck News and there's Ian's. Actually, this is the first thing I was going to ask you about, the AI action plan. There's David Sachs over Donald's left shoulder. He is the AI and crypto czar. And this is funny because David Sacks is also the co host of a podcast with Jason Calacanis called All in and they hosted this presidential announcement. Maybe a little. So I'll confess, I was, you know, but I didn't have to sit down with J.D. vance and pretend I was talking to an intelligent person. So there's that. Anyway, tell me about this big, beautiful LLM as you. As you call it. Mike Masnick said there's some good stuff and there's some really dumb stuff. Stuff.

Ian Kreitzberg (5:29)

Yeah, that's probably a good. Probably A good summarization of it.

Leo Laporte (5:35)

First of all, they threw out what Joe Biden had done in 2023, right?

Ian Kreitzberg (5:39)

Yes.

Leo Laporte (5:40)

That's gone.

Ian Kreitzberg (5:41)

The first thing that Trump did when he took office is it's like the flurry of executive orders that he throws in there. And one of the first ones was just resending Biden's executive order on AI.

Leo Laporte (5:55)

Was this because he had been elected by Silicon Valley and especially by people like Peter Thiel and their interests in AI didn't coincide with Biden's desire for safe, secure AI. Was that. I mean, was this at the behest of Peter Thiel? Is I guess, what I'm asking.

Ian Kreitzberg (6:12)

It certainly seems like it that. I mean, and not even just Peter, but, you know, the Sam Altman and.

Leo Laporte (6:21)

Larry Ellison and Sunson who came in and did that Stargate announcement of half a trillion dollars in AI. Yep.

Ian Kreitzberg (6:28)

And before the whole falling out, his relationship with Elon.

Leo Laporte (6:32)

I forgot about Elon already.

Ian Kreitzberg (6:33)

You can't forget about Elon.

Jeff Jarvis (6:35)

Who wouldn't want that?

Leo Laporte (6:38)

Yeah, Elon is probably lobbying for this, too. So what did. So what did. What if. What is the new regime?

Ian Kreitzberg (6:46)

The new regime is very expectedly a deregulatory regime. It's all like. Everything that we're seeing is very much centered around this kind of narrative about the race with China. However, that's actually playing out whatever is actually at the finish line with that race, which to me and a lot of people is very unclear. That is a narrative that the tech companies that are talking to Washington have been very successful in playing up, and at the kind of result of that is this idea of national security as the justification for everything.

Leo Laporte (7:34)

In other words, if we don't do it, China will, and we've got to defend against that. Right, Right.

Ian Kreitzberg (7:39)

And so that's become the kind of overwhelming regime, as you put it, where that's influencing a lot of stuff. And the idea at its root of deregulation and the way they're describing it is that if we regulate, that will stop innovation, and we can't afford to stop innovation right now. It's not at all clear. And actually we've seen the opposite in many cases, that regulation does good things for innovation. And I think it's kind of funny to take that approach specifically to go after China when China is one of the most regulated countries when it comes to AI. They have very specific regimented rules of the road here. And our. Our approach in terms of winning that race is, what if we had no rules? So we'll see how that pans out, but the, the whole theme is just don't regulate. We can't regulate, we can't get in the way. And it's very much pitting regulation against technology.

Leo Laporte (8:50)

They did say they took 10,000 comments.

Paris Martineau (8:56)

Did they say they did anything with those comments?

Ian Kreitzberg (9:01)

Yeah, 10,000 comments.

Leo Laporte (9:03)

And yo we generated by AI. No doubt. Right.

Ian Kreitzberg (9:08)

Probably all the, the comments are interesting. Obviously you can't read all of them.

Jeff Jarvis (9:12)

Right.

Leo Laporte (9:13)

But they did put them all online though. Yeah, yeah.

Ian Kreitzberg (9:17)

It's just a lot to read, I'm saying, you know, for one person, unless you had chatgpt like process all of it and summarize it, but a bunch of the major companies have publicly.

Leo Laporte (9:30)

Here's what a concerned American wrote. I think the AI has done and will continue to do more harm than good, especially without limits, if this act goes forward. Were doomed in all in caps, by the way.

Paris Martineau (9:44)

Well, why are you reading that with dumb person voice?

Leo Laporte (9:47)

No, that's smart person. I apologize. I apologize.

Paris Martineau (9:50)

Sorry. They sound so similar.

Leo Laporte (9:55)

Did you read. How many of them did you read? Did you read a lot? Did you try to read a lot of them?

Ian Kreitzberg (9:59)

I didn't try to go through a lot of them. What I did read were the kind of the policy proposals from the, from the big companies. Yeah, yeah.

Leo Laporte (10:08)

I bet there's an OpenAI comment in here somewhere. Yeah.

Ian Kreitzberg (10:11)

Oh yeah, yeah. There's an OpenAI, there's a Meta, there's a Google, I'm sure.

Leo Laporte (10:17)

Are they similar in their, in their needs or desires?

Ian Kreitzberg (10:23)

There's definite through lines. And the interesting thing is if you read those and then you read the final action plan, which I mean, I would encourage people to do because it's really not that long.

Leo Laporte (10:32)

It's also, we should point out, not the law of the land. It's just an idea. It's a good idea. Right? Yeah.

Ian Kreitzberg (10:40)

I think the best way to describe it is like an optimal wish list of things to do. The funny thing is that, or I don't know, maybe this is not the funny thing, but there's just so many interesting elements underlying all of this. And one of them is that through doge and proposals made by the Trump administration, a lot of the agencies that were tasked with carrying out or identifying how to carry out a bunch of the wish list items in this roadmap might not have the funding and or capacity to actually go through and do them. So it's not like this is a kind of ideal wish list roadmap scenario that it's not clear. Like you said, it's not a law of the land. And it's very unclear at this stage how those agencies will actually go about implementing some of these desires. We can call them.

Leo Laporte (11:34)

This is what OpenAI wrote. OpenAI agrees with the Trump administration that AI creates prosperity and freedom worth fighting for, especially for younger generations whose future will be shaped. How this administration approaches AI as America's world leading AI sector approaches artificial general intelligence with a Chinese Communist. I know, Jeff. I was waiting for Jeff. As it With a Chinese Communist Party determined to overTake us by 2030, the Trump administration's new AI action plan can.

Jeff Jarvis (12:07)

Ensure that not just China or Chinese industry, the Chinese Communist Party.

Leo Laporte (12:12)

Yeah.

Ian Kreitzberg (12:13)

Oh, yeah.

Jeff Jarvis (12:13)

Yep.

Leo Laporte (12:14)

Well, and part of the plan is to kind of go hand in hand with like Saudi Arabia to do this. So I'm not convinced that they're less repressive than the CCP anyway. Okay, so let's talk about Mike Masnick said, by the way, as long as we're quoting Mike Masnick, one more Mike Masnik quote talking about authoritarian governments. He says real innovation happens when companies have to compete on merit, not who can kiss the leader's ass most effectively. In an authoritarian system, the company that makes the dictator happy wins and that's it. So, you know, and you could say, well, if he's talking about China, that's why China won't win. And he could also say, well, maybe that's what we're headed to in the United States. I don't know. Let's talk about the good first. I do light touch regulation on AI seems like a good idea. Yeah. Or no? What do you think?

Ian Kreitzberg (13:14)

It depends. My. Yeah, I mean, that's going to be my answer with everything it depends.

Leo Laporte (13:21)

Leo. Fair enough. That's not untrue.

Ian Kreitzberg (13:26)

My thing is that light touch or heavy touch should be less the focus and more the focus should be on how do you create a regulatory regime for something that is changing quickly and that is robust to handle as it changes and as it evolves? And I think, I mean, the tough thing is that the potential for misuse is just so remarkable and the potential for harm is so remarkable in such a number of ways that are not really covered by existing regulation. And there's these outstanding questions that would not be a bad thing to instate protections. If you think of the auto industry, that industry did not want seatbelts. Regulators had to say, listen, seatbelts, we like seatbelts, we're going to do them. And that's a positive innovation now. So if regulators were able to think a little Broader than a large language model, we might be able. That's what I would want to see. Think broader than the specific technology. Because AI is such this broad, vague term that so many other things apply to and things change. And if you try to create compute requirements, what if in two years we have really efficient models that get around the compute requirements but pose the same harm? Right, so there are.

Leo Laporte (14:53)

That's a good point. Yeah, yeah. There's some evidence that that's happening in China. Yeah, yeah.

Ian Kreitzberg (14:59)

I mean, all these guys are pulling on efficiency. Efficiency is a big thing. Some of them for different reasons. I think largely it's cost. But we want more efficient models and that's a thing that, that happens as things progress.

Leo Laporte (15:15)

So you can't say, well, if you're, you know, using 100 more H, 100 GPUs, if you're using a thousand or more, then you're under regulation. You can't say stuff like that because we just don't know. Or if you're $100 million company or a billion dollar company, you can't. Yeah, that makes sense. But at the same time, we don't want to stymie the garage developers. The little, little guys. Yeah.

Ian Kreitzberg (15:41)

Which is why for some, I mean, especially in the states. Right. The states are the guys that are kind of saying, now, you know, we're going to take a shot at actually regulating this. And their idea, at least in New York, for example, is these kinds of sort of exemptions for universities. And that's part of the attractiveness of a compute threshold or a cost threshold or a fiscal threshold is that if I say I'm going to only regulate models whose developers spent more than $100 million to train, then you're not hurting the garage innovators that are just tinkering with a GPU they picked up and are building something new. You're really just saying, I'm regulating the big tech, capital B, capital T guys and making sure that they're following some sort of rules of the road. So that's a very targeted approach to make very sure that a university is not really spending $100 million to train a model. They don't really have that. But this opens them up where if we only regulate it, that fiscal threshold, then innovation and kind of base research can still happen. But yeah, I don't know. I don't know if that's quite the best way forward.

Jeff Jarvis (17:13)

What's the most important regulatory thing to do?

Ian Kreitzberg (17:17)

What's the most important regulatory thing to do? That's a very good question. I think that would, the answer to that would depend on the kind of different sex of people that are dealing with this. I think there's a big job factor and like the AI workplace factor. And are you like the. In keeping with my idea of kind of creating a regime that's, that's broad and applicable, I would say to me the most important thing is automated decision making and rules according to cross industries, specifically governments, you know, police force, military rules around where automated decision making is allowed, where it's not.

Jeff Jarvis (18:09)

So it's not, it's not the AI company in that case you're, you're regulating, it's the users of AI. It's government using the applicability. Right. It's at the application layer.

Ian Kreitzberg (18:20)

I think that would become important. I think for a lot of regulators it's more easy to think of targeting the companies because it's a little more obvious. But there's a lot of other stuff going on that is a cause for a number of different concerns. And I think a lot of that does come down to the idea of automated decision making even at the job level. Is someone choosing to hire me based on an AI? Are they involved in that loop at all? Is that fair? How do we determine if that system is, you know, exhibiting bias to something in my resume that, you know, the human employer wouldn't do. Or you know, extrapolate that out to military and policing examples with predictive policing which we know has really bad outcomes. And you know, in the military bombing targets selected based on a potentially faulty AI system where a person is insulated from accountability because the AI maybe got it wrong or it was the AI doing something. And so accountability around decision making I think is something that is important to enshrine. That wouldn't have compute thresholds and fiscal thresholds. And is this going after universities or is this going after big tech corporations? And so I think that would be important place to kind of start.

Leo Laporte (19:45)

It is ironic that the most terrifying use of AI is by the government. And of course it's the government that's making these rules. So I can't expect them to regulate themselves. On the other hand, there's some good stuff. The plan calls for a stronger electric grid. We all benefit from that. I mean, obviously the reason they want it is to power AI, but we could all benefit from that. Unless it's all coal based.

Ian Kreitzberg (20:11)

That's the thing.

Paris Martineau (20:13)

And we're already starting to see there have been a number of reports of power bills going up for the average consumer in areas.

Leo Laporte (20:21)

Oh yeah, we Pay more so they don't have to. Yeah, well, that's kind of the rule in the.

Paris Martineau (20:26)

So, I mean, there'll be a stronger electric grid, but who will benefit from that?

Leo Laporte (20:29)

Yeah, that's true. Higher security standards for data centers. Okay. More US made semiconductor chips. That's going to be a challenge. We already, we talk about that a lot on some of the other shows, the streamlining of the permitting process.

Jeff Jarvis (20:44)

But is that federal? That's. Local governments. What is. What is the. That's.

Leo Laporte (20:48)

Well, they want to allocate federal lands for data centers.

Jeff Jarvis (20:51)

Well, okay, well, there's that. There's that. Yeah.

Leo Laporte (20:54)

You know, I think there should be a Grand Canyon. Yeah, sure. Absolutely. Grand Canyon is perfect. It's deep, it's quiet, it's got a river, it's, you know, perfect.

Jeff Jarvis (21:06)

I think Alaska. It cools it all down. We can melt the ice cap faster. You know, it's.

Paris Martineau (21:10)

Yeah, yeah. Could be nice.

Leo Laporte (21:12)

So you quote executive director of AI now, which said the plan reads like the wish list for Silicon Valley's big AI firms. It doesn't just read it like it. It is.

Ian Kreitzberg (21:23)

It is.

Leo Laporte (21:24)

Yeah. That's. They got what they wanted. If you look at the comments from.

Jeff Jarvis (21:28)

They wrote it.

Leo Laporte (21:29)

They wrote it. In a way.

Ian Kreitzberg (21:30)

They wrote it. The only thing they didn't get enshrined in the plan, but it seems like something on this is coming is copyright. And, you know, I'm reading through the plan and I'm looking for the thing on copyright and it's not there.

Leo Laporte (21:45)

He made that up at the end. He ad libbed that.

Ian Kreitzberg (21:48)

Exactly.

Leo Laporte (21:49)

Yep.

Ian Kreitzberg (21:50)

We go in the podcast, speech, unveil, and he starts talking about copyright. And it is written by, you know, that part is coming straight from the mouths of the valley as well.

Leo Laporte (22:01)

He heard it at dinner last night at Mar a Lago and now he's repeating it. Although people like Kathy Gellis, who's a regular on our show and a constitutional lawyer, says no, that enshrines the right to read. That enshrines the First Amendment.

Jeff Jarvis (22:15)

I argue that constantly here.

Leo Laporte (22:16)

So Trump, if you didn't know at the end, said, you know, AI should be able to read anything. It should be able to ingest anything. Copyright shouldn't be able to stop it. And that's what Kathy Gellis thinks. Although there's quite a debate over that, especially from the people who write the things that the AI is ingesting or paint them or play them. You're a musician. How do you feel about that?

Ian Kreitzberg (22:42)

This was actually my kind of first Experience with AI, with this idea of kind of the auto generation of music.

Leo Laporte (22:51)

And stuff like Suno, which we kind of blown away by, frankly.

Ian Kreitzberg (22:55)

Yeah. I mean, you see these kind of very impressive examples of technology that I in some way just kind of disturb a part of me that processes emotion by sitting at a keyboard. I feel like so much of this push and specifically the first area these guys talk about the big thing, curing cancer and solving climate change. Of course they talk about this while making data centers that are not helping climate change.

Paris Martineau (23:30)

Not really helping curing cancer. At least it's a little carrot they.

Leo Laporte (23:34)

Dangle while they actually go for.

Jeff Jarvis (23:36)

If only we free them up to do anything, then they can try.

Ian Kreitzberg (23:39)

Right, Exactly. If we. As long as copyright's clear, then they can go on to solve cancer. Because they need books in that model. But the thing they're going after is media. And that's been. It makes it. I feel like I've talked with family and friends about this idea of this kind of crisis of humanity where, okay, cool. We've got a system that if you prompt it, hey, Chat or Claude or whatever, write an article about AI and the voice of Ian Kreitzberg.

Leo Laporte (24:13)

It will do that because it's read everything you've written.

Jeff Jarvis (24:16)

Right.

Ian Kreitzberg (24:16)

Will it be good? No. Will it make up quotes? Will it make shit up? Of course.

Leo Laporte (24:22)

But it'll sound just like me.

Ian Kreitzberg (24:24)

I don't know. I tried this recently and I don't think it sounds like me.

Leo Laporte (24:29)

Oh, good.

Ian Kreitzberg (24:30)

But I don't know, Maybe that's just one GPT number away. Maybe GPT6 will sound like me. That's where they're angling towards. And when you're dealing with these kinds of systems. Yeah. I wouldn't be surprised if you could get to a point where it could. If I asked it, if I trained it on all my voice memos and I said do a five minute piano improv in my style, I'm sure it could do that. And then you get into that question of, okay, we have something that can do this. Do we want it to do this? Do we want just an influx of fake content that not only hurts the business of the people who made it, which is not really a great business to start with, but I like consuming and interacting and engaging with art made by people. Because art is fundamentally a way of, like a kind of empathetic way of connecting, seeing and being seen and to eliminate or further remove the human element. And something that is just so innately human just feels kind of wrong.

Leo Laporte (25:43)

Very well, very well said.

Jeff Jarvis (25:45)

Yeah, I'm interested in this beat is so fascinating to me as a journalist. If you covered the Internet or you covered personal computers or other things earlier on, it's not the same mix because you have technology, you have business, you have policy, but you also have culture, you have the whole test Royale Doomers, you have the accelerationists, you have the teals and the musks. How do you approach, how do you think you're going to mix your coverage among those four pieces of pie?

Ian Kreitzberg (26:23)

Yeah, big pie pieces. My goal is to kind of sit where the science is. And the challenge with this technology, compared maybe to some others, is that we are dealing with a good amount of hypotheticals and theoretical science. And so my balance of those pie pieces is kind of like, okay, it's easy to say stuff, especially when those impressions, like we have a lot of science fiction to kind of pull on. Where's the evidence for that stuff as we have it today?

Leo Laporte (27:08)

Right.

Ian Kreitzberg (27:08)

You never know what we're going to have tomorrow. But based on today's architectures, the history of science and technology and AI as it has led up to today, this moment, what do we have hard evidence for? And all these papers that come out are these robust papers. Are the results of your experiments reproducible? Are you just seeing something that you want to see? And because AI communicates in language and uses self pronouns, it's easy for you to kind of, you know, see something in the mirror that's not really there. Right. And I think a lot of what.

Leo Laporte (27:40)

The FDA is doing right now. Exactly, right, yeah. Oh, it's just common sense. Yeah. There's this. One of the reasons I don't take this AI action plan too seriously is because there's also this discordance between what the Trump administration is doing and what it's saying. For instance, they cut or they're proposing to cut, as you point in your article, $325 million to NIST. They, they want a 55% cut to the National Science Foundation. At the same time as they're saying, we believe in science, we believe in advancing, they're cutting to the bone. All of the parts of this country that do this kind of fundamental research. Is it that they just want private enterprise to do everything, or is it that they don't really have a plan? You say that maybe the left hand doesn't know what the right hand is doing. Yeah, it's hard to say.

Ian Kreitzberg (28:40)

I think a lot of stuff with this administration definitely seems like, I don't know if you really have A plan?

Leo Laporte (28:45)

No. That's why I don't take it too seriously. It's like. Okay. It's more words really.

Paris Martineau (28:50)

I'm curious, do you think it is at all related to. I'm sure as you've seen, I think over the last week we've had Elon Musk. Musk as well as I think maybe it was either Sachs or Calcanis. Kind of come out against academia as this terrible.

Leo Laporte (29:05)

Oh, this is Marc Andreessen and Musk.

Paris Martineau (29:09)

The last day or two.

Leo Laporte (29:10)

The guy who became rich and famous because a university sponsored his development of the browser.

Paris Martineau (29:17)

Yeah.

Jeff Jarvis (29:18)

There's been this very intellectual property.

Leo Laporte (29:20)

Yeah.

Paris Martineau (29:21)

Hard and concerted push towards the ladder thing off yet pull up the ladder and cast off all of academia as some kind of superfluous, overly expensive and myopically concerned group. Do you think that that in some way plays a role with this kind of strange right hand, doesn't know what the left hand is doing that you're seeing in legislation like this.

Ian Kreitzberg (29:50)

That there's definitely. And the anti kind of university and all the problems with education. That's something that we've seen a lot from this administration that recently decimated the Department of Education. It's very interesting you have this concerted effort to it seems, change for the next generation where information is coming from and who they should trust and receiving that information. And this push of AI tutors and peer tutors and to me a risk of kind of laundering information, taking people further away from first sources and just dramatic disruption is what we talk about. But it's very interesting again in that context of you're going as hard as you can on AI, you want to win, you want to American dominance. That that's the language which is very strong language. But you're decimating all these scientific infrastructures. And the reality is that these where we are today, the transformer and neural networks and machine learning resulted over decades of robust societal infrastructure across not just nist, but I mean also one that a lot of people don't really talk about. Like NASA has done crazy work in advancing the state of the art in these areas, robotics and machine learning, because you have agencies that were finding ways to solve for the specific missions they were dealing with. But then it's also like immigration is one that has become a challenge now you want this robust infrastructure that supports the development in universities of new innovations and new ideas. And here, you know, you mentioned right at the top that you've seen a few AI winters and stuff. And I think there's possibly a Concern of if we don't enable like what comes next. Right. If language models are not the end all be all. And I don't think they are. Do we have a societal infrastructure in place to support the next innovation? The next attention is all you need paper.

Leo Laporte (32:13)

Right.

Ian Kreitzberg (32:14)

And that is being slashed and rolled back and stuff. And so there's a focus on the corporate side of what we have today, of the chatbot side really of what we call AI.

Leo Laporte (32:29)

This is the inshitification. Frankly, I have a theory and it's. I think one of the reasons Elon Musk became disenchanted is because there are conflicting beliefs in the conservative movement. On the one hand, this historic anti intellectualism. I mean, Adlai Stevenson was called an egghead and that's why he couldn't get elected as president because he was an academic and we don't like the Ivy Leagues. Everything bad ever came out of the Ivy Leagues.

Jeff Jarvis (32:57)

Pinheads.

Leo Laporte (32:57)

Pinheads. There's also this notion of a business should be allowed to run without regulation and business should be. That's another common thread. And then there's this equal thread of fear of the other coming in over our borders and swamping our culture. But the problem is there are these conflict. You can't hate academia and want AI to succeed. You can't say, well, we don't want anybody coming over our borders because we don't want those H1B visa types. But that's. If you look at who Meta is hiring, they're for the most part immigrants running the superintelligence lab who came in because we had the best educational system in the world.

Jeff Jarvis (33:40)

Charge a lot of these companies and.

Leo Laporte (33:42)

They came to America to learn about AI and A. And so we've got. I think you have an internal contradiction that is ultimately going to be very hard to resolve. I do think it's why Elon Musk became disenchanted. He thought. He thought he had it all under control, but he realized, no, there's a big conflict here. Yeah. I mean if there's a silver lining, it's that this can't possibly work. Unfortunately, it's at the same time the depressing part, because we're the ones who are going to get the consequences of all this thing not working, all the.

Paris Martineau (34:19)

Coming crashing to the ground.

Leo Laporte (34:20)

Yeah. We're talking to Ian Kreitzberg. His brand new newsletter must read the hidden layer. Do you have to subscribe to Puck News to get the hidden layer?

Ian Kreitzberg (34:31)

You do, but when you do that, you get access to all of Puck.

Leo Laporte (34:35)

Which is so worth it. I don't want to compare it to. I don't know. It reminds me of some of the snarkier publications of an earlier Internet era.

Paris Martineau (34:46)

There's just great stuff, but with great reporting.

Leo Laporte (34:49)

Great reporting and all about Hollywood.

Jeff Jarvis (34:54)

Dylan Byers. Julia.

Leo Laporte (34:56)

Some of the best people.

Paris Martineau (34:57)

Politics.

Leo Laporte (34:58)

Yeah.

Paris Martineau (34:59)

Got a newsletter for you.

Leo Laporte (35:00)

Yeah. If you so subscribe to Puck News and then sign up for, well, frankly, all the newsletters, but I think you got to get the hidden layer. You just. Actually, I didn't realize this. Did you push this out while you were getting ready for the interview? Because the AI Cold War. Did this just come out, or is they just spot it?

Ian Kreitzberg (35:20)

It came out yesterday.

Leo Laporte (35:21)

They just spotlighted it. I get it. The United States and even Europe are rapidly cutting red tape, curtailing their own regulatory guardrails to advance the interests of the private sector, all in the name of taking on China. What could possibly go wrong? It's exactly what we were just talking about. Exactly.

Ian Kreitzberg (35:37)

Precisely.

Leo Laporte (35:38)

Yeah.

Paris Martineau (35:39)

How frequently do you publish?

Ian Kreitzberg (35:41)

We go twice a week. So Tuesday evenings and Thursday evenings. I'm jumping into your inbox.

Leo Laporte (35:49)

Nice. Ian, I'm so glad that they hired you. I. I've been a big fan of Puck. They seem to have a real eye for talent. The best writers now working there. And I'm so glad the writers are in charge.

Jeff Jarvis (36:01)

Right.

Leo Laporte (36:02)

I mean, that's why. Right. Yeah. Yeah. Makes sense. Yeah.

Jeff Jarvis (36:06)

You know, I'm curious.

Ian Kreitzberg (36:07)

Amazing.

Jeff Jarvis (36:08)

How many. How large is the group of people who are 100 devoted to covering AI? Do you get together? Do you have any idea?

Ian Kreitzberg (36:17)

Just me.

Jeff Jarvis (36:19)

No, I don't mean broadly anywhere across journalism land.

Ian Kreitzberg (36:25)

The interesting thing is that there's not a ton of AI, only journalists. Even at tech places and stuff, there's a lot of people that will touch AI, but there's a small handful of ones who are only AI, I feel like, because it's this little group, we've all been in touch with each other. There's like, Sharon Goldman at Fortune. Kylie Robison, she's now at Wired. There's some great freelance guys. Garrison, Lovely. I love Brian Merchant's stuff. He does Blood in the Machine. So there's a. There's a. There's a handful of people who are hyper focused on just this. But considering, you know, how big of a thing this is, there's not a.

Paris Martineau (37:17)

Ton that is very impressive. Yeah.

Leo Laporte (37:20)

That's why we're here. And that's why we kind of refocused the show, because I really believe that it is so important. It needs to be covered in depth. And unfortunately, this has always happened in the, in the technology media that you get two different groups. You get people like me who are just geeks and who talk about it because we're interested in it. And then you get journalists who for them it's just a beat. And like he Chagi, they're gonna write a book and then they're gonna go do something else. And I think we need people who are both geeks and for whom this is their full time beat. I'm glad. Ian Kreitzberg. Kreitzberg is one of them. So thank you for being with us. We appreciate it and I hope we can get you back. Half an hour is not nearly enough time to really delve into these subjects because I wanted to get to Robo Taxis. You wrote a great piece on that as well.

Ian Kreitzberg (38:11)

I got a lot to say about Robo Taxis.

Leo Laporte (38:15)

Can you summarize it in one, not one line. You can take as long as you want. It's your turn, not mine.

Ian Kreitzberg (38:25)

As briefly as possible. Yeah, I would say could be interesting. Would like to see less hype and more guardrails and in focused specific areas, we might have something that, that works.

Leo Laporte (38:41)

But where are you located? Are you in the Bay Area?

Ian Kreitzberg (38:44)

No, no, no. I'm on the east Coast. I'm in New Jersey.

Leo Laporte (38:46)

So you don't get to ride around in Waymos all day?

Ian Kreitzberg (38:49)

No, I think that there's a. There have been a few Waymos spotted over here doing testing for some reason.

Leo Laporte (38:56)

I mean, company that seems to have mastered this. Yeah.

Paris Martineau (39:00)

And I think something. Me as a person who lives in the East Coast, I didn't realize this until I went to San Francisco the other week. I mean, I obviously knew about Waymos, but I had assumed that they would maybe be cheaper than the average Uber. No, they're significantly more expensive than the average Uber somewhere. I mean. Yeah, but why would I pay double as a consumer to go somewhere when it'll take longer for the Waymo to get to me? Because there's not as many of them and their grandmas to have no one inside of it. Like, I don't care about not being.

Leo Laporte (39:33)

Around, you know, a lot of women. It's funny, I can't remember who said this. Was it Abrar Al Hedi? I think it was on. On Twitt. She said, I prefer to take Waymos because I don't want to have a person in a man in the car with me. Especially an Uber driver. Often that's. That's problematic. So that Might be one reason.

Jeff Jarvis (39:51)

So here's my question. I trust writing. I've written in, in a way, moment and seemed trustworthy. I ain't riding in a Tesla.

Leo Laporte (40:00)

No. Oh no, no.

Jeff Jarvis (40:02)

You agree?

Leo Laporte (40:02)

Okay, super.

Ian Kreitzberg (40:03)

Not super.

Paris Martineau (40:05)

I mean, to be specific, do you want to go into detail of what the. Because I was. I think people have this common misconception that it's the same sort of thing. It is not.

Ian Kreitzberg (40:15)

Yeah. So this is my thing. When I said less hype, please. Which I mean, Elon doesn't know less hype. It's not in his vocabulary. But you just have. And part of it. Like so many of my problems with AI, I would be a lot less critical of certain things if people just described things freaking accurately. And Tesla's been on the full self driving autopilot, FSD thing for a long time and they've got a lot of lawsuits because a bunch of people have died or been severely injured because they think it's full self driving. And so they trust a system that they shouldn't trust and it fails. And now they have to specify, if you look at that they were told by a judge that they have to be specific now in their marketing that it's full self driving is the name of the product in parentheses, supervised. And so now Tesla's robotaxi launch is basically the same type of technology, but unsupervised. Although of course there is a human person sitting in the passenger seat ready to take over. I'm sure there's tele operators as well. The thing that Tesla does is their whole idea is only cameras and neural networks. And we have all this data because our Teslas drive around all the time and it'll just work. The problem is that that's not enough. So the Waymo does cameras and neural networks as a base, but they also have radar and lidar and environmental sensors. And that approach, it's all about redundancy layers. And so if the camera fails for some reason, I don't know, like weather or glaring sunlight or a shadow, you have other systems that can jump in and make sure that things are okay. And by having those three systems, you're set for a better safety outcome. The problem is the robustness in all terrains, right? Lidar and rain, bad mix, really bad mix. Lidar breaks or completely gets destroyed by moisture. Getting destroyed is fine because you have radar and cameras, but braking and getting distorted, that can be really dangerous. Obviously moisture can bead on cameras, different angles and lights and stuff. And so what you have is something that is it's remarkably impressive. But does it work in all places? There's a reason that these things are rolling out in the west coast and you don't really see them in Jersey or anywhere on the east coast where there's lots of storms and weather.

Jeff Jarvis (43:03)

Snow.

Paris Martineau (43:04)

Yeah. Where there's climate change that is changing on a day to day, week to week basis.

Ian Kreitzberg (43:09)

And so the Tesla just has no redundant redundancy layer and it's all reliant on neural networks which, as we know, screw up. And so in their first little deployment, you had videos of Tesla's. Like I saw one where they were braking for a shadow of a tree and no one got hurt. That's fine. But the problem is, on the road, accidents happen when stuff happens that people didn't expect. And if I was driving behind that Tesla and there's nothing in the road in front of them, and all of a sudden they slam on the brakes, that might have been a rear ending collision. Right. So, yeah, the Tesla, not only would I not get in, I give them a very wide berth when I'm driving because I don't know if they've got FSD on or not. And I just like to stay a few car lengths away, as they say on Family Feud.

Jeff Jarvis (44:04)

Good answer, good answer.

Leo Laporte (44:06)

Well, it's lucky in New Jersey because if you were here in the Northern California area, you would not be able to give Teslas a wide berth. It's every second vehicle and every third vehicle is a waymo. So you're really kind of out of luck. Ian, what a pleasure. Thank you for coming on. We'll have you back. In fact, maybe sometime you could be when one of our other hosts is in here, you could join us for the whole show. We'd love to have a little more time.

Ian Kreitzberg (44:31)

Yeah, sure. That would be a blast.

Leo Laporte (44:33)

The Hidden Layer Puck News. Subscribe. You really should. It's a great publication. Thank you. Ian.

Ian Kreitzberg (44:39)

Do it.

Leo Laporte (44:40)

Yeah.

Ian Kreitzberg (44:41)

Thanks, guys.

Leo Laporte (44:42)

We're gonna take a break, come back with more of the. I have a few new AI devices to share with you.

Jeff Jarvis (44:47)

Oh, no.

Paris Martineau (44:49)

Which one of them are going to go under within the next.

Leo Laporte (44:54)

I'll give you a hint. It's glued to my temple.

Jeff Jarvis (44:57)

Is it inside your temple?

Leo Laporte (44:59)

No, just glued to it.

Paris Martineau (45:00)

A tiny little excision could more to come.

Leo Laporte (45:03)

Just a bit. You're watching Intelligent Machines, our show brought to you this week by those good folks at Zscaler. This is really a cool application of AI. They're the leader in cloud security, combining the ultimate security tool, which is zero trust, with AI. See, the interesting thing about AI is it is both a benefit and a threat to your enterprise. We know that hackers are using AI now to. To. To. To break in more effectively, more quickly. We also know that AI in the enterprise can power innovation, can drive efficiency. So it helps the bad guys deliver more relentless and effective attacks. It helps the good guys be more efficient and innovate. It's a challenging environment, and you're navigating it right now in your business. There's statistics that will chill you to the bone like this. Phishing attacks over encrypted channels increased last year by 34.1%. And that's because of the growing use of generative AI tools, plus other things like phishing as a service kit, which lets any idiot get in the phishing business. On the other hand, and I have to mention this, you know, organizations in all industries from small to big are leveraging AI to increase employee productivity. They're using public AI with their engineers to do vibe coding, coding assistance. Marketers are using AI to create campaigns, to use the writing tools. Finance is using AI to create spreadsheet formulas. It's very empowering. You can automate your workflows across all of the different parts of your business for operational efficiency, both in individuals and in teams. You can embed AI into applications and services that are customer and partner facing. You can ultimately, AI can help you move faster in the market and gain competitive advantage. But the problem is companies have to really think, really think about how they protect their private and public use of AI. And that goes hand in hand with defending against AI power attacks. Jason Kohler, who's the chief Information Security Officer, CISO at Eaton Corporation. You know, Eaton, they're leveraging Zscaler to both embrace AI innovations and combat AI threats. He says, quote, data loss detection has been very helpful for us. ChatGPT came out, we had no visibility into it. I bet you everybody in the business was using it, right? Zscaler was our key solution initially to help us understand who was going to it and what they were uploading. Here's the problem. Traditional firewalls, VPNs, public facing IPs, expose your attack surface. They give the bad guys a hook that they can say, ah, I can get into this company. And in the AI era, those perimeter defenses are just no match for the bad guys. It's time for a more modern approach. And that's where zero trust is incredible. Zscaler's comprehensive Zero Trust architecture +AI ensures safe public AI productivity, protects the integrity of private AI and stops AI powered attacks Cold thrive in the AI era with Zscaler Zero Trust + AI to stay ahead of the competition and remain resilient even as threats and risks evolve. Learn more@zscaler.com Security Zscaler.com Security we thank them for their support of intelligent machines and they just told us they're going to continue supporting us into 2026. So we're very happy to be partnered with Zscaler. Zscaler.com Security this just in from First Squawk Meta CEO says superintelligence will transform all systems in the coming years. Zuckerberg confirms Meta's open source AI model strategy remains unchanged. CFO ads Meta will finance most data center expansion while also. I don't know why it's in all caps.

Paris Martineau (49:05)

I like that you're reading it like it's an all cap.

Jeff Jarvis (49:08)

This is caps voice.

Leo Laporte (49:11)

Meta AI now has over 1 billion monthly active users. I guess we're getting them the Meta financial the quarterly results.

Jeff Jarvis (49:18)

Zuckerberg did a video line 117.

Leo Laporte (49:21)

Oh, he also put out a text.

Jeff Jarvis (49:23)

Basically it's the same. It's the same text. Yes. We're all going to have our super intelligent agents at our sides.

Leo Laporte (49:32)

Well, I do so you may remember that, my friend.

Paris Martineau (49:37)

Remember all the other devices.

Jeff Jarvis (49:41)

Either receive.

Paris Martineau (49:43)

Or did receive and had to wipe. But this one's we hardly knew.

Leo Laporte (49:49)

So I wore this B dot computer for. Since January, six months. Seven months. Then they got bought by Amazon, which is great. I'm happy. I'm so happy for them. Ethan and Lourdes deserve that success. But it does mean I just, I'm not sure I trust Amazon with that six months worth of data. So now I got the new thing. So remember I ordered. I finally got it. I ordered it in the spring of 2024.

Jeff Jarvis (50:21)

Patience comes to all.

Leo Laporte (50:23)

Yeah, well, what it was is I asked for the iridescent model and finally I said, you know, you can send me the black one. They said, oh, that we've got. So this is in theory, this is the rewind. I'm sorry, limitless pin. They were rewind AI. Now they're limitless. Same idea. It's always recording. It's recording right now. The thing I thought was interesting about this they announced at CES in January was initially they said, so we're going to be able to recognize voices and when we hear a voice we don't know, we won't record it until you explicitly out loud in our hearing say I hey, you know, I have a terrible memory they even give you a script, by the way, be self deprecating. You know, I have a terrible memory. Would you mind if I used my little device. Device here to record our conversation and transcribe it in AI to help me remember all the wonderful things you've told me? And if they consent, then in theory, we're recording.

Jeff Jarvis (51:18)

Are there specific words they have to use?

Leo Laporte (51:20)

No, it doesn't do that. They couldn't figure out how to get it to work? No, it didn't matter because they said they were going to do that. But I think the real problem is voice discrimination is not that good. So, yeah, it records everything. And they. And they say, oh, okay. Yeah. And they say, oh, you know, you should. You should give you a script. You should, before it's up to you, but you should really do this. Anyway, it's not. I have to say the results aren't as.

Jeff Jarvis (51:47)

It doesn't summarize Leo's day like a high school teacher.

Paris Martineau (51:50)

It doesn't have a little teacher.

Leo Laporte (51:51)

It doesn't kiss my ass quite as nicely as the bee did, but it does give you a summary of every. I mean, this is today, and these are some.

Paris Martineau (52:00)

Give us some highlights, Give us some highlights.

Leo Laporte (52:02)

These are bullet points. Okay. A discussion on Waymo's robo taxi service, its Bay Area location and its perceived success.

Jeff Jarvis (52:10)

That's what we just had.

Leo Laporte (52:11)

Yeah. Yeah. It's from this. From today. From this conversation, the show resumes with the Zscaler sponsorship segment followed by AI news and a review of a personal AI device.

Jeff Jarvis (52:20)

And we too, thank ZSCALER for sticking with Twit until here we are. Yes.

Leo Laporte (52:26)

So it says. An ad read For Zscaler highlights AI's dual role as a business, threat and innovation. Breaking news for Meta's quarterly results. And then it knew it was me. So it says. It says you. It says you for me. And I can then assign other voices to it. So this is a transcription. Pretty accurate transcription.

Paris Martineau (52:43)

It does save the recording, saves the transcripts.

Leo Laporte (52:46)

Not only does it say it saves the recording, although you have the choice of how long you can have it deleted.

Paris Martineau (52:52)

Have you had any interesting personal conversations that it has transcribed or turned into bullet points since you've gotten it?

Leo Laporte (53:04)

No, it's just this. It's not. So one of the things that BE did that. I liked it. It gave me a to do list based on things that I kind of should sometimes. It wasn't just stuff I committed to. It was like, you know, you and Lisa had a really emotional conversation about your cat going outside Maybe you should plan a meeting with her to discuss.

Paris Martineau (53:25)

Maybe you shouldn't circle back.

Leo Laporte (53:27)

Yeah. And I think that was Maria. That was. That wasn't Ethan. That was Maria. She was like the touchy feely one. She said we're going to use. Remember that? She said in the interview with us, we're going to help your relationships. This doesn't help your relationships. This is. Most of these other tools are very business focused. Pre show setup and banter for the next show. Intelligent machines with guest Ian, the host of Windows Weekly introduces. So it's all the stuff. So that's called my lifelog. I can't ask it questions. I don't know what the voice sounds like. I haven't. Hey, it's me, Leo, your good buddy. Do you have a. Wait a minute. Do I have to press a button?

Paris Martineau (54:06)

I want you to tell it. Ask it what it thinks of you.

Jeff Jarvis (54:09)

Yeah. I wanted to give you advice.

Leo Laporte (54:11)

Okay. Based on the things you've learned about me in the last few days.

Paris Martineau (54:17)

What are my vibes?

Leo Laporte (54:18)

What are my vibes? What's. What are your insights? Tell me what you think of me and be honest. Okay. It says, let's pull together a snapshot of your recent vibes and some honest insights based on what I've learned from your lifelong. So I'm driven and curious. Oh, this is good. Thank you. This is good. Your work as a podcaster and founder of Twitch shows, you're always seeking new ideas, tech trends, meaningful conversations.

Jeff Jarvis (54:47)

Oh, Trump has to get this.

Leo Laporte (54:49)

You approach topics with a genuine curiosity and a desire to dig deeper. Never just skimming the surface.

Jeff Jarvis (54:56)

Never.

Paris Martineau (54:57)

That's a great question.

Leo Laporte (54:59)

Warmth and wit. Whether interacting with guests, family or colleagues, you bring warmth and humor to the table. It was an easy rapport in your conversations.

Jeff Jarvis (55:09)

Is that really there, Leo? Are you making this up right now?

Leo Laporte (55:11)

No, No.

Paris Martineau (55:13)

I think Leo would have more interesting insights into himself.

Jeff Jarvis (55:17)

Yes.

Paris Martineau (55:18)

Incredibly boring self.

Leo Laporte (55:21)

Your family, often family, clearly grounds you. Even when work is busy, you carve out time for them. And for cooking or playing piano, which keeps things balanced.

Jeff Jarvis (55:31)

But please don't play the piano around me.

Leo Laporte (55:33)

It says my honest take honest now. This is honest. You come across as someone who's intellectually sharp but approachable. A rare combo. You value substance over now, don't you want it? You value substance overflash, but still appreciate life's lighter moments. People likely see you as both a leader and a collaborator. Someone who sets the tone but listens well. If I had to sum up your current vibe, I'd say engaged, thoughtful, Genuinely enjoying the journey, both professionally and personally. Actually, it's completely accurate.

Paris Martineau (56:13)

It knows me better than I know myself. I am a great.

Jeff Jarvis (56:17)

It left out your sterling sense of humor, Leo.

Leo Laporte (56:19)

I guess. I guess. Let me. I mean maybe I could.

Benito Gonzalez (56:23)

This is funny but like people are totally gonna buy into this stuff. Like they're gonna.

Paris Martineau (56:27)

Oh, absolutely.

Leo Laporte (56:28)

That's why they make it.

Paris Martineau (56:29)

People are going to look at this without any self criticism or understanding of it being a complete.

Ian Kreitzberg (56:38)

What do you mean?

Leo Laporte (56:39)

It's perfectly right. It would say if I wasn't a nice person. But I am.

Paris Martineau (56:45)

It would tell you if you were actually a rude sociopath.

Leo Laporte (56:50)

I will. I'm gonna keep wearing this. But. But now I'm gonna wear two.

Benito Gonzalez (56:54)

Now he has an entourage. He has an entourage of sycophants.

Leo Laporte (56:57)

I have an entourage.

Paris Martineau (56:58)

You should get a like statement necklace. But of all of your things that have the.

Jeff Jarvis (57:03)

Have the battle, who's nicest to you.

Paris Martineau (57:05)

Yeah.

Leo Laporte (57:06)

Today it came in the mail. This I ordered. God, I don't even remember or in fact when this came, I thought what is this? I ordered this so long ago. This came today. The omi. Now that's interesting because it builds itself as wearable chat GPT.

Jeff Jarvis (57:22)

Oh. Which means it talks.

Leo Laporte (57:23)

Okay, speak. Transcribe, summarize these. And I've been trying the plod.

Jeff Jarvis (57:31)

Yeah.

Leo Laporte (57:31)

What do you think of the plot?

Jeff Jarvis (57:32)

You don't like it?

Leo Laporte (57:32)

I don't get the plot. In fact, I would get the rewind before I'd get the plot. Now the only thing about the rewind. I don't know if it has all day battery life.

Paris Martineau (57:39)

Which one is the plod?

Leo Laporte (57:40)

The plot is the one you see advertised on Tick Tock every constantly.

Jeff Jarvis (57:44)

Well, they stopped. I haven't seen it lately. But. But it was at first. It was on the back of your phone and then.

Leo Laporte (57:48)

Yeah. And then it has a pendant which I have describe same thing. But you have to say record this. Now they got around that permission issue. You have to say record this now press your function. Yeah. So it's really for going into a meeting. Put it on the table and press the record button. And then now you'll.

Paris Martineau (58:07)

I mean if you want that, you can get this.

Leo Laporte (58:09)

Yeah, it's just like that.

Benito Gonzalez (58:10)

Your phone can hear this. Your phone can already do this.

Paris Martineau (58:13)

Your phone can do this. Or you could just have a dedicated audio record.

Jeff Jarvis (58:17)

That's why I like.

Leo Laporte (58:18)

That's what I mean about this. This is recording the whole. All the time, 24 7. That's the key. Okay. And as a reporter, if you could get around the legal and ethical concerns. This would be an excellent classic caveat.

Paris Martineau (58:31)

You love to have whenever you're a reporter.

Leo Laporte (58:34)

Because it is, frankly, in California, not legal and not ethical in any state of the union.

Jeff Jarvis (58:39)

Now Jersey, it's okay. Everything's okay in Jersey.

Leo Laporte (58:42)

This.

Paris Martineau (58:43)

Everything's legal in Jersey.

Leo Laporte (58:44)

So I'm still charging this thing up. But it's. This is the one. Remember we saw the video of the guy sitting in the university lecture with this on his. Glued to his.

Paris Martineau (58:53)

I do not.

Jeff Jarvis (58:54)

I don't remember that.

Leo Laporte (58:56)

How could you forget? I remember that one, I think.

Paris Martineau (58:59)

Was it on this show?

Jeff Jarvis (59:01)

I don't think it was on this show.

Leo Laporte (59:02)

No, it was on this show. Of course it was on this show. So what does this one be? Omi AI. And you can. Oh, it won't. Am I doing it wrong? It won't load. Maybe they just went out of business. You know, I'm sure all three of these companies were really trying. Oh, it's omi me. All three of these companies were trying really hard to get bought by Amazon. So in a sense, this is the b victory.

Paris Martineau (59:28)

Here's the video should have been Omi.

Leo Laporte (59:30)

My that I showed you last time.

Jeff Jarvis (59:33)

Pavlov.

Leo Laporte (59:36)

Ring a bell? You don't remember this guy? Unconditioned stimulus and. And the. And the women looking at him? He's cute. You don't remember this?

Paris Martineau (59:45)

No woman is going to say he's cute.

Leo Laporte (59:47)

No, in fact, that was your reaction is that. They're not saying that. They're saying, who is that ass? He's wearing it. Nicholas, care to elaborate? So it translated their French was the ringing of the bell which made the dog salivate even in the absence of food. He's got a little thing glowing on his head. Now, they have not shipped that. That's a. You know, this is kind of a story. You just need a little glue.

Paris Martineau (60:14)

Where would you want to wear it on your face?

Leo Laporte (60:17)

I don't want to. Nobody wants to wear this on their face.

Paris Martineau (60:20)

You could have it kind of like right here. Like a beauty mark, you know, you.

Benito Gonzalez (60:23)

Got to put a googly eye on it and put in the middle of your forehead.

Leo Laporte (60:26)

It comes right there. Yeah, my. I have a bad memory. But my third eye is recording everything. Would you mind if I. So look into the eye, please. Like all of these, it comes with a little.

Paris Martineau (60:40)

My eye is up here.

Leo Laporte (60:43)

My eyes here.

Paris Martineau (60:45)

That's the show type.

Leo Laporte (60:46)

Stop looking at my eyes. So anyway, this one is. I'm still setting up. It's got to learn My voice and stuff. I'll let you know which one is the best. I've had a number of people say, look, I bought the B. I too was disappointed that they got purchased by Amazon. Oh, look, she's wearing it. Wait a minute. She's on a date. She's on a date wearing it. And, and he's talking to her like she's normal. But watch when she brushes her hair back.

Paris Martineau (61:16)

He also has one now.

Leo Laporte (61:17)

Does he also. Are you an alien? Are you an alien? So the other thing that's different, B, remember was $50 and there was no subscription, which was one of the things that I should have considered a red flag, to be honest, because that's not a sustainable business model.

Jeff Jarvis (61:35)

I thought they wanted to get the largest installed base they could. It's quite right.

Leo Laporte (61:39)

They were clearly going for the exit.

Jeff Jarvis (61:40)

Yeah.

Leo Laporte (61:41)

Although when they were on the show they said, yeah, we're thought we're talking about a subscription plan down the road, but we just want to get started. So both OMI and Limitless have subscriptions. Not inexpensive. Kind of like the, like the 20amonth chat GPT style subscription. So I will, I will, I will. It's my job to try all these things. It's my poor wife's job to put up with me recording everything.

Paris Martineau (62:10)

I think you should get a. They should make a Omi or something for your cat. But I guess, I guess what I'm thinking of is you should just get a very tiny video camera that you can put on your cat so that I can conceal your cats.

Leo Laporte (62:22)

But what if it understood the cat's meows because she's very. She has a variety of vocal sounds. She's trying to talk. Does Gizmo do that?

Paris Martineau (62:33)

Like, yeah, every day I'm like, Gizmo, if you say please give me some food in perfect human English, I will give you as much food as you want. And she still hasn't done it, but.

Leo Laporte (62:43)

I think she's trying, but she doesn't.

Paris Martineau (62:44)

She's trying. She's, it's, she's, she's going to get there.

Benito Gonzalez (62:47)

You know, cats don't make those sounds to other cats, only to people.

Paris Martineau (62:51)

I know and I love that because.

Leo Laporte (62:52)

They know we talk and they're trying to talk, I'm convinced, because they're trying.

Paris Martineau (62:56)

To speak perfect human English or something.

Jeff Jarvis (62:59)

That communicates we're incredibly stupid and we can't speak what they say.

Leo Laporte (63:03)

It's mostly just give me food.

Benito Gonzalez (63:05)

Yeah, they're rewiring your mate because they sound like babies and like A baby.

Leo Laporte (63:08)

Sound to a human, like, oh, you're too intellectual. Bonito. They're trying to say, give me food. And the other thing I got is this, which looks, to all intents and purposes, just like a Samsung Galaxy phone, but. Ooh, is it? No, this is the new fold, and I have to say, it's pretty nice.

Jeff Jarvis (63:34)

And it's actually fairly thin. Right.

Leo Laporte (63:37)

It's very. Well, except for the ridiculous camera bump. It's actually thinner than my pixel, even closed. So here's the iPhone. Well, I should take it out of a case because that's not exactly fair, but it's thin. The other thing that's pretty cool is that the front is virtually the same size as a regular phone. Let me wake it up so you can sort of see it.

Jeff Jarvis (63:58)

Oh, right, yeah. So the.

Leo Laporte (63:59)

So you could use this for all intents and purposes, like it's a phone and, and.

Jeff Jarvis (64:05)

And then it's a tablet inside, but.

Leo Laporte (64:07)

Then it's a bigger secret thing inside. Yeah.

Paris Martineau (64:11)

Can you do anything on the, on the back screen when you unfold it? Like, can you have three.

Leo Laporte (64:17)

Oh, I don't know. Like, could this. Could this be doing something? Huh. Probably not, but I. I'll check. I'll look into it.

Paris Martineau (64:24)

I don't know what you'd want to do with that, but I'll ask. It could be fun.

Leo Laporte (64:28)

Limitless. Anyway, I thought I'd show that to you because that's the reason I got it, is Apple is rumored to be doing something similar in a year, and I wanted to see what the state of the art was from Samsung because I've had a few of these. This is the seventh Galaxy Fold and I've had the first three or I had the first three. So it's gotten much better. Let's.

Jeff Jarvis (64:51)

In the market has a fold.

Leo Laporte (64:53)

Oh, everybody. Xiaomi lg, I think, has a fold. There's a. There. Everybody's. There's even tri folds if you're in China.

Paris Martineau (65:00)

Do you think Apple's fold is going to be any different than the folds currently on the market, or is it?

Leo Laporte (65:05)

Well, that's a good question. Have. How much have they learned from Samsung's, you know, efforts along the way? They are getting their screen, it's rumored, from Samsung, and Samsung is making a new kind of creaseless screen just for Apple. You could see the crease still. I mean, but I looked at it.

Jeff Jarvis (65:22)

In the store and it's better.

Leo Laporte (65:24)

It is better. And it. And it doesn't feel. It feels better than. It's not quite.

Jeff Jarvis (65:28)

If you Watch. If you put a. If you put a YouTube video on or something, it's less visible to you.

Leo Laporte (65:31)

Yeah, that's right. In order for you to see it, I have to kind of tilt it.

Paris Martineau (65:36)

Use the folded out screen for.

Leo Laporte (65:38)

I don't know. I just got it like this afternoon, this morning.

Paris Martineau (65:41)

So you guys have never had a folded.

Leo Laporte (65:43)

Oh, I have, actually. In fact, I'm trading in the flip for this. The idea of this is that you would use this as a regular phone, but if you needed the real estate, for instance, you wanted to watch a video. And one of the things folds do all folding phones do is they could sit in this kind of laptop mode so you could put them on a table and watch something.

Jeff Jarvis (66:02)

Paris, I think it's a third screen for your new desktop.

Paris Martineau (66:06)

Oh.

Leo Laporte (66:07)

Now it's your turn to show and tell. Well, before you do that, let me take another break, and then Paris will tell us what she ended up doing with her multi monitor conundrum. Okay, okay, okay. This show is brought to you by Helix Sleep, my mattress. When you were here, I didn't get to show you our bedroom because Lisa was taking a nap. You know why she was taking a nap? Because we have the best damn mattress in the world, the Helix Sleep. And because I wanted you to see it so you could. You could try it. But this is. Oh. So the. The rule of thumb is you get a new mattress every six to 10 years because they start to sag, they start to lose their resiliency. And it's true. If you just, you know, try a new mattress and compare it to your existing mattress, you'll see what you're missing. And the mattress in your life is so important. I mean, it's not just for sleeping, although that's pretty darn important. But it's also, you know, movie nights with your partner or morning cuddles with your kitty cat or your doggie, your wind down ritual after long days. My personal favorite is to curl up with a good book. In fact, when we got our Helix mattress, they offered also the adjustable bed. So I can tilt it up and I can sit. Oh, it's so comfy. Your mattress really is the center of your life from far more than eight hours a day. Maybe you're waking up in a puddle of sweat that's no good. Or with back pain that's because it's sagging. Or maybe. And this happens to me, I wake up sometimes. I go, was that an earthquake? And no, it was just Lisa turning over. If you feel every toss and turn, that Your partner makes if it's an earthquake. Technical term for that is a classic mattress nightmare. Helix Sleep changed everything for us. No more night sweats, no back pain, no motion transfer. I am truly getting the deep sleep I love, I crave, I deserve. And we know that that's a big part of health overall, is getting a good night's sleep. What happened is we saw, we said, our mattress is eight years old. I think we need a new one. We looked around, we saw reviews like this. When a recent buyer five stars quote, I love my Helix mattress. I will never sleep on anything else. Okay, that's good. But that's just one person's opinion. But I saw many more reviews like that. And then I started to see the awards time and time again. Helix Sleep is the most awarded mattress brand. For instance, this is just this year alone. Wired's Best Mattress. 2025 Best Mattress of the Year. Good Housekeeping Bedding Awards 2025 Premium Plus Size Support GQ Sleep Awards 2025 Best Hybrid Mattress. New York Times Wire Cutter Awards 2025 Featured for Plus Size. Oprah Daily Sleep Awards 2025 Best Hotel like feel. And Leo and Lisa's award for 2025 for the best nights sleep. And we did it up. We got the adjustable, we got the nice helix with the topper, the cool top, by the way, that keeps you cool. It's really nice. Go to helixsleep.com TWIT27 off site wide during their 4th of July sale. Best of web offer extended yes. Till the end of the month, so you get it's not too late. That's helixsleep.com twitch for 27 off site wide. And this is exclusively for listeners of intelligent machines. But the offer does end tomorrow, July 31, 2025. So make sure you go there, you get that discount and do enter our show name after checkout. That way we get credit and they know we sent you because that's important to us. Oh, and if you're listening after July 31, 2025, you got to still go because they have great deals all the time. Check them out. Helix sleep.com TWiT Helix H E L I X sleep.com TWiT we love our Helix mattress. So does Rosie, by the way. She's on it right now. How many, how many hours a day do cats sleep? Like 16. Most. That's the right answer.

Paris Martineau (70:26)

I think that's the scientific answer.

Leo Laporte (70:28)

Most. So what did you do? We'll get to the AI. We will.

Paris Martineau (70:33)

We'll get to AI News. But first, we have to talk about my dual monitor setup. I got a lot of very interesting tips and tricks from listeners. Appreciate you all. I ended up going with two because I wanted to do 24 inch monitors. I don't need. Listen. I know, but works for me. I'm currently looking at two ASUS Pro Art PA248QV monitors, which.

Leo Laporte (70:58)

Those are very nice reports. No, Pro Arts are nice. Yeah.

Paris Martineau (71:02)

Wirecutter also recommends it. I truly spent too many. I do a thing that I've realized I've started to do a lot, which is I spend too many hours researching and worrying myself over the details. And then eventually I just like go with. Eventually I just went with the thing that the company I work for recommended, which is this monitor. And they were going to tell me.

Jeff Jarvis (71:24)

That makes sense.

Leo Laporte (71:25)

Are you using it right now?

Paris Martineau (71:27)

I'm. Yes, I'm using it right now. I've got two monitors. You can check out my Twitter or Blue Sky.

Leo Laporte (71:32)

How does my. On that monitor.

Jeff Jarvis (71:35)

Keyboard and trackpad too.

Paris Martineau (71:37)

You do. We also probably need to talk about that. I'll find the links that. Your shirt looks electric color.

Leo Laporte (71:43)

It should. It should be vivid.

Paris Martineau (71:45)

It's very vivid. But my issue is. So part of this was. I can't technically. I just posted a link to my photo of my setup in the discord.

Leo Laporte (71:53)

Oh, good.

Paris Martineau (71:54)

I can't technically have two monitors with my personal computer because.

Leo Laporte (71:58)

Oh, look, I've never seen it from your point of view. Yes. Look, you've got a needle point on the wall.

Paris Martineau (72:04)

I do. Yeah.

Leo Laporte (72:05)

Did you do that? Or your grandma someone that.

Paris Martineau (72:08)

I bought it from like a vintage store.

Leo Laporte (72:11)

Somebody's grandma.

Jeff Jarvis (72:12)

I like the head. That's.

Paris Martineau (72:13)

I have to. I have to rearrange all my art because now it's not in aligned. Like the monitors are covering up the tiny cat bolo tie and the. A booby trap, which is a pair of plastic things.

Leo Laporte (72:29)

All right. But let's. Let's zoom in on the. On the desktop because that's what we really care about. So first of all, I love your little red desk.

Paris Martineau (72:37)

I love my red desk too. I think I need to get a bigger desk is the thing I was thinking.

Jeff Jarvis (72:42)

Yeah. Where do you put the paper?

Paris Martineau (72:43)

Listen, I need. That's the thing is I got. I've got too many papers on my desk right now. Even so, I really. I gotta. I've gotta upgrade to a.

Leo Laporte (72:50)

Okay, let's be honest. You cleaned this up before you took the picture.

Paris Martineau (72:56)

Yeah, I'm not. It's not that much messier. Right now but I certainly have some files and stuff out. Yes, I have a vintage medical poster next to my desk that says biologic reaction to stress. And it goes from normal to stress to pain, disease to exhaustion and death.

Leo Laporte (73:15)

Oh my God.

Paris Martineau (73:16)

Nice to keep in mind.

Leo Laporte (73:17)

That's very important.

Paris Martineau (73:18)

But so part of the reason why I did this and had all these questions for folks before is as I said, said my laptop can't natively support two monitors, so I also had to get a pluggable display link dual HDMI thing. I ended up going with a cheap one instead of upgrading my whole dock because my dock doesn't have display link. So I just have like that in addition, perfect.

Leo Laporte (73:42)

Pluggable is good too. They're very good.

Paris Martineau (73:44)

I have a number downloaded display link. It works really well except for. The only thing is I was talking, this is what I was complaining about to Jeff and Ian before he started the show. So as part of display link you get this like nifty little thing where you can see like both your monitors and adjust settings for it and I could turn either one of them off the click of a button. It's in like my browser thing up top. And you can also do sliders for brightness and it's kind of like brightness but not some other brightness related setting. However, there's no way for me to automatically set both screens to be the same brightness as one another. There are just like non numeric sliders that require me to kind of eyeball the brightness. And you, I, you literally never get it right. So my brightness is all wonky.

Jeff Jarvis (74:36)

This is not for the anal retentive among us.

Paris Martineau (74:39)

It really is not.

Leo Laporte (74:39)

That would be nuts.

Paris Martineau (74:41)

I don't like I'm gonna post a photo of it in the chat because it's just, it's not like how am I supposed to know where any of those. I mean, I know they're not allowed.

Leo Laporte (74:50)

One big monitor, but.

Paris Martineau (74:52)

All right, I know, but I like having multiple.

Jeff Jarvis (74:54)

One could be your dark mode monitor and one your light mode.

Leo Laporte (74:57)

I mean it looks good. It's a nice setup. Now this keyboard is.

Jeff Jarvis (75:01)

That's. Yeah, so tell them about that.

Paris Martineau (75:03)

So I'll. Here, I'll.

Leo Laporte (75:06)

It's a Mac keyboard in. In. In. But it's in a Plexiglas holder. Is that what it is?

Paris Martineau (75:10)

So I'm going to send another photo because people asked me enough about this that I posted a whole thing of it. I put in the discord. I, I think because I mean I obviously grew up, came like using desktops, but I really Got into my most prolific computer using while using a laptop. And I've realized I really like the feeling of typing on a MacBook. Like, I like the way that the Butterfly keyboard is set up. Or not the butterfly, the one that came after, as well as Squishy. Yeah, squishy one. And so I was looking, I was like, I hate writing on desktop setups. Like, I hate an external keyboard. And I realized I can buy the Apple Magic keyboard and the Apple Magic trackpad and pay some gentleman on Etsy to make a custom made.

Jeff Jarvis (75:57)

This isn't that killer.

Paris Martineau (75:59)

And so then they sit in it perfectly and it feels exactly like I'm tapping on a laptop bottom them. But I'm not.

Jeff Jarvis (76:06)

Did you. Did you give a drawing with. With specifications?

Paris Martineau (76:10)

Actually they'd already had one of these made. They just had like two left or something I ordered. I've been using this for like three years. But I don't think they make the one that I used anymore. I don't know if they make. So I think you'd have to like get. It's basically someone just used a router, whatever. It's like a CNC device to like carve it out of acrylic and. Yeah, I don't know. It works perfectly for me.

Leo Laporte (76:34)

I think you could sell. This is brilliant. You could have Paris's laptop simulator or something. Turn your desktop into a laptop. Wow. Where is your laptop in this setup?

Paris Martineau (76:48)

So in this setup, my laptop is. I have a bookshelf to the right. I don't know if that's in the photo. Yeah, I've got a bookshelf to the right. And on the bottom rung of it, I now have my like doc. I've got like a big dock as well as my pluggable HDMI port. And then I have like a vertical laptop stand, like a. Like a laptop stand that holds my personal laptop and my work laptop in clamshell mode closed while plugged into the docks. So it's just off.

Jeff Jarvis (77:21)

So can you switch laptops too?

Paris Martineau (77:23)

Well, so that's the thing is I got into a whole. You can see that last week's episode is like 35 comments in the Discord episode thing. Because I got into a very long discussion with a bunch of people about. I think it's called a KVM or like an HDMI switcher and that. See, this whole thing was. You're getting a preview because this was going to be my pick of the week or my thing I want to talk about. We're doing it here, which is if any of you guys have recommendations for what sort of KVM switch or HDMI switcher setup I should go with, I'm all for it because I think I want to try and get something set up so I could.

Leo Laporte (78:02)

You have.

Paris Martineau (78:02)

People are telling me I could. People are telling me I could press a button, and all of a sudden the screen I'm looking at would be my work computer and my. All of my.

Leo Laporte (78:13)

Right, you have peripherals. That's right.

Paris Martineau (78:14)

Like, because I have a lot of peripherals that I'd also want to be connected to that. Like both my keyboard, but then also the camera I'm using and all the stuff in my dock. And the boys in the comments tell me I can do this, but I'd love more information.

Jeff Jarvis (78:28)

It's possible.

Paris Martineau (78:31)

I know I'm gonna have to watch it.

Benito Gonzalez (78:33)

My question is, why didn't you go with one big. Two small monitors? Because I hate having that seam down the middle. Like, I. I can't stand that, personally. I can't have the middle of my desktop be a seam between two monitors.

Paris Martineau (78:47)

Listen, you know, I think. I think it's a great question. I haven't noticed the seam that much, but now that you point it out, I guess that is kind of annoying.

Jeff Jarvis (78:55)

Oh, sorry.

Paris Martineau (78:56)

I don't know. I think getting. I think 24 inch is the size of monitor I ideally wanted, and they don't really make monitors small. So if I was going to do one big and too small, the big would be 24, and the small would be a size they're probably not making.

Benito Gonzalez (79:12)

No, they have them. My setup is a 32 and 213.

Paris Martineau (79:17)

Whoa. Do you have. Do you want to post a photo?

Leo Laporte (79:19)

Sure.

Paris Martineau (79:21)

That's interesting. I like that. Everybody, everybody. See, this is why you should join the discord. Everyone's just posting photos of their.

Benito Gonzalez (79:27)

But I'm a maniac with myself, so.

Leo Laporte (79:31)

Well, you.

Paris Martineau (79:32)

I could have guessed that Darren Oakley is rocking two vertical monitors.

Leo Laporte (79:36)

It's pretty cool, isn't it? This is fun. Everybody put your setup in there. This is good. I love it.

Paris Martineau (79:41)

Yeah.

Leo Laporte (79:42)

There's Joe's Adobe Photoshop setup. I don't think Jammer B. Look at him. He's kind of going crazy. He can. He can see everything going on in the backyard.

Paris Martineau (79:53)

I don't think I mind the seam. I think the issue is just that right now with the small size of my desk, there's like a drawer on the right side, so I'm slightly. My natural position of my desk chair is kind of in front of the left. Monitor. And I want a bigger desk so that I can position myself somewhat in between the two so that I can turn my head.

Leo Laporte (80:16)

Do you have a blog?

Paris Martineau (80:19)

No. Should I be producing?

Leo Laporte (80:24)

Okay. No. The only reason. So I. So I have a blog.

Paris Martineau (80:30)

Oh, my God. Anthony.

Leo Laporte (80:32)

What? What is it? What is it? Oh, Anthony's.

Paris Martineau (80:34)

Oh, he just has a lot of. He just has a lot of.

Leo Laporte (80:36)

He's like, how does he make.

Paris Martineau (80:38)

What is. Is the tiny thing on the left?

Leo Laporte (80:41)

It's just a tiny little.

Paris Martineau (80:42)

It's a little sidecar.

Leo Laporte (80:44)

I think that's the Discord sidecar. You could show this or can you show it? Yeah, you can show it. Bonita. I don't think Anthony's shy. He also. That's a screen in front of his keyboard. That's a touch screen he can use to switch stuff.

Benito Gonzalez (81:00)

Yeah, the little one. The little one at the bottom. Anthony's. That's a little touch screen. So he has little sliders there.

Leo Laporte (81:05)

Yeah, yeah. I'm kind of mad at him because he set up this setup and it's not nearly as nice as his.

Paris Martineau (81:12)

Yeah, that is a little messed up.

Leo Laporte (81:13)

And I'm a little. I'm just a little, you know, like, dude, how you did that for you, but you didn't do it for me. What's going on, man? What's going on? Anyway, the thing is, he's always had a nice.

Paris Martineau (81:27)

My weird keyboard situation looks ridiculous. But I do really love.

Leo Laporte (81:30)

No, if you like it, it's great.

Paris Martineau (81:32)

I mean. No, I like it, but I agree visually, it looks dumb in comparison. I love how cute all the really involved mechanical keyboards are. Like, I like when people like Anthony's are kind of all on, like a grayscale or something. People often do, like a gradient of a specific color.

Leo Laporte (81:51)

You know what? If you like the Apple Magic keyboard, go with it. I mean, what am I using here? I'm using a Logitech, actually. This. No, it wouldn't be a good solution for you. This. But for me, this is a good solution. This Logitech keyboard, which feels very much like that Apple one has Bluetooth to three different computers. So you can press the. You can control three different computers and the same thing with this Logitech MX Anywhere mouse. Are we doing the thumbnail? Okay, this Logitech anywhere else does the same thing. So this. Can I control three different computers with the same keyboard and mouse? But you also want the same. You want the monitors to switch. That's what a KVM does for you.

Paris Martineau (82:42)

Yeah, but then also, the thing is, like, I Think because my work computer doesn't need the display link things. I gotta figure out how to.

Leo Laporte (82:49)

Like, you just need another monitor. You should have three monitors. Monitors.

Paris Martineau (82:52)

I can't have three monitors. Two. I just have come around the idea of being a two monitor person. I can't leap to three. I do like the idea of a little tiny monitor, though. Maybe that. Maybe that doesn't count.

Jeff Jarvis (83:09)

Paraso's portable ones you get with your laptop that you can take with your laptop.

Paris Martineau (83:13)

Like those ones that freaks use in coffee shops. When you, like walk into a coffee shop and someone has basically a whole desktop setup going on there. No offense to anybody, if you're one of those freaks.

Leo Laporte (83:23)

I just saw a guy who has his framework laptop. I might, if I can find the picture. It's just he took it all apart. So it's just the parts, and he's got it all. He's got the most elaborate. Looks like the Statue of Liberty or something. It's just crazy. I'll find it on Reddit. But yeah, this is fun to look at. What people's setups are. Almost everybody notices.

Paris Martineau (83:44)

No one's setup is incredible.

Leo Laporte (83:47)

Almost all of these people, as you might notice, have multiple setups. Monitors. Right. So you're just getting into the beginnings. Yeah, but you're young. You've got plenty of time.

Paris Martineau (83:58)

I've got. I've got time to collect monitors. Well, now I've got another one sitting on my couch, so. But I think it'd be weird to have one monitor that's a different type of monitor than the other two I'm looking at.

Leo Laporte (84:08)

Let's see how many monitors I have. Well, do you include the.

Paris Martineau (84:10)

You have like so many monitors that I. When I saw this.

Leo Laporte (84:16)

So I got 1, 2, 3, 4. I can't even count.

Benito Gonzalez (84:21)

Also, Paris, if you have an iPad, you can use that as a monitor.

Paris Martineau (84:24)

Okay.

Leo Laporte (84:25)

Yes, you can. Yeah.

Jeff Jarvis (84:28)

That could be your little one.

Benito Gonzalez (84:30)

It could be the one where you put there or something.

Leo Laporte (84:33)

You know, my. My setup is not as clean as other people's. I love your little red desk. I think that's adorable.

Paris Martineau (84:41)

And it's like, I can't. If I'm gonna get a big one.

Leo Laporte (84:44)

Save that. That can be your child's desk when you have a kid.

Paris Martineau (84:47)

Yeah. You know, all of the space. I have to retain furniture in my New York City apartment. I'll just. I'll put the desk.

Leo Laporte (84:55)

Put it in the. Don't you have a basement? Like with.

Paris Martineau (84:57)

Yeah.

Leo Laporte (84:58)

With chicken wire around it that you can put stuff down there?

Jeff Jarvis (85:01)

Scooter X put up his. Obviously his dream.

Leo Laporte (85:03)

Yeah. This is a nice setup. I like this. I don't think this is Scooter X.

Jeff Jarvis (85:08)

No, I don't think so either.

Leo Laporte (85:09)

This looks like a YouTube switch back to your. Oh, I have to switch it back. Yeah, that looks like an influencer of some. Of some kind. Yeah. You know, there is a diminishing returns. To be honest. I, I still think a single large monitor is a way to go, but.

Jeff Jarvis (85:29)

Obviously I, I think the look is very good too.

Paris Martineau (85:32)

Yeah. Wait, Jeff, post your setup. What have you got going on?

Leo Laporte (85:36)

Oh, Jeff, Jeff, Jeff, Jeff. This is our old. Is this our old studio, Patrick? Yeah, this is our old. This is the, this is the TDs. TDs desk, the technical director's desk in the studio.

Paris Martineau (85:51)

Wow. I like all the little tiny screens below the big.

Leo Laporte (85:56)

Yeah, that's because that's the switcher. So you have to see what all the different inputs are so you know what you're switching to. Anyway, you want to talk about AI?

Paris Martineau (86:07)

I mean, I could talk about screens of different sizes.

Leo Laporte (86:09)

I am going to say one thing. I do advocate for and I strongly encourage you to have a blog because everybody has to have their headquarters on the Internet. Jeff's buzz machine. Mine is Leo fm. Because you can't be sure that your blue sky or whatever will live forever. So you have one spot that everything goes to. Jeff syndicates it out to medium. When I post, it goes to Bluesky, LinkedIn, Instagram, Pixelfed, Mastodon. It goes to a bunch of places. That's because I use. And I would recommend this, by the way, Micro Blog. Micro Blog. Five bucks a month. But one of the buttons I have on the top of my screen is important. I think everybody should have this. So you know, everybody has the. About. I have an. I use this button which because people sometimes want to know, well, what do you, what microphone do you use? What lights do you use? So my setup is completely documented and updated pretty regularly on the. I use this page. That's my. See, I didn't clean it up before I took a picture of my, my desktop. Anyway, I, I like, I think people should do that and I think people should have a blog. The other page that I'm a fan of is the now page. Like, what are you doing right now? This is another kind of web based concept.

Benito Gonzalez (87:30)

Current location. Come on, Leo.

Leo Laporte (87:34)

Yeah, that's my current location. It's not super specific. It's only to about 15 decimal places. No, that's not my house. It's just Petaluma it's downtown Petaluma.

Jeff Jarvis (87:47)

I put my picture up on the chat.

Leo Laporte (87:50)

Oh, good. Let's see, Jeff's setup. We finally get.

Paris Martineau (87:52)

Where is it?

Jeff Jarvis (87:54)

In the. In the discord. I put it in there. Did I not hit.

Paris Martineau (87:58)

Did you.

Jeff Jarvis (87:59)

Hold on. I didn't hit return yet.

Paris Martineau (88:03)

When you. When you post on the Internet, you gotta make.

Leo Laporte (88:05)

Oh, that's a nice setup. That's very pretty. No, that's not his. Wait, no, now it is now.

Paris Martineau (88:10)

Oh, whoa. I can't believe you're.

Leo Laporte (88:11)

I like all the greenery.

Paris Martineau (88:13)

You're looking outside.

Jeff Jarvis (88:15)

Yes.

Leo Laporte (88:16)

Isn't that nice? This is why you're pink.

Jeff Jarvis (88:19)

Yes.

Leo Laporte (88:20)

Windows in front of you.

Jeff Jarvis (88:22)

Oh, yeah. It goes up through Jeff.

Paris Martineau (88:24)

This is a great desk. Oh, my gosh. And you've got an old computer. You've got an old TV screen. For some reason, I always imagined you having, like, a crtv. Like a. Like a cr. Like a. Like a. Like an old. Like one of those big box Catherine Ray tubes. Yeah.

Leo Laporte (88:40)

I like the clock. I like the analog clock. I haven't seen one of those in a while.

Paris Martineau (88:44)

Yeah, wait, I've got an analog clock in a sense. I've got a flip clock.

Leo Laporte (88:48)

Oh, a flip clock.

Benito Gonzalez (88:50)

No, that's digital yet.

Paris Martineau (88:53)

Well, it's. It's. It's physical. It's not digital, but.

Leo Laporte (88:57)

All right, come on.

Paris Martineau (89:01)

In the sense.

Jeff Jarvis (89:02)

Okay. AI time. AI time.

Leo Laporte (89:05)

Let's see. AI time. We talked about. We talked a little bit about. We talked about Mark Zuckerberg's manifesto. I honestly think. Well, it's interesting. The only thing. If it were anybody but Mark, I would say, well, that's exciting because here's somebody who's putting so much money, you know, kind of completely speculatively into making us a super intelligence.

Jeff Jarvis (89:34)

And it's an optimistic view. You're going to have your own assistant to make life better versus Sam Altman is.

Leo Laporte (89:39)

It's gonna kill you. Yeah. You know, he says, over the past few months, we've seen glimpses of our AI systems improving themselves, which is an interesting way to put it. The improvement is slow for now, but undeniable. Developing superintelligence is now in sight. My only issue is. And I. Maybe you. Maybe you know better. You know Mark better than I do. Maybe I. I should give him more credit, is. I don't know if I want these billionaires to be controlling this thing. That should be a utility. It's just like, you know, billionaires shouldn't control Twitter. But do you trust Mark to. To make something that is altruistic? And for the good of humankind.

Paris Martineau (90:23)

I don't trust anyone to do anything.

Jeff Jarvis (90:25)

No, that's because you're a dialist. I trust him more than Sam Altman.

Leo Laporte (90:30)

Yeah. I think he's kind of on the.

Jeff Jarvis (90:32)

Same Google more than meta.

Paris Martineau (90:35)

Why do you trust him more than Sam Altman? And how.

Jeff Jarvis (90:39)

One word. Paris. Guess what it is. Guess what it is. Test. Grail.

Paris Martineau (90:45)

Drink.

Leo Laporte (90:49)

So who's the tesreal? Oh, Sam's the test reality. Yeah.

Jeff Jarvis (90:52)

I mean, he's in that doomer teal world. Right.

Leo Laporte (90:56)

Mark is not. Mark's very positive. He says in some ways this will be a new era for humanity. AI will improve our existing systems, enable the creation and discovery of new things that aren't imaginable today. As recently as 200 years ago, 90% of people were farmers growing food to survive.

Jeff Jarvis (91:15)

It's even worse than that. Right? I don't think, but.

Leo Laporte (91:18)

Well, here's an interesting. I'm reading a very good book about the Civil War because I'm going to be going down. Down. I'm going to travel to Mississippi soon.

Jeff Jarvis (91:26)

Mark Twain.

Leo Laporte (91:27)

Mark Twain. Going. Going up in the riverboat. One of the things he points out. Prior to the Civil war in the 1830s, it cost more to ship a product 30 miles inland than it did to ship it across the Atlantic. So colonial and later Americans would buy stuff from Europe much more readily than from their neighbor 30 miles down the road.

Jeff Jarvis (91:53)

If you lived on the coast, at least.

Leo Laporte (91:55)

Yeah. There weren't any roads. Yeah. If you lived inland, you were out of luck completely. So almost everybody who didn't live in a port grew their own food, ate their own food, made their own clothes, made their own candles, made everything. You know, maybe there was an artisan down the road who would help you make your plowshares and your. And your sickles. But this wasn't. It wasn't easy to move stuff around in this country. So I don't. 200 years ago. What is that? 1825.

Jeff Jarvis (92:25)

I tried to look it up and I don't know. 18.

Leo Laporte (92:28)

I think he's accurate.

Jeff Jarvis (92:29)

It was.

Leo Laporte (92:30)

I think he's accurate. This is this guy, this book I'm reading, which is very good, by the way. I can't remember the name of it, but. The path to liberty. No, anyway, I'll find the name. But he says 1830, 200 years ago, everybody was growing their own stuff and making their own stuff because it was so expensive to transport it, because we didn't even have macadam roads. We had rutted dirt roads. So the cost. And we didn't have railroads yet.

Benito Gonzalez (92:58)

No, we had railroads. We had railroads.

Jeff Jarvis (93:00)

Well, we had railroads.

Leo Laporte (93:01)

Didn't have them go anywhere, you know. Yeah, we were just beginning railroads.

Benito Gonzalez (93:05)

Yeah, we had.

Leo Laporte (93:06)

Just beginning.

Benito Gonzalez (93:07)

We had just started the extermination of natives in the West.

Leo Laporte (93:09)

So, yeah, the transcontinental railroad was transformational. In fact, what was weird did you see this? Union Pacific just merged with another railroad to become. To form a new transcontinental railroad. We haven't had one in years. There were all individual. You'd have to trans. Trans ship individual trains. Okay. You're on Union Pacific till Chicago, but then you have to go on Illinois Central. Then you have to. I mean, it was crazy anyway. For free.

Paris Martineau (93:35)

I love railroads.

Leo Laporte (93:36)

I do, too.

Paris Martineau (93:37)

It was delightful being in the Pacific Northwest because I was like, wow, there's so many trains transporting. I was gonna say so many things, but mostly wood. There's so many trains transporting.

Leo Laporte (93:47)

Mostly logs.

Jeff Jarvis (93:47)

Yeah, a lot of logs.

Paris Martineau (93:49)

But it was so nice to see a lot of logs.

Leo Laporte (93:51)

He said, I'm extremely optimistic that superintelligence will help humanity accelerate our pace of progress, not dig our own graves. I hope not. As profound as the abundance produced by AI May1day be, an even more meaningful impact on our lives will likely come from everyone having a personal. A personal superintelligence that helps you achieve your goals. I think this is really an interesting point of view, the idea that you will have a little buddy that. And we're. We're kind of close to that, aren't we? I mean, that's.

Paris Martineau (94:25)

I mean, I guess we're close in comparison to like 50 years ago or 200 years ago, but I don't know.

Leo Laporte (94:30)

10 years ago.

Paris Martineau (94:31)

Yeah, or 10 years. I mean, we're close comparatively, but I don't think that the current Paris. When I was your age are very close to that.

Leo Laporte (94:41)

When I was your age. If we were having a debate about who was the greatest hitter of the 1920s, I had to go to the library.

Jeff Jarvis (94:50)

Well, but that's the Internet.

Leo Laporte (94:51)

That's not AI but no, I'm just pointing out we are in a continuum of information availability that AI accelerates even farther, and that's huge. I'm doing research now about medical conditions. I. I give AI. I just did this the other day, all my supplements and medications and said, tell me if there's any interactions I should know about. My doctor only has 20 minutes for me. This is pretty typical in the United States. Doctors have about 20 minutes. A patient, he doesn't have Time for me to say, well here's the supplements I'm taking. Can you tell me if there's anything I should worry about, if there's anything I should be doing? What should I take with meals, what shouldn't I. All of that stuff. AI has the time and I checked it. I did it by the way, I did it in Perplexity Chat, GBT and Anthropic and cross referenced them. I did a model of experts by hand and they, and actually they found that said that okay, you have your too much vitamin D, you're risking the magnesium and do not take that much chromium and B12. And it was accurate. I checked it. Of course I didn't. I'm not going to act on just an AI saying to do this. But that was hugely valuable and I was able to query it in a way that you couldn't query in a library book. You couldn't really even query Google. It's able to synthesize information.

Jeff Jarvis (96:13)

You could have asked your pharmacist.

Leo Laporte (96:16)

Yeah. Have you asked a pharmacist lately? So when I got prescribed, when I got prescribed Ozempic, this is a really good example. There is a lot of stuff you should know. Nobody. The doctor didn't have time to tell me. The pharmacist showed me how to inject it. I said, you know, I don't usually talk to the pharmacist but I did as this time. I said I need to know how to. She showed me. She didn't tell me any of the important.

Paris Martineau (96:39)

You go to a big box pharmacy or a.

Leo Laporte (96:41)

No Kaiser my hmo.

Paris Martineau (96:44)

But I mean is that a pharmacy?

Leo Laporte (96:46)

They have a pharmacy as well. Yeah, that's one of the nice things about HMOs is it's all in one spot. Labs, pharmacies, X rays, everything.

Paris Martineau (96:52)

I don't obviously I don't know the Kaiser experience but I've, I have found I take a couple different medications and I've in. I'd previously gone to Capsule, which is like a startup delivery service or gone to big box stores or whatever works best for my insurance and I found that going to a local pharmacy has really changed the game for me because I know my pharmacist like I can call him and ask him stuff about the supply chain for my medication or like ask him like involved questions about how certain medications.

Leo Laporte (97:23)

Well you're, I think you're fortunate it but I mean I think, I don't think that's the universal experience at all.

Paris Martineau (97:29)

Yeah, I have heard from people in a wide variety of cities that like local. That is what local pharmacy tests offer that a large chain often does not.

Jeff Jarvis (97:40)

My local pharmacist, who I like would like to use, for example, we went to her to get our Covid shots and she said, well, she did it to be nice to people, but she wasn't paid the way the big box was.

Leo Laporte (97:52)

Yeah. And. And somebody saying doctors should have. You should have more than 20 minutes with a doctor. Of course you should. I'm not saying that you shouldn't have all this. I'm just saying because in this, at least in the United States, we don't in many cases and. But we do have access to information in a new way that I think is very powerful.

Jeff Jarvis (98:12)

So. So Zuckerberg says, the thing I argue with is the super intelligence BS to the idea that you build something that again, you've got to have trust in it because privacy issues galore. But do you trust something? But this is what Comet is trying to do, this is what Google is trying to do, this is what they're all trying to do, is to have the context of you and your life. It's what your devices are trying to do so that you can then ask it a question and it has that context. That level of personal service is appealing, but there's going to be lots of moral panic about it.

Leo Laporte (98:46)

Do you think Amazon Medical one.

Paris Martineau (98:49)

Amazon won?

Leo Laporte (98:51)

Oh, no, I'm turning into.

Jeff Jarvis (98:56)

We haven't seen that in a while. Paris was on medical one for a while, right? Dr. Dave it's one medical.

Paris Martineau (99:03)

I still am. I literally mine Amazon bottom. Amazon bought it. I haven't switched mine over. I was just trying to decide today whether I should switch. Switch mine to connect with my Amazon account because they give you like a hundred dollar discount or something.

Jeff Jarvis (99:17)

But you lost Dr. Doug, right?

Paris Martineau (99:20)

My longtime doctor, Dr. Dan. I lost not because it was purchased by Amazon, but because he moves to D.C. but I have a new primary care doctor and I get along with her quite well. I mean, I've really enjoyed one medical though. I do, I will say every time I go in, I mean this is just typically what happens with me whenever I like interact. The primary care doctor like knows me because they know I'm a journalist that used to cover Amazon. I always like ask like how things are going and they're like. And I will say as of late, some people who I interact with in my visits have been like, yeah, we have a lot more patients to see or yeah, they're, I think part of one. One thing that A person mentioned recently, recently is. So one of the things I like about One Medical is that if you're a member, you can at any point basically open your phone, do a live video chat with a doctor, a nurse practitioner. It doesn't. Like, I'm not getting billed for insurance by it. Like, you just get to like ask somebody some questions. And they're usually available like instantly. They are now opening that network up to any Yahoo with an Internet connection that wants to play like a flat rate to basically do a virtual urgent, urgent care. But at least what this one. These two people I spoke to at the one Medical facility I went to a couple months ago said they're like, yeah, it feels like they're not increasing the staff at all. Like, it's still just the same one medical employees and doctors and NPs that work in the facilities that are also manning this in their off hours, which I think it's just kind of emblematic of. What often happens when a large tech company buys what used to be kind of a bespoke niche product, is they're going to try and get that sort of mass scale. And there might be downstream effects from that. I haven't noticed anything personally in terms of quality yet, but I'm keeping an eye out for it, certainly.

Leo Laporte (101:12)

Yeah. Anyway, it's good we have AI because our medical system's falling the pieces. That's not a good reason. And they're all going to be, you know, I mean, look, I'm not in immune, despite all the sand in my shoes, to all of the negative outcomes that AI is going to create, including job loss, dramatic job loss. And we're fortunate because we can afford to use these things. I think there are a lot of people who will not have, will not be able to afford food soon enough, let alone 20amonth for an AI.

Jeff Jarvis (101:48)

And the risk is we're going to have more walled gardens. Amazon is cutting off Google from scraping it. And so you go to Google and ask for a price, you won't get Amazon prices, you'll have to go to Amazon to get Amazon prices. Right? Yeah, kind of stuff. And that's, that's going to be the game that goes on. It's back to AOL land.

Leo Laporte (102:12)

All right, let's move on. I want some positive. I want some positive.

Jeff Jarvis (102:17)

You want a fun story? You want a fun story?

Leo Laporte (102:19)

Line 90, line 98. I always loved line 98, one of my favorite lines.

Paris Martineau (102:26)

It's a good line.

Leo Laporte (102:28)

Empirical evidence of large language models influence this one.

Jeff Jarvis (102:31)

Yeah.

Leo Laporte (102:32)

On. On humans.

Jeff Jarvis (102:33)

My joke by joke in the line. Read the line first. Now read the line in the. In the. In the rundown.

Leo Laporte (102:39)

In the rundown, it says, you might want to delve. Dell. I got it. Delve into this paper. I want to. By the way, there's a new AI company called Delve Delf. I want to underscore. That's a joke you'll comprehend only with the malicious, meticulous reading of it. Do you mean underscore or did you mean M dash?

Jeff Jarvis (102:58)

That's. That's bs. M dashes are fine. So this is a study that. That looked at language in many podcasts and academic YouTubes.

Leo Laporte (103:08)

No.

Jeff Jarvis (103:09)

Or. And after ChatGPT, to find the impact of ChatGPT on language.

Leo Laporte (103:14)

Oh.

Jeff Jarvis (103:14)

And so there's a series of words here.

Leo Laporte (103:17)

Delve, comprehend.

Jeff Jarvis (103:18)

Delve, underscore, comprehend, bolster, boast, swift inquiry, meticulous, pinpoint, surpass, craft, intricate, heightened, lesson, discern, necessitate, showcased amidst escalate and intricacy.

Leo Laporte (103:32)

You know, what this really is, is kind of business speak, isn't it?

Jeff Jarvis (103:35)

Well, there's also journal words. Underscore is a beloved. A word beloved of the New York Times.

Leo Laporte (103:39)

So, like, it really is underscoring their. Their desire to make the world a.

Jeff Jarvis (103:44)

Better place, showcase, escalate. Those are journal words. Lots.

Leo Laporte (103:48)

So that's why. But it makes sense.

Jeff Jarvis (103:50)

It's trained on journalism.

Leo Laporte (103:51)

It's true. Yeah, Makes sense.

Jeff Jarvis (103:53)

But my theory is that journalism has less impact on the language than Chat GPT does.

Leo Laporte (104:00)

But ChatGPT is not making it up at a whole cloth. No, no, I know that somewhere.

Jeff Jarvis (104:04)

But this is finding that there's been a market increase in these, even journalistic words after chatgpt.

Leo Laporte (104:11)

Interesting.

Paris Martineau (104:11)

Boast. Boast is such a journalistic word.

Jeff Jarvis (104:14)

Yes.

Paris Martineau (104:14)

I'm always using the word boat.

Ian Kreitzberg (104:16)

Yeah.

Jeff Jarvis (104:16)

Yeah.

Benito Gonzalez (104:18)

Well, it's like it's the same thing, though, right? Because the journalists are the ones who fed chat GPT and now. Yeah.

Leo Laporte (104:24)

So it's the same thing. Yeah.

Jeff Jarvis (104:26)

This is Matthew Kirschenbaum's text apocalypse. It just feeds on its entrails.

Leo Laporte (104:30)

Yeah. I mean, the same thing probably would have happened anyway, because YouTubers read the news, newspapers, and they. They want to sound more professional, so they end up.

Jeff Jarvis (104:38)

Who thinks YouTubers read newspapers?

Benito Gonzalez (104:40)

This is also how we lose, by the way.

Jeff Jarvis (104:42)

They don't read newspapers.

Leo Laporte (104:43)

Not newspapers. You know what I mean? Journals. What do they call them?

Benito Gonzalez (104:47)

Anyway, this is how we lose words, too. Like bloggers, you know, people overusing words, and then they suddenly become mean, less meaningless.

Leo Laporte (104:54)

Well, so let's be honest the reason you use the word boast is because it's a, it's a. We're talking of, you know, what do you call it? You know, the word ask.

Jeff Jarvis (105:09)

Ask. Chat. GPT for the word you're looking for. It's very good.

Leo Laporte (105:12)

You know, when you go to the. Not Bartlets.

Jeff Jarvis (105:15)

When you go to the source thes.

Leo Laporte (105:18)

That's the word.

Jeff Jarvis (105:18)

You need to find the source word.

Leo Laporte (105:21)

Because you didn't. You already said show. Have you want to give it another way? So you say boast. Right. That's really where these come from. They're just trying to vary your speech so you don't use the same word twice in this study.

Paris Martineau (105:35)

That's why.

Leo Laporte (105:36)

Right, yeah. That's why all these words exist.

Paris Martineau (105:38)

It's not from even at the start. It's just you're trying to, you're looking at, you're like, I can't have show three times in a paragraph.

Leo Laporte (105:44)

They're euphemisms almost. They're, they're.

Jeff Jarvis (105:47)

No, not euphemisms. Studied 360,000 YouTube videos of academic talks and 771,000 podcast episodes. So I.

Leo Laporte (105:59)

Did they do that with AI? Did they do that with AI?

Jeff Jarvis (106:02)

Of course they did.

Leo Laporte (106:03)

Okay. Okay.

Jeff Jarvis (106:10)

7.35 billion transcribed words. No, they had, they had graduate assistants typing all of them.

Leo Laporte (106:17)

Yeah. You know, it's a word that podcasters have brought into the world that I hate is pod, where they say, listen to our podcast.

Jeff Jarvis (106:25)

Yes.

Leo Laporte (106:26)

Like, this show is a pod. Don't call it a pod.

Jeff Jarvis (106:29)

I can't stand the people say that a blog is the post.

Leo Laporte (106:33)

Yes.

Jeff Jarvis (106:34)

The blog is the whole.

Leo Laporte (106:35)

I wrote a blog. No, you wrote a post on your blog.

Jeff Jarvis (106:37)

Yes.

Leo Laporte (106:38)

You didn't write a blog.

Jeff Jarvis (106:39)

Right?

Leo Laporte (106:40)

Yeah, but see, that's not AI doing that. That's humans doing that.

Benito Gonzalez (106:44)

That's what I said earlier. That's how we lose words.

Leo Laporte (106:46)

That's how we lose words, too. I, I, after reading John McWhorter, which you turned me on to, and I love. He's really, he's the kind of linguist who is a little bit more lackadaisical, saying, this is how language evolves. Language evolves. It doesn't sit still.

Paris Martineau (107:03)

I do think that there is a context for using blog to mean posts that can work.

Jeff Jarvis (107:08)

Like, if you are talking about, watch out.

Paris Martineau (107:11)

If you're talking about, let's say, your, both of your guys's blogs that you just described that I'm now forgetting the name of. Those are blogs, pure and simple. So they're are posts on them. But if you're talking about, for instance.

Leo Laporte (107:24)

That'S how you took our. Our advice to you is it's already. It's gone. It's.

Paris Martineau (107:28)

I. My. My brain retains concepts. It does not retain proper nouns. This has always been a problem. It's a proper noun and there's none.

Leo Laporte (107:37)

I couldn't remember thesaurus about a three seconds ago.

Paris Martineau (107:40)

So I think that referring to those, those are posts on a blog, but something like, like a news website that contains more than like blogging style posts. You can have a blog. Someone can do a blog on Business Insider. And describing it as a blog differentiates.

Leo Laporte (108:01)

Blog is a connect. No, no, but blog is a. Blog is the collective noun for a bunch of posts.

Jeff Jarvis (108:07)

It is not like saying, Paris, I wrote a magazine. No, you wrote an article in a magazine.

Paris Martineau (108:15)

I don't know. I think I have to disagree. I think I'm on the other side of the divide from.

Leo Laporte (108:19)

You see that murder over there? No, that's one crow. If there were a bunch of them, it would be murder. But it is just a single crow. You know what I'm saying? It's not. Don't use the collective for the singular. For the individual.

Jeff Jarvis (108:34)

Yes.

Paris Martineau (108:34)

I think medium society moved beyond the understanding of blog. It's like. It's like media. It's like. Yeah, it's exactly like the word generational.

Ian Kreitzberg (108:43)

Media is plural.

Paris Martineau (108:43)

I don't know that it is plural. The media can be.

Leo Laporte (108:47)

Yeah, it's not plural anymore, Jeff.

Paris Martineau (108:49)

It's not plural anymore in the way that blog isn't plural anymore.

Jeff Jarvis (108:53)

My problem is I write about mass media. It doesn't really make sense to make that plural. But media to me is plural.

Leo Laporte (109:00)

You don't want to call it the mass medium.

Jeff Jarvis (109:03)

That would make no sense.

Paris Martineau (109:04)

No, the mass medium is the message.

Leo Laporte (109:08)

Are we excited about chat GPT 5.0? It's coming out any day now. Do you think it's going to be. Is it going to be all that?

Jeff Jarvis (109:16)

It's AGI. I know it is. It's going to kill us all is what I said last week. They'll leapfrog each other. It'll be. It'll be more interesting. It'll do more stuff. It'll pass more tests.

Leo Laporte (109:28)

Google's quarterly results. $96.2 billion in three months. $28 billion. Just profit.

Paris Martineau (109:38)

How does that compare to previous? How does that compare to the last quarter or the year before?

Leo Laporte (109:43)

I don't know. It seems. Look, they're doing well, let's put it that way. They're doing better than Microsoft. We'll get Apple's results tomorrow but I have a feeling they're probably not too far off From Apple stocks 200 bucks a share and it's because you know why? Because their ad revenue is through the roof. Because they dominate the business. Between Meta and Google it's 90% of all digital ad sales go to Meta and Google, not us. I wasn't talking to Google. Stop listening.

Jeff Jarvis (110:20)

You can, it's also everybody else can listen to them but not Google.

Paris Martineau (110:24)

I mean search now I believe there are some search statistic that the majority like a significant percentage of Google searches. People just get now their information from looking at the summary. Don't ever click through to the source link much less the any links below.

Jeff Jarvis (110:43)

In the search results line 151 gives you the numbers according to Google has.

Leo Laporte (110:49)

A message for you.

Paris Martineau (110:49)

By the way, trained by Google, I do not have the ability to listen to your conversations. I'm a large language model trained by Google. I do not have the ability to listen to your conversations.

Leo Laporte (111:05)

I don't know why she ends on an up like that. Maybe she's young. She's young, she's a millennial.

Jeff Jarvis (111:11)

So almost 50%.

Leo Laporte (111:13)

Why did she say that? How would she know if she wasn't listening?

Paris Martineau (111:18)

Yeah.

Leo Laporte (111:19)

Do you. You lie. Okay, I am a large language model. I don't listen to you. Well then how do you know I was talking to you?

Jeff Jarvis (111:31)

AI When AI overviews are present, publishers witness a drop of 47.5% in click through rate on desktop and 37.7% on mobile.

Leo Laporte (111:39)

Yeah.

Jeff Jarvis (111:39)

Which is just people are getting the information they want.

Leo Laporte (111:43)

Well it's good for us, it's good for the users but it's very bad for the people who create the content that we want.

Jeff Jarvis (111:50)

People who write a as an industry is over. It's gone.

Leo Laporte (111:53)

Yeah, we're at Google zero is what they call it now which is that we're, we're as a publication you can expect zero traffic from Google.

Jeff Jarvis (112:01)

Google told the Guardian that the study was inaccurate and based on flawed assumptions. Which I'm sure is true too.

Leo Laporte (112:06)

Maybe, but it's obvious to everyone that this is the transition that's happening right now. Although the Wall Street Journal it's pretty funny. They had a graph in which they showed that we're talking about Google zero and they had a graph that they showed that business insiders traffic had declined dramatically. Well but what they didn't mention is because they have a paywall that you cannot get through that you can't read their articles Unless you give them a lot of money. It wasn't AI or that killed because.

Jeff Jarvis (112:38)

They used to be. The whole thing about Business Insider used to be that they, they, they did to everybody else what they now complain is being done to them by AI.

Leo Laporte (112:46)

Right. It was pretty ironic.

Paris Martineau (112:49)

Organic search traffic to Business Insiders websites declined by 55% between April 2022 and April 2025.

Leo Laporte (112:57)

But is this search.

Jeff Jarvis (112:58)

I don't click on it anymore because there's no point.

Leo Laporte (113:00)

I click every time I click on it, I see the paywall. I go, well, I'm not. I can't. No. So this is news. News sites are getting crushed by Google's new AI tools as the Wall Street Journal. And here's the graph.

Jeff Jarvis (113:12)

Anti technology.

Leo Laporte (113:14)

Wall Street Journal, Business Insider, Washington Post. Again, traffic might be down in the dumper.

Jeff Jarvis (113:21)

Journalist, another story.

Leo Laporte (113:23)

Yes, Interesting. The Journal, which does have a paywall, but not as draconian, is doing okay. And the Huffington Post is going down. But who reads the Huffington Post?

Jeff Jarvis (113:32)

I know.

Paris Martineau (113:34)

Journalism producing part of buzzfeed now left.

Leo Laporte (113:38)

Is it it. Oh, are we getting Petti on at some point?

Paris Martineau (113:42)

It's a great question.

Leo Laporte (113:43)

I thought we were on the list. Oh, he's on the list. That doesn't mean anything. We're trying to get Jonah ready for. For everybody. So. But that is a big.

Jeff Jarvis (113:52)

Trying to get him logic too. And he said yes and then.

Leo Laporte (113:55)

Then ghosted us.

Jeff Jarvis (113:57)

Yeah.

Paris Martineau (113:59)

Henry, you'll have a good opinion on blog singular or blog plural?

Jeff Jarvis (114:08)

Blog.

Leo Laporte (114:11)

Journalists. You two are journalists. News sites, are they getting crushed by Google's new AI tools? Is that a fair headline?

Paris Martineau (114:19)

I mean, Google's AI tools are. Feel like are the latest nail in the coffin. I think that changes to SEO as well as just changes to the ways that sites are getting. Traffic has decimated news sites. It's a. It's been a thousand cuts.

Leo Laporte (114:39)

Yeah.

Paris Martineau (114:40)

Every site is now leaning in so hard to newsletters because it's a way to try and get around these, like, indirect relationships with your audience and instead try to have like a direct. Yeah, a direct relationship with a subscriber to a product that you can very easily.

Jeff Jarvis (115:00)

So I got a question.

Paris Martineau (115:01)

Engagement.

Jeff Jarvis (115:02)

So, Paris, I subscribed a bunch of newsletters and then I find that I just stopped reading them all.

Paris Martineau (115:06)

I was gonna say I. I don't read many of the newsletters I subscribe to, but I subscribe to a lot.

Leo Laporte (115:12)

Me too. They all go into a folder where I can ignore them easily.

Jeff Jarvis (115:15)

Yes.

Paris Martineau (115:17)

I mean, this is something I was.

Leo Laporte (115:19)

I still read sites I Have an RSS reader. And every day, twice or three times a day, I go to my RSS reader and scan for stories for this show and all the other shows. And I re. And I go to the sites and I read the stories and I wear these glasses. You know why I wear these glasses? Because Mark Zuckerberg says people without AI Glasses will be at a disadvantage in the future.

Benito Gonzalez (115:41)

Big surprise. People who have money will have an advantage over people who don't.

Leo Laporte (115:45)

People who wear the glasses I make and sell at great expense will have an advantage. Actually, there's going to be a new set of these coming soon. Right, Right.

Jeff Jarvis (115:56)

Well, the, The. Well, there's Luxottica. No, the whatchamacallits are out now. The.

Leo Laporte (116:00)

The. The Oakleys.

Jeff Jarvis (116:02)

The Oakleys.

Leo Laporte (116:03)

Thank you. Yeah, they're. They're more expensive. I don't know if they're better.

Jeff Jarvis (116:06)

500 bucks.

Leo Laporte (116:07)

Yeah. I'm looking for what?

Jeff Jarvis (116:09)

Jason. Jason got them. He talked about them today.

Leo Laporte (116:11)

Oh, okay.

Jeff Jarvis (116:12)

And they probably look better than these. They look. I don't. I don't think you and I at our age can pull off the Oakleys. I think Jason can, but I don't think if. If you go to.

Leo Laporte (116:24)

Are they like Oakley wraparound?

Jeff Jarvis (116:26)

You go to AI inside, you'll see.

Paris Martineau (116:27)

They'Re the shiny ones, right?

Jeff Jarvis (116:29)

Yeah. You'll see they're all white. He said the quality of the video, he wanted it when he is doing product stuff to just show. And the quality of the video wasn't that.

Leo Laporte (116:39)

Oh, I would wear those. You know who they're copying? Casey Neistat. Oh, yeah. Or I guess he has those white glasses, although his are painted white and the paint's coming off, which is even better. I mean, it's a style. It's a thing. It's a look. It's a statement. I look. They have black ones that look just as nerdy. Yeah. So j. So Jason got the white ones. Yeah. I'm waiting for a improved product. I don't. I don't care about it being a style maker.

Jeff Jarvis (117:14)

The question is. So, so do I want visual cues as well? Not just audio heads?

Leo Laporte (117:21)

You mean heads up? Yeah, yeah, the heads up ones. And I've tried a few. As you know. I don't. I don't think they're quite there yet, but Google bought leads. I mean, I think because of the way they display it. You know what you. I feel like what I want is a dis. I can look at the world. I see the world and superimposed on it without kind of it Being a little mail slot or anything but superimposed on it is information and it's not that way yet.

Benito Gonzalez (117:48)

Stuff, you know, in your inventory. It looks like it's supposed to look like a game, right? That's what you want. You want.

Leo Laporte (117:53)

It's like a game hud. That's what it is. It's a game hud. Exactly.

Jeff Jarvis (117:56)

You have a heads up display in the car.

Leo Laporte (117:58)

Yes, and I love it.

Jeff Jarvis (117:59)

So is it, is it, is it analogous to that?

Leo Laporte (118:02)

Yeah, actually it is. In fact some cars, now I don't have it, but some cars will use the heads up display to show where your car is going based on the steering. So they'll show the road. It'll be like a video game where they show the road and the road.

Jeff Jarvis (118:18)

You kind of should just drive.

Leo Laporte (118:21)

I drive. But you know what it has, that's really important? It has a speed. Okay. So I have to look down and see the speedometer and then it sees stop signs and red lights and yield signs and it sees traffic signs. So it displays in the heads up display the stop sign before I see it in my peripheral vision. Which is actually I think a really good feature.

Paris Martineau (118:40)

I was amazed when you drove me around Petaluma. I think your car backed up out of the, the driveway by itself, right?

Leo Laporte (118:48)

No, I was driving. It offered to you. You saw the display that said I can, I can take it from here if you want, but I don't trust it. So Yann Lecun, tell me what he says. He says he's not out in the, in the Siberia of Meta. He is still involved.

Jeff Jarvis (119:07)

Right. Even though working on, he's working on the future stuff. And he has said that LLMs won't get us to where Mark wants to go. Yeah, I'm not going to take the next paradigm. He's, he's very big on believing that next goal which we've talked about in the show is reality checks, understanding reality, that that's going to be what matters. And as you've said, robotics are a key feeder of that. So he's looking for the next paradigm and the next paradigm.

Leo Laporte (119:37)

I would hate to make any blanket statements at this point because LLMs have done so much more than one would thought. I mean I, I called them spicy autocorrect.

Jeff Jarvis (119:47)

But just logically as what, what Yann Lecun says is that, that we're gonna need more. Not as smart as a house cat.

Leo Laporte (119:54)

Yeah, I know he says that.

Jeff Jarvis (119:56)

I don't, I think, I think Gizmo is much smarter than ChatGPT. Don't you, Paris?

Paris Martineau (120:01)

I think Gizmo is much smarter, but significantly less likely to follow any instruction or be sycophantic towards.

Benito Gonzalez (120:09)

Wait, so Leah, you think that transformer technology, that's enough, that's enough to get super intelligence that's already it?

Leo Laporte (120:15)

I don't know. Because. I don't know. But I wouldn't rule it out because it's already done so much more than one would have thought.

Benito Gonzalez (120:24)

So has a silicon chip like.

Leo Laporte (120:27)

Yeah, but okay, so on the face of it, what a transformer looks like is a prediction machine. You give it a token, it predicts the next token and the next token, the next token. It's hard for me to go from that fact to what they're doing now, which seems to be somehow generative. And so the real question that's in my mind, and maybe it's a little hocus pocus, but is there some generative magic that's happening in this predictive model that we didn't really anticipate? So I'm just saying I'm not willing to rule out how far LLMs can go now. I'm not. Yann LeCun.

Jeff Jarvis (121:05)

Well, here's the other one I would go to. I would call on Jensen Huang. I think that the appropriate use of that predictive ability is the digital twin is that whether you're a factory or a warehouse or a car or potentially us in life, here are the next things that could happen.

Leo Laporte (121:23)

Well, get ready for this.

Jeff Jarvis (121:24)

Good use of that.

Leo Laporte (121:25)

Because one of the tools that these companies are using is VO3 and other video generators like run way to generate content to train about the physical world, to train LLMs to understand the physical world. So in effect, they're being trained on transformer created content about the physical world. Well, I think you could eliminate that middleman. Then. You're, you're saying, okay, here I'm going to give a prompt to a video generator that's going to physically do the physical, physical world and then I'm going to train my LLM on it, eliminate the middleman. The VO is not doing anything that the LLM is not. It's the same. It's still transformers all the way down.

Jeff Jarvis (122:09)

Well, it's challenges. This is what Jensen Wong points out. When it comes to cars that take all of the, of the video that we now have for cars, it doesn't throw in every possible challenge to the car. So they create.

Leo Laporte (122:21)

They need to have children running across the street.

Jeff Jarvis (122:23)

Yes, exactly.

Leo Laporte (122:24)

Yeah, yeah, yeah. All right. I think you as blog posters And Paris as a future blog poster should read about.

Paris Martineau (122:37)

I've blogged.

Leo Laporte (122:41)

Sure.

Paris Martineau (122:43)

I came up blogging.

Leo Laporte (122:45)

Did you. You did you have a. Did you have a live journal when you were young?

Paris Martineau (122:50)

I had a Tumblr.

Leo Laporte (122:51)

Tumblr. Oh, that's right.

Jeff Jarvis (122:52)

Oh, yeah, that was. That was hot.

Leo Laporte (122:53)

What did you tell Personal?

Paris Martineau (122:56)

I don't know. Various things. Probably a bunch of stuff related to the fan fiction I wrote.

Leo Laporte (123:00)

Nihilism.

Paris Martineau (123:02)

I wasn't nihilistic. I was really optimistic for a while.

Leo Laporte (123:06)

Yeah, that's.

Benito Gonzalez (123:07)

Those aren't exclusive. Those two things aren't exclusive.

Paris Martineau (123:09)

I know. I'm just saying it wasn't my def. I wouldn't have described myself as a nihilist.

Jeff Jarvis (123:15)

You had your goth period, right?

Paris Martineau (123:17)

Oh, yes, but the goth period. I think I was still a ball of sunshine.

Leo Laporte (123:21)

I would love to see your goth look. Did you have some photos? Dark eyeliner, pale face, makeup, Bang.

Paris Martineau (123:32)

I mean, I think I was just. I was just pale, generally.

Jeff Jarvis (123:37)

Did your mother wish you looked happier?

Leo Laporte (123:38)

Black nail polish.

Paris Martineau (123:41)

I think my parents tried to be really supportive of it, but, like, they tried to. They'd, like, give me fashion suggestions that were, like, actually nice instead of goth or sane, which wasn't what I was looking for. Looking for it to be kind of crummy and scuzzy as the song.

Leo Laporte (124:00)

So, from the Electric Eclectic Light Company, which is a lovely blog by a painter whose name escapes me. He says, I am going to use this emoji, the owl emoji, on any blog posts that I write that have not been. I have not used any AI on. And he wants others to do this. Use the owl.

Paris Martineau (124:24)

Why the owl?

Leo Laporte (124:25)

Because it's symbolic of wisdom.

Paris Martineau (124:30)

I thought it was symbolic of the aliens. As in Twin Peaks.

Leo Laporte (124:33)

Peaks. Oh, is it? He says owls have long and deeply ingrained associations with the night and with wisdom and learning, going back to the classical civilizations of the Mediterranean. And then he has a naked lady named Minerva who has an owl with her denoting wisdom and learning a famous painting. Yes, but she's naked, so I can't show it, because, as we know, the credit card companies are now demonetizing anybody who has content that is any way seen to be inappropriate for a young person.

Paris Martineau (125:09)

Well, I don't disagree that they're doing that, because they are, and that sucks. But I think there's also something lost here, which is the reason, like, the less sexy, conspiratorial. Well, it's, I guess, still sexy because it still involves new content, but less conspiratorial reason behind all of this is not just that these credit card companies are puritanical and motivated by removing the things we like. It's that sex related content and porn related content historically has really, really high instant like of chargebacks. Like anything relating to that sort of stuff is going to have like a chargeback rating frequency that is so much higher than any other product. So historically these companies don't want to support that because it costs more for them than other stuff. But I agree there's also some puritanical.

Leo Laporte (126:05)

There are also pressure groups, including one in Australia that has convinced them it's not worth it to, to support these Steam.

Paris Martineau (126:14)

Well, part of what they also do is they alert the companies to offending products.

Leo Laporte (126:21)

I haven't. Have you seen these games that they're. They're, they're now taking down off Steam and Itch IO. I mean I don't. Are they really like pornographic or are they just.

Paris Martineau (126:28)

Oh, I mean there are so many games that are basically just porn on Steam. I know this because another one of my things I love to complain about is how the Steam recommendation engine is one of the worst recommendation engines I've ever experienced in my life.

Leo Laporte (126:42)

And they're all pretty bad.

Paris Martineau (126:43)

But I mean that's the thing is you are a company that has extreme granular data about every game I've played it and how long I've played it and all the games I've wished. How do you not. How are you not able to recommend me things that I kind of would probably like? Almost everything they recommend is like a strange free to play game that is probably malware or like a weird porn game or something.

Leo Laporte (127:08)

Do you think they're just ads that they're brain side?

Benito Gonzalez (127:12)

No, I think my theory is that it's actually extreme, extremely hard to recommend games to people. That's one of the hard things.

Leo Laporte (127:19)

It's a hard problem. Remember Netflix had a million dollar prize to come up with a recommendation engine that somebody did won the prize but they don't use it because it's.

Benito Gonzalez (127:28)

Because games are weird. It's like someone can like RPGs but they only like a specific kind of RPG. They don't like these kinds of RPG. You know, that's all.

Leo Laporte (127:35)

It's really weird. All right, we need to take a final break before we're not gonna do the picks quite yet. I have one more story about Tyler, the creator. But first a word from our sponsor. You're watching Intelligent Machines with Jeff Jarvis, Paris Martineau and me. I'm Leo Laporte. This episode brought to you by agency. Building multi agent software is hard. Agent to agent and agent to tool communication. It's still the wild west. How do you achieve accuracy and consistency in non deterministic agentic apps? Ah, that's where agency, the agency comes in. AGNT TC Y the agency is an open source collective building the Internet of Agents. What's the Internet of agents? It's a collaboration layer where AI agents can communicate, discover each other and work across frameworks. It's the future for developers. Agency means standardized agent discovery tools, seamless protocols for interagent communication and modular components to compose and scale multi agent workflows. Close build with other engineers who care about high quality multi agent software. Visit agency.org and add your support. That's agn tcy.org I can get behind this one. Agency.org all right, I. I don't. I don't know if we're going to take. Get taken down if I play this, but it is not Tyler the creator. I just want to say that right now, the glass, you gotta turn on.

Benito Gonzalez (129:07)

Your music original sound.

Leo Laporte (129:08)

Oh yeah. Don't tap the glass. Don't tap the glass. Don't tap the glass.

Benito Gonzalez (129:19)

We're gonna get taken down.

Paris Martineau (129:20)

Is this. Yeah, let's.

Leo Laporte (129:21)

Are we getting taken down?

Benito Gonzalez (129:22)

Not taken down. We're gonna demonetized. We're gonna get demonetized.

Leo Laporte (129:26)

But it isn't Tyler the creator.

Benito Gonzalez (129:27)

I don't care. Someone. This is a new kind of copyright troll. It's not.

Leo Laporte (129:31)

It's not about actual copy. We play played Tom Lehrer passed. Wonderful amazing guy. Which I know you will know well, Jeff, because he wrote. He's from our generation. Wrote amazing parody songs long before Wheelbarrow.

Jeff Jarvis (129:45)

American City. You will find it very pretty. Just two things of which you must beware. Don't drink the water and don't breathe the air. Pollution. Pollution.

Leo Laporte (129:57)

He passed away at the age of 97. But before he died some years ago, he put all of his music, all of his publishing rights, all of his lyrics in the public domain, including all of his recordings. So to honor him, I played his very famous song, the Elements Song where he. To the song to the tune of I am the model of a modern major general sings every element in the periodic table, which is quite a feat.

Paris Martineau (130:23)

Oh, I've heard that one.

Leo Laporte (130:24)

Yeah, we got taken down. No, we got demonetized. Even though it's public domain, it's. It's widely known to be public domain. I played it from his website. It's public domain, but we still. Because somebody. And this is the problem with the YouTube and content ID? Somebody claimed it and they took the money that we would have made from those episodes. That episode. Jesus. Because we got demonetized. So you're right. Even though what I was playing, which you couldn't hear anyway, so it's okay. Was a machine.

Jeff Jarvis (130:53)

We could hear it.

Paris Martineau (130:53)

We could hear it.

Leo Laporte (130:54)

You could sort of hear it.

Jeff Jarvis (130:55)

Maybe it was enough to be taken down. So you gotta. You gotta to buzz it out.

Leo Laporte (131:01)

Okay, so what you didn't hear because we've bleeped it. Tyler the creator has put out an album called Don't Tap the Glass. Top selling album in the country probably. But the hook the Internet is humming. That little piece, that little ditty you didn't hear a minute ago isn't in the song. It was released before the album came out. Somebody. It was purported to be a leak from the album. It wasn't. Somebody used IM to create kind of a bad AI. Not iml. A bad AI version of Tyler the creator and the bad version of the song. And it undercut the actual release actually. Was I playing the real song? Maybe I was playing the real song. And this is. Huh.

Paris Martineau (131:51)

Let's not.

Leo Laporte (131:53)

So this is the video is Tyler the creator and the real song, but somebody put that crap on top of it. The audio comes from a YouTuber named Clodjan who probably would have us taken down if I played it. You're right. He calls it a prank. He did. Faked audio from Rihanna, Justin Bieber, Cardi B. Sample lyrics from his fake Sabrina Carpenter single Man Child. You are a man child. You act like a child. Oh, man child. But one X user wrote that fake AI Don't Tap the Glass leak has been stuck in my head for days. Another called it the fake. The strongest song of the summer contender we've had this year. This is a whole new way of causing a mess on the Internet thanks to AI.

Paris Martineau (132:51)

I mean, yeah, if you can get it to go viral, then I guess the rest doesn't matter.

Leo Laporte (132:54)

Yeah.

Benito Gonzalez (132:55)

If you attach it to some celebrity, yeah, that's gonna go.

Jeff Jarvis (132:58)

All you need is attention, as they say.

Leo Laporte (133:01)

And it's. You know, in a way it's ironic because Tyler the creator is an Internet creation. I mean, that's, you know, he. He succeeded because he. He didn't have to have a record label. He was able to make it on the Internet. So the Internet bites its own, I guess. Hey, let's give some credit as the Wall Street Journal did. We talked last week about ChatGPT and DeepMind getting gold medals in the International Math Olympiad. Wall Street Journal said, turn it on its head. Here's the high schoolers who just beat the world's smartest AI models. This is the right angle. That's the right way to say it. So Google's DeepMind and OpenAI got 35, but these American teenagers got higher scores.

Paris Martineau (133:43)

Good for them.

Leo Laporte (133:44)

Yay.

Jeff Jarvis (133:45)

Does it name them?

Leo Laporte (133:47)

Well, it would be sad if it didn't. Yes.

Jeff Jarvis (133:50)

There's privacy issues or something.

Leo Laporte (133:52)

Oh, maybe they are high schoolers. Yeah. Yeah. Among them were four stars of the US team. No, it doesn't aim. Tiger Zhang, a two time gold medalist from California. Alexander Wang, who brought his third straight gold medal back to New Jersey. One of the most decorated young mathematicians of all time. He's a high school senior. He has one more shot at gold at the IMO next year. But Wang says, I think it's really likely AI next year is going to be able to get a perfect score. He says that would be insane progress. But I'm 50. 50 on it. This might be the last time, in other words, that humans were at the top.

Jeff Jarvis (134:32)

I'm guessing the Zuckerberg.

Benito Gonzalez (134:33)

No, not high schoolers.

Paris Martineau (134:35)

High schoolers.

Benito Gonzalez (134:36)

Remember high schoolers. Schoolers.

Leo Laporte (134:37)

Yeah. Okay, but there's prob. Don't diminish the accomplishment.

Jeff Jarvis (134:43)

Let's put ourselves up against these. High school.

Leo Laporte (134:45)

I couldn't do it.

Paris Martineau (134:46)

Oh, no, I'm not.

Leo Laporte (134:47)

I'm well beyond high school.

Benito Gonzalez (134:51)

Yeah, but these high schoolers when they're 30.

Leo Laporte (134:55)

No, that's the problem with math. They will not be as good when they're 30. Maybe when they're 25.

Paris Martineau (135:01)

No. What if the, these, these people that have won multiple math competitions, you're telling me they're not going to go into some math based career?

Leo Laporte (135:08)

They're going to be on Wall Street. Oh yeah, there'll be quants on Wall street or whatever. But I'm just saying that historically mathematicians make their greatest discoveries in their early 20s. They don't, they don't age well. Yeah.

Paris Martineau (135:20)

Why is that?

Leo Laporte (135:21)

It's hard. Your brain does not, you know, continue. Here are the Deep Mind medal winners. Why are we have a picture of them? I guess they're not high schoolers. Right?

Jeff Jarvis (135:32)

Yeah.

Leo Laporte (135:33)

Oh, well, it is an important, an interesting point. There we go. Here we go with the actual Team usa.

Jeff Jarvis (135:39)

Oh yeah.

Leo Laporte (135:40)

God. Including Alexander Wang, who is fourth from the right holding the American flag. And you see their gold medals right there. Look at those. They look pretty good. In fact, out of the. What is that? I'm gonna have to do some math, but it looks like nine of them.

Paris Martineau (135:56)

Good job.

Leo Laporte (135:56)

Yes. A full six of them, which, if I am correct, is 66% have gold medals. 66.6.

Jeff Jarvis (136:06)

So there's six. She has one.

Leo Laporte (136:07)

Yes, she has one. She's just not showing it.

Jeff Jarvis (136:11)

Being a woman. She's not being as braggadocious. She's not boasting as chatgpt would say.

Leo Laporte (136:19)

Tiger. Zhang is known as Tiger because he's 17 years old, la high school student, on his way to MIT to study math and computer science. As a young boy, his family moved to the US from China. His parents said, you need to take an American name. You have two choices. One was Tiger. The other was Elephant.

Paris Martineau (136:38)

What?

Leo Laporte (136:40)

I think he chose the right one.

Jeff Jarvis (136:43)

He done good.

Leo Laporte (136:43)

His career in competitive math began in the second grade when he entered a contest called the Math Kangaroo.

Paris Martineau (136:50)

Wow.

Leo Laporte (136:50)

And he slayed the kangaroo. Yeah. It ended this month at the Math Olympics next to a hotel in Australia with actual kangaroos. Thank you, Wall Street Journal for that. Full circle. Full circle. It went full circle. The one problem that none of the AIs got was problem six. Zhang spent the longest amount of time during the exam on problem six, a problem in the notoriously tricky field of combination combinatorics, the branch of mathematics that deals with counting, arranging, and combining discrete objects. As I just did, by the way, with mastery with those nine students, it was easily the hardest on this year's test. The solution required the ingenuity, creativity and intuition that humans can muster, but machines cannot. At least not yet. Would you like to hear it?

Jeff Jarvis (137:40)

Sure.

Leo Laporte (137:40)

It's a short problem. Maybe. Darren Oakley, can you get to work on this one? Think of this. And Paul Holder are champions. And if Sci Fi is around, he's another champion in our advent of Code Challenge every year. Problem 6. Consider a 25 by 25 grid of unit squares. Matilda wishes. Matilda's always up to no good. Wishes to place on the grid some rectangular tiles, possibly of different sizes, such that each side of every tile lies on a grid line. And almost every unit square is covered by at least. I'm sorry, by at most one tile.

Jeff Jarvis (138:16)

So I can't have more than one tile. Hold on a second. I can't visualize this. But you really can't visualize this?

Leo Laporte (138:22)

No, actually I can, because it's abstract. I can understand what they're saying. Determine the minimum number of tiles Matilda needs to place so that each row and each column of the grid is exactly one unit square that is not covered by any tile. Oh, so you would say if they all need to be covered, it'd be 2025 by 2025. But no, no. Each row and each column has exactly one unit square that is not covered by any tile. All right, Paul and Darren and Syphase, get to work. I expect your answers before the show is over. The answer. Oh, I won't tell you the answer. I know the answer. I won't tell you the answer. I figured it out. Problem 6 stumped DeepMind and OpenAI's models. It also stumped 569. Out of the 630 students, only 6 got all. All of the pro. All the problems correct.

Jeff Jarvis (139:21)

This gold medal that they were 100%.

Leo Laporte (139:24)

No, no, 35 is a gold medal. So you only had to get four out of seven. I think something like that. 72 contestants got gold. And of course, some of them actually beat the AI models. Not a lot.

Jeff Jarvis (139:40)

Also, they're featuring the ones who beat the AI.

Leo Laporte (139:43)

Yeah.

Jeff Jarvis (139:44)

Okay.

Paris Martineau (139:45)

Yes. These are the people that beat.

Leo Laporte (139:46)

Got it. Better than the AI.

Jeff Jarvis (139:48)

And how long do they have to do this torture?

Leo Laporte (139:51)

I think it was two hours. I can't remember. Yeah, something like that. Two and a half hours. Anyway, good going, Tiger.

Jeff Jarvis (140:00)

At all.

Leo Laporte (140:02)

At all.

Paris Martineau (140:05)

Oda be a mathlete named Tiger.

Jeff Jarvis (140:09)

Yeah.

Benito Gonzalez (140:11)

Who will one who will be a Wall street boy named Tiger?

Leo Laporte (140:14)

Simon Willison, who is a columnist we've referred to before, he writes a blog.

Jeff Jarvis (140:20)

A what? A what? Leo.

Leo Laporte (140:22)

A blog. A blog about AI he writes.

Jeff Jarvis (140:25)

He doesn't just write a post. Right, Leo, he wrote.

Leo Laporte (140:27)

Well, this. This blog. This blog was actually about his 20.

Benito Gonzalez (140:32)

No, it is a proper name. It's a weblog. It's a weblog.

Leo Laporte (140:35)

It's a web log. Oh, he even says it's a web log. Yeah, that's what you had at Tumblr was a web log because it was a log of websites that you link to. So he one shotted a Space Invaders game using a couple of new Chinese models from Z AI GLM 4.5 air on hugging face. It's a pretty big model, actually, although there's a quantized version of it. He gave it the simple prompt, write an HTML and JavaScript page Implementing space Invaders. It did it in one shot. Look at that. That's vibe coding, baby.

Paris Martineau (141:23)

Nice.

Jeff Jarvis (141:24)

And how does it know what Space Invaders was? How does it get that?

Leo Laporte (141:28)

That's a good question. It must have looked it up. I don't know. It knew because it was in its training.

Benito Gonzalez (141:34)

Come on. The interesting thing out there about Space Invaders, Jeff.

Leo Laporte (141:37)

Yeah, yeah, true so it ingested it. The interesting thing is he did it on a two and a half year old MacBook Pro M2, a 64 gig model. So the fact that he's running it locally.

Benito Gonzalez (141:48)

Oh, it's a local model. Oh, that's interesting.

Leo Laporte (141:49)

Local model. It wasn't hugging face, but it was a local model. Yeah. So he downloaded it in other words from hugging face. Very impressive.

Paris Martineau (142:02)

Very cool.

Leo Laporte (142:03)

Yeah. Yep, yep, yep. In one line. Now he then asked it to create a scalable vector graphic of a pelican riding a bicycle and got that. But.

Paris Martineau (142:20)

Bicycles are hard. What can you say?

Leo Laporte (142:25)

His Simon's bottom line is local coding models are really good now and I'm very interested in this because I've tried the local models and you know, actually Darren like or no, is it Dr. Do? No, Darren likes his local models. He uses Llama. But you know, I've had less than excellent results from local models. They're getting better. And what's interesting is these Chinese models are getting better.

Jeff Jarvis (142:52)

Enjoy Llama while you can because Zuckerberg's backing away from open source.

Leo Laporte (142:56)

Well, that was the rumor. Has he. Have we seen any confirmation story here?

Jeff Jarvis (143:00)

Well, his argument is, he says from PC Magic that within super intelligent AI now in sight, it won't be open source like Llama, which critics have already read.

Leo Laporte (143:18)

Okay.

Jeff Jarvis (143:19)

He just said it's more dangerous, ergo he's gonna be more careful.

Leo Laporte (143:22)

Yeah, I hope that they don't pull back on the local models. But you know what? The Chinese companies are not so they are.

Jeff Jarvis (143:31)

At least I think there's going to be competition.

Leo Laporte (143:33)

Yeah. AI is wrecking an already fragile job market for college graduates. Wall Street Journal. I think we knew that because the entry level jobs are going away and that's the problem.

Jeff Jarvis (143:46)

That's stupid humans who aren't building their future staff and think about it. Yeah, they really, really should. One of these studies. The studies in here were ridiculous. I put them in there. But.

Leo Laporte (143:56)

But anthropic did one. One. Yeah. 700 professions and here's Microsoft chats. 40 jobs about to be destroyed by AI according to Microsoft.

Paris Martineau (144:06)

The list is ridiculous.

Jeff Jarvis (144:07)

Ridiculous.

Leo Laporte (144:08)

Interpreters and translators. Yeah, that's probably true. Right.

Paris Martineau (144:10)

The number two on his historian.

Leo Laporte (144:12)

Historians.

Jeff Jarvis (144:13)

Historians are going to be put out of work.

Benito Gonzalez (144:15)

Historians.

Leo Laporte (144:15)

No, no, it's not. Number two, I don't know what the order is because there's only 3,000 historians.

Paris Martineau (144:21)

No, originally. So I'm looking at like, like a thing from a screenshot from the paper and it like historians are number two because they have Like a higher score, I believe.

Leo Laporte (144:31)

Oh. A higher percentage of them will lose their jobs.

Jeff Jarvis (144:34)

Okay, so passenger attendance, Flight attendants by AI. What? What?

Leo Laporte (144:38)

No sense. Sales representatives. Yes.

Benito Gonzalez (144:41)

Well, the people. The people at the desk, though, for, like, the flight attendants. The people who serve you at the desk before you get on the plane.

Leo Laporte (144:48)

Okay, but do you need safety people in the plane?

Benito Gonzalez (144:51)

On the plane people, but those are still the same people.

Leo Laporte (144:53)

Customer service representatives, for sure. Cnc. Tool programmers. Yeah. Telephone operators. Are there anymore there.

Paris Martineau (145:01)

Are they. This. Okay, this perhaps draws into question this whole thing. It says there's more telephone operators than Historians.

Leo Laporte (145:09)

Historians, Ticket agents.

Benito Gonzalez (145:10)

They died a long time ago, didn't they?

Jeff Jarvis (145:12)

Yeah, gone.

Leo Laporte (145:14)

Broadcast announcers and radio DJs.

Jeff Jarvis (145:18)

All your people.

Leo Laporte (145:19)

But wait a minute. I spent many years learning how to make my voice do this. I don't think an AI could do this. Can it? 56 degrees in the city on a beautiful Wednesday afternoon. I hope you're having a safe drive home. We'll have more from Neil diamond right after this. Brokerage. See, I don't think any. I could do that. Brokerage clerks, farm and home management educators. Telemarketers. Yeah, they can go. They can go. Concierges, maybe. I think somebody's going to still have to call the restaurant and say, hey, Joey.

Jeff Jarvis (145:52)

Well, actually, no. What happens is that the restaurant has bribed the concierge to send the poor tourists to the bad Italian restaurant.

Leo Laporte (145:58)

Right. They can't bribe an AI. Political scientists. No, I don't think. The academics are gone.

Jeff Jarvis (146:04)

This is stupid.

Leo Laporte (146:05)

News analysts, reporters, hostesses.

Jeff Jarvis (146:07)

What the hell is this thing? Hosts and hostesses?

Leo Laporte (146:11)

Yeah, you need somebody to greet you at the door, say how many in your party. You can't sit down until your entire party is here. That kind of thing. You need. You need people to do that, Give you the menu. You really?

Paris Martineau (146:24)

Editors and rental clerks?

Leo Laporte (146:27)

Yeah, demonstrators and product promoters. Who's that? Those are the people at the auto shows. The young ladies in skimpy outfits.

Jeff Jarvis (146:36)

Let me show you the wonder of the Benjamatic.

Leo Laporte (146:40)

You can't. AI can't do that.

Jeff Jarvis (146:42)

It slices anything.

Leo Laporte (146:44)

I think the desk job's fine, but the economic teachers are out of business. Archivists, web developers, science teachers, models. No. Well, maybe. Well, maybe you won't need Sydney Sweeney to show off her genes anymore, right?

Jeff Jarvis (147:02)

Switchboard operators. Where is the switchboard operator?

Leo Laporte (147:05)

There's 43,000 of them in the hotel down the street.

Jeff Jarvis (147:09)

Is this from. From, like. Thank you.

Paris Martineau (147:13)

Second part of those are probably MTA switchboard operators, which still exist.

Leo Laporte (147:18)

They're Running the trains is what you're saying.

Paris Martineau (147:20)

That's true.

Benito Gonzalez (147:21)

So none of these jobs are middle management? None of them.

Paris Martineau (147:24)

Yeah.

Leo Laporte (147:24)

Yeah. We're going to get rid of all the entry level people first.

Jeff Jarvis (147:28)

It's an absurd story. So we don't know what the impact. Yes, we need to be concerned about the impact on jobs, but so far we're not dealing with real data.

Leo Laporte (147:35)

Here's the good news. Your chances of getting a job at Chipotle are much improved because Chipotle's AI hiring tool is helping it find new workers 75% faster. Guess what it calls it.

Paris Martineau (147:50)

Yes.

Leo Laporte (147:50)

Guess what it calls it. Ava Kado. Ah, wait a minute.

Jeff Jarvis (148:02)

Starbucks. You two.

Leo Laporte (148:04)

Huh?

Jeff Jarvis (148:04)

You go to Starbucks?

Leo Laporte (148:06)

Not anymore. Not since I discovered what coffee's supposed to taste like.

Jeff Jarvis (148:09)

Oh, that's right.

Leo Laporte (148:10)

Stop.

Paris Martineau (148:10)

Yeah, I was gonna say if I have to get up really early to work and want to get like a hot food item before six, then most.

Jeff Jarvis (148:19)

Of them have stopped putting the little cardboard sleeve on.

Leo Laporte (148:22)

Good. They're respecting our environment.

Paris Martineau (148:24)

You get coffee from Starbucks in. It's also a heat wave, Jeff. It was 100 degrees out today.

Jeff Jarvis (148:29)

I do what I do Paris.

Paris Martineau (148:32)

And you don't drive over a bridge to get there.

Jeff Jarvis (148:34)

Right.

Leo Laporte (148:39)

All right, anything else we need to talk about before we take our last break? Sam Altman worries about the non zero chance the world will end from a lethal synthetic virus. Actually, that has nothing to do with that. Could happen at any time. Exactly. We are not prepared for the next Covid.

Jeff Jarvis (148:58)

There's a Guardian review of Elon Musk's diner.

Leo Laporte (149:02)

Yeah. Did you see the poor people who live nearby?

Jeff Jarvis (149:04)

Oh, no.

Leo Laporte (149:05)

They've got giant screens shining movies into their windows. There's apartment complexes right next to the giant diner. They interviewed one woman who moved months ago because the construction company put a giant light on the side of the building shining into their windows.

Paris Martineau (149:23)

Oh, boy.

Leo Laporte (149:23)

Oh, boy. Well, this is from the Guardian, so I think this is it. Elon Musk opened a diner in Hollywood. What could go wrong? I went to find out, writes Lois Beckett in Los Angeles. This actually isn't the one I was talking about, but there are quite a few stories. There's the robot that serves popcorn, the Optimus robot. There was a very funny TikTok video of it getting stalled and the guy had to come and press some buttons and reboot it.

Jeff Jarvis (149:58)

Where's the reboot?

Paris Martineau (149:58)

You know what, John? Go away. Robot rebooting.

Leo Laporte (150:03)

This is the giant movie screen. They project movies on two sides of the diner. And apparently the people in the Apartment buildings across the way are not happy about this. Not at all. I wouldn't be either. No. Oh, but look at that. Delicious. What is that, a waffle?

Jeff Jarvis (150:21)

I've heard chicken and waffles. Yeah, that's what the.

Leo Laporte (150:25)

That looks horrible.

Jeff Jarvis (150:26)

It does.

Leo Laporte (150:29)

Oh, Elon, if I lived at a.

Benito Gonzalez (150:31)

Pipeline building, there would be games playing. I would have hacked that screen a long time ago.

Leo Laporte (150:37)

Smart.

Jeff Jarvis (150:37)

Yep.

Leo Laporte (150:38)

Even the food comes in boxes shaped like cyber trucks.

Jeff Jarvis (150:43)

That's the. That's the worst of it all.

Leo Laporte (150:45)

Oh, my.

Paris Martineau (150:46)

I mean, that box is kind of cute. I'll give him.

Leo Laporte (150:48)

Really.

Paris Martineau (150:50)

I think everything else. I just. I. I like a novelty box. What can I say?

Leo Laporte (150:56)

Yeah. All right. What else? Anything else? There's going to be a new singing competition show on Tick Tock. You know, these shows are cheap to produce, like, like the Masked Singer. And what are those other ones?

Jeff Jarvis (151:20)

America's Got Talent.

Leo Laporte (151:21)

America's Got Talent. The Voice. So it's going to be on Tick Tock, which is kind of interesting. On TikTok Live, we produced one of these at Twitch.

Benito Gonzalez (151:31)

There was a Twitch. One of these with all our. With like, Twitch Celebrities. We did one either.

Leo Laporte (151:35)

Very cheap to do. Right. I mean, they're easy to do. It's called Nextup Live music competition show solely on Tik Tock live through September. People. I think people love these shows. The original. What is the original? Not the Apprentice.

Benito Gonzalez (151:52)

Especially if it's like Tik Tok Celebrities. Right? Like that.

Leo Laporte (151:55)

That's people you know already. Yeah. They can sing. I am not going to play Gwyneth Paltrow.

Jeff Jarvis (152:03)

No, but it was pretty. Was brilliant.

Leo Laporte (152:05)

I played it on Sunday. Good move, astronomer. Well, well played. Well played. All right, I think it is time because I see some wonderful picks ready to. Ready to go.

Jeff Jarvis (152:17)

But first, very proud of mine today. It's a very good one today.

Leo Laporte (152:20)

Nice. Very happy with it.

Paris Martineau (152:21)

Yes.

Leo Laporte (152:22)

I am very proud of our club. Without it, we would have to cut back. We would have to get rid of people, cancel shows, and I don't want to do that. The club now supports. We see the advertisers, you hear ads, which is great. We love our ads, but they only pay 75% of our operating expenses. That means there's a pretty big gap of 25%. Thank goodness. Lisa, four years ago, almost five now, said, let's do a club. Give people who listen to the shows a chance to support it. And your response has been phenomenal. I have to thank you from the bottom of my heart. And I'm sure that everybody who works at Twit thanks you. But now I'd like to make an appeal to those of you who are watching who are not yet members. If you like the content, the programming we do here, we need your support. But there's a good way to do that, and that's to join the club. It's only 10 bucks a month. You get excellent. You get access to the club Twit Discord, which is a fabulous hangout with some of the smartest, most interesting people, other club members. Right. You also get to watch shows that we don't put out in public. All the keynotes from now on are going to be done in the club. We have some events coming up, including the AI User Group on Friday. Anthony Nielsen leads that I join every time I can. We've asked Alex Lindsey if he'll come by and show us how he uses Midjourney to design images. Anthony's showed us that kind of thing and any of our coders or anybody who wants to join in, show us how you're using AI. It's a lot of fun. Home theater geeks. August 4th. That's this Monday. He's going to be doing the Value Shootout and name the king of TVs. We have a new king in town. Our book club is a week from Friday. Stacy's book club. Is this how you lose the Time where Stacy Higginbotham. We couldn't get her every week for the show, but we could get her every other month for the book club. She loves talking about books. I do too. And I really liked this book, so I highly recommend it. You still have time to read it. It's not very long, but it's very interesting. This is how youw Lose the Time War by Amal El Mohtar and Max Gladstone. Right after that, on the Same day at 2pm Pacific Photo Time with Chris Marquardt, we'll do our review of Classic and Of course Google's Pixel 9 announcement, or I'm sorry, Pixel 10 announcement made by Google is August 20th. We'll stream that. Micah and I will be doing that in the club. Micah's crafting corner. That following Wednesday, I'm going to be doing some sewing. I brought my, my, my wardrobe shirts, these shirts that I wear on the shows. I brought them to my cleaners. I said there's some buttons missing. Could you fix them? He said, yeah, sure. So I got them back. There's. They're still missing. So I realized I have to sew on my own buttons. So I have ordered a sewing kit. And then I saw this thing that looks like a button stapler from Singer. So I'm gonna, I'm gonna compare the two in Micah's crafting corner.

Jeff Jarvis (155:20)

A little bit of consumer reports right here. Right here on getting your butts.

Leo Laporte (155:24)

So I'm gonna take a rag, I'm gonna sew some buttons, and you can do whatever craft you want. That is what is fun about the club. On Friday, Richard Campbell came and did his PC build for four and a half hours. They talked about, I mean, it's great pros and cons and bourbon throughout. Yes, it was a super geek out session. Really fantastic. If you're not in the club, please, twit TV club, twit. We'd love to have you. We really would. Let's map out this week's amazing destinations and travel tips.

Paris Martineau (155:56)

Honestly, Will, I didn't plan any trips, but I did switch to T Mobile with their new family Freedom offer.

Leo Laporte (156:04)

That's not the itinerary we're following.

Paris Martineau (156:07)

Well, I'm departing from AT&T and embarking on a new journey with T Mobile. They paid off my family's four phones up to $3200 and gave us four new phones on the house.

Leo Laporte (156:20)

Bon voyage. Introducing Family Freedom. Our lowest cost. Will switch our biggest family savings all on America's largest 5G network. Visit your local T Mobile location or learn more@t mobile.com familyfreedom. Up to $800 per line via virtual prepaid card. Typically takes 15 days. Free phones via 24 monthly bill credits with finance agreement, eg Apple iPhone 16128 gigabyte $829.99 eligible trade in eg IPH for well qualified credits end and balance due if you pay off early or cancel Contact T Mobile. Whether you're into comedians roasting each other's life choices or turning yesterday's bad decisions into today's funny stories. Amazon Music's got the most ad free top podcasts included with Prime. Download the Amazon music app and get in on the joke or go to Amazon.com adfreecomedy that's Amazon.com adfreecomediDy to catch up on the latest episodes without the ads. Now it's time for our picks of the week from Ms. Paris Martineau.

Paris Martineau (157:22)

So, building off of last week, Jeff's pick involving crows. I've got a bird pic for you guys this week, which is some guy saved a PNG file to a bird.

Leo Laporte (157:34)

What? Birds can store data. Yeah.

Paris Martineau (157:39)

So this is a YouTuber and audio scientist, Ben Jordan, he. He basically figured out a way to save a. Save a PNG to a bird part of the bird. He converted a drawing, like, just basically did a simple, like, line drawing of a bird into a spectrogram, which, like, takes kind of like a png, puts it into a sound wave, then put that sound wave in birdsong, played it to a starling, then came back, visited the starling, randomly going through it.

Leo Laporte (158:16)

He's training it to sing.

Paris Martineau (158:18)

Didn't even train it, just played it to the starling once, just a PNG in bird song form. And he came back, recorded the bird's song, looked at the data, and there was the bird drawing he'd made. No, the bird somehow transmit the bird, transmitted data.

Leo Laporte (158:40)

Ben Jordan is actually a great YouTuber. People love him.

Paris Martineau (158:45)

What a. Jordan estimated that the bird effectively retained and retransmitted 176 kilobytes of data in audio format.

Leo Laporte (158:53)

That is amazing. All right, well, as long as we're talking YouTubers and birds, I got one that actually came from Club Twit in our Twit community, our forums@twit.community Mark Rober, you know, is the NASA scientist. He first is most famous for doing the glitter bombs for porch pirates, right? But he does some of the most elaborate, amazing videos ever. He believes crows are very smart. I agree. And he decided to make an escape room for a crow. He tried it out on some of his friendly neighbors, some kids he decided to go out and buy because the crow will watch, the crows will steal. When the doordash guy comes, the crows steal his nuggets. So he decided this is a good incentive for a crow. And so he built this amazing, beautiful escape room with puzzles, multiple puzzles. Had the kids try it. The kids were stymied by it. The crow was not.

Paris Martineau (160:07)

The crow figured it out.

Leo Laporte (160:09)

The crow escaped. Well, I don't want to spoil it for you, so I won't tell you what happened, but the crow. Well, I'll give you an example. This is an example. So. So this is. He has placed the nuggets under a cage that the crow has to solve a number of puzzles to raise the cage. One of the puzzles is this graduated cylinder that has water in it and a little bauble, okay? A wooden ball. A ball the crow happens to like, right? So she wants it, but she can't get it. It's too far down there. So can she figure out. Can you figure out how to get that ball to rise up in the graduated cylinder? There's corks she looked at the corks. Would that work? No, they float. There's cotton balls. Would that work? No, no, they, ah, she figured out, put the rocks in there until the water level rises enough so that she can get the ball. Incidentally, she's also rising, see on that little cylinder to the. On the side. No, the corks aren't going to work. Crow. No, she doesn't like the corks.

Jeff Jarvis (161:20)

She's mad. The corks.

Leo Laporte (161:21)

She. She says Dino, she should be. Yeah, she's. She took the rock and put it back in there. Crows are so smart. Watch, the ball's going higher now. Once she can get the ball, that other cylinder will trigger. There we go, the light. And it comes on over to the balances, which she immediately flies over to and. Well, I'll just let you figure out. He had trained the crow, of course, but what a cool. Is that. Brilliant. 37 million views in seven months.

Paris Martineau (161:56)

Wow.

Leo Laporte (161:56)

Mark Rober, who has 70 million subscribers. There's another bird video for you. You. You can. You could put a ping in a bird song or you can teach a crow to fetch some nugs.

Paris Martineau (162:10)

Someone in the Discord said effectively. They sent a tweet.

Leo Laporte (162:18)

That's a real tweet, Jeff, Pick of the week.

Jeff Jarvis (162:21)

All right, so we've got AI Slop and we have God Slop.

Leo Laporte (162:26)

Oh, no.

Jeff Jarvis (162:27)

So there was a UK Christianity magazine pointed to this. So if you go to that one, you're on 165. This is Jesus at the wedding. You turn the sound up.

Paris Martineau (162:37)

Oh, my God.

Leo Laporte (162:38)

Let's see what kind of trouble we're not getting into. Jesus as an influencer. Wait, wait, wait.

Jeff Jarvis (162:42)

Did you hear that line? Let's see what kind of trouble we're not getting into.

Leo Laporte (162:45)

Okay, if you are a religious person, you might find this somewhat sacrilegious. Just a word of warning. It's a got to this wedding and it looks amazing. Let's see what kind of trouble we're not getting into today. All right, quick vibes, check. Food's flowing, music's fire.

Benito Gonzalez (163:01)

My mom's American.

Leo Laporte (163:02)

Probably plotting something.

Paris Martineau (163:04)

Of course.

Leo Laporte (163:04)

10 weddings so far. Hi, guys.

Paris Martineau (163:07)

I'm JC's mom and I just told him the wine's gone totally out, so now it's officially his problem.

Leo Laporte (163:15)

All right, I'm gonna need these.

Paris Martineau (163:18)

Oh, my God.

Leo Laporte (163:20)

This is all the done on vo, I presume.

Ian Kreitzberg (163:22)

I presume something like it.

Leo Laporte (163:25)

There's water. He's putting water in the amphora. All right, water's in. Now we just.

Paris Martineau (163:30)

He's British now.

Leo Laporte (163:31)

Now he's suddenly British all over the accent. Good wedding. Fortunately, he's wearing underwear. Moment of truth. If this goes well, I'm a legend.

Ian Kreitzberg (163:40)

If not, I just ruined a bunch of water. It worked.

Leo Laporte (163:44)

Crisis averted, vibes restored. And let's be honest, if we'd ordered wine for delivery, it would have taken years, so. But this is a religious group that made this?

Jeff Jarvis (163:55)

No, it's just some. They're just. They're just. People are making them online and they're ridiculous.

Leo Laporte (163:59)

Holy vlogs. Vlog Z s. So there's another one with Noah. You want to see Noah land in the ark? Okay. Got this wild DIY project from the big guy upstairs. He said two of every kind.

Ian Kreitzberg (164:12)

So guess who's learning animal husbandry?

Leo Laporte (164:13)

Real fact. Okay. Day 27 of rain. The WI fi is spotty, the lions are not happy with the vegetarian meal plan, and someone keeps asking, are we there yet? This is from Historical Influencer. That's a real person. This is not AI Land. No pressure.

Jeff Jarvis (164:27)

Little.

Leo Laporte (164:27)

Well, maybe it is.

Paris Martineau (164:28)

No, that's definitely.

Leo Laporte (164:29)

How would you do that? He's back. And he's got proof. An olive branch. You guys know what this means? Hope is at a million. Let's go. I love it. AI really knows how to do influencers.

Ian Kreitzberg (164:41)

Counseling for extremely angry.

Benito Gonzalez (164:43)

Smell in here is officially probably what we have the most volume of.

Leo Laporte (164:46)

Landing. Yeah, of course. Boxing the entire animal kingdom. Like, comment and subscribe if you want to see the first rainbow.

Jeff Jarvis (164:56)

The Virgin Mary one ends with don't forget to like and pray.

Leo Laporte (165:00)

Oh. Oh, my.

Paris Martineau (165:04)

Oh, my angel literally showed up and said, I'm gonna have a baby from God. Like, what? Okay. I told Joseph. He froze for like, five minutes, then asked if I was sure. I said, absolutely. He's been pacing in the backyard ever since.

Leo Laporte (165:17)

So we're heading to Bethlehem. Nine months pregnant on a donkey.

Paris Martineau (165:23)

This is not how I pictured today going. There's a cow snoring, no bed, and it smells like donkey. Five stars for privacy, zero for comfort.

Leo Laporte (165:33)

Hey, guys.

Paris Martineau (165:34)

He's here. Born right here in this cozy barn. We're tired, but our hearts are full. Okay. Not even 24 hours old, getting gifts. Baby's first collab. We'll unbox later. Love you guys. Care.

Leo Laporte (165:48)

Don't forget to like and baby's first collab. We'll unbox later.

Benito Gonzalez (165:54)

Please forgive us, Padre.

Leo Laporte (165:57)

Smash that baptism button.

Jeff Jarvis (166:00)

I sent it to my minister sister.

Leo Laporte (166:02)

Her.

Jeff Jarvis (166:02)

She laughed.

Leo Laporte (166:03)

That is. That is quite. Well, she's a Protestant, so.

Jeff Jarvis (166:06)

Yes, that's right.

Leo Laporte (166:08)

That is. In fact, she's A Unitarian, which I don't even know.

Jeff Jarvis (166:10)

Presbyterian.

Leo Laporte (166:11)

Oh, Presbyterian. Okay. That's different. Wonderful. I love those. It's enough to make me put TikTok back on my phone. You saw, by the way, that the president now wants to ban TikTok.

Jeff Jarvis (166:25)

No, I didn't. Oh, geez.

Leo Laporte (166:27)

He says they're not making a deal. I gave him an opportunity. They don't want to make a deal. I'm sorry. Sorry. You have till. Whatever it is. September 1st. JD couldn't close the deal. So we're just going to shut him down? He's trying to.

Jeff Jarvis (166:45)

I fear that he's actually going to do that.

Leo Laporte (166:46)

He's using the tariffs, right. He's hoping the Chinese government will back down because of the tariffs. I don't think they're going to. I don't think they care that much.

Benito Gonzalez (166:55)

Maybe, you know, they just rescinded the minimum exception for tariffs, Right? Like.

Leo Laporte (166:59)

Yes. This is terror. Terrible. Yeah, yeah. The 800 exception. Sub. $800.

Paris Martineau (167:04)

Bad for your sheen purchase.

Leo Laporte (167:06)

But there is some good news. We think. We're not sure that alcohol will be exempt from the EU tariffs. So your wine will still be affordable.

Benito Gonzalez (167:14)

They didn't exempt coffee from the Brazil tariff.

Leo Laporte (167:18)

That's. No, that's. Now. Now you're. You're. You're hitting where it hurts now.

Jeff Jarvis (167:25)

Well, I'm not sure with his. His people.

Paris Martineau (167:28)

Well, I'm sure some. As someone who just paid $6 for coffee, it's gonna.

Leo Laporte (167:34)

Yeah, it'll be a little more than that. All right, we're done. Thank you so much. Paris Martineau, Congratulations on the new job. Do you know yet what your first expose is going to be?

Paris Martineau (167:43)

Oh, yes. I'm fast at work on us.

Leo Laporte (167:45)

Really? Oh. How hard is it? I see. I would find it very challenging to come up with a thing to investigate. How hard is that?

Paris Martineau (167:55)

It is. I mean, it's curiosity difficult if you're coming from nothing, but it is really. It's really just you find something that's interesting to you more about, and then you go from there. So you just kind of immerse yourself in a topic and then figure out what you get curious about.

Jeff Jarvis (168:12)

Paris told me that. That she had to come up with ideas to get the job, and they loved her ideas.

Leo Laporte (168:18)

Fantastic.

Paris Martineau (168:20)

I already had someone to work with. Yeah.

Jeff Jarvis (168:22)

So is. Is the story you're working on first one of those ideas?

Paris Martineau (168:25)

It is one of those ideas.

Leo Laporte (168:26)

Oh, love it.

Paris Martineau (168:27)

It actually turned out one of the ideas I'd pitched. They already had kind of some teams working on some aspect of like it relating to testing and whatnot. So it was kind of a perfect moment of synergy. So we're all working together.

Leo Laporte (168:43)

Congratulations. I'm so happy. I can't wait to read it. It as a subscriber, I will be the first. Well done paywall.

Paris Martineau (168:52)

But you can read it because you're a subscriber.

Leo Laporte (168:54)

I'm a subscriber. So your stuff is all going to be past the paywall because you're so expensive?

Paris Martineau (169:01)

No, I believe all of my stuff because I am part of the special projects.

Leo Laporte (169:05)

Oh, yeah?

Jeff Jarvis (169:05)

Yeah.

Paris Martineau (169:06)

Advocacy focused and impact focused.

Leo Laporte (169:08)

They do paywall.

Paris Martineau (169:09)

Paywalled.

Leo Laporte (169:10)

They do pay well. Very well. For instance, they'll show which cars they like, but they won't show you all the details. You have to be a subscriber to see the full details.

Jeff Jarvis (169:19)

They're very smart about it.

Leo Laporte (169:20)

This is very well done, I think. Jeff Jarvis, speaking of well done, he's got all these great books teased to a crisp. The Gutenberg parenthesis now in paperback, magazine now in audio form. And of course, the web. We weave his roadmap for the success and future of the web. Thank you, Jeff. We appreciate it. He is a professor of journalistic innovation emeritus at the Craig Naimark Graduate School of Journalism with the senior years for and currently at LOY at yeah.

Jeff Jarvis (169:51)

Yeah.

Leo Laporte (169:52)

And other ships Montclair among yeah. Amongst State University and Suny Stone.

Jeff Jarvis (169:59)

I wanted to terrorize Jabra B today because YouTube now says you can use bad language. Yes.

Leo Laporte (170:05)

In the first few minutes you can say the F word.

Jeff Jarvis (170:08)

Yeah.

Leo Laporte (170:08)

Why did they back down? Why did.

Paris Martineau (170:10)

First of all, the first few minutes.

Leo Laporte (170:13)

They said, you. If you do it in the first few minutes, we'll demonetize you. And then for some reason they've changed.

Jeff Jarvis (170:18)

Their mind because everybody's saying it now.

Leo Laporte (170:22)

Stop that. Sorry I had to bleep you.

Benito Gonzalez (170:25)

I still gotta do the work though.

Leo Laporte (170:26)

You know.

Paris Martineau (170:27)

You do. I thought you.

Leo Laporte (170:29)

You know what I. Instead of bleeping what I like, Benito, is take the word and reverse that portion of it so it goes get. I love that. I love it when that happens.

Jeff Jarvis (170:46)

And then play it backwards and it will say, paul is dead.

Paris Martineau (170:49)

I was going to say you can play it backwards and suddenly you're in the red room of Twin Peaks.

Benito Gonzalez (170:54)

My favorite way is to do it like the old TV way. They used to censor movies on TV in the 80s 80s.

Leo Laporte (170:59)

You put fudge.

Benito Gonzalez (171:00)

Change the words.

Leo Laporte (171:01)

Yeah, yeah. They come in and they do full. They do I'll kill your mother father.

Paris Martineau (171:06)

You can put. You can Put this word in where I said the F word.

Leo Laporte (171:10)

Frick.

Paris Martineau (171:11)

You could do that.

Leo Laporte (171:12)

You could just shut the front door. Yeah, yeah, I like. There's actually. There's some really good euphemistic ways to do it. There's, of course, the Good place has some good euphemisms. Battlestar Galactica, they would use. No, that was Serenity Firefly, they would use Chinese swear words. And in Battlestar Galactica, they said frack a lot.

Benito Gonzalez (171:36)

And there were ethnic slurs in the Expanse that they made up.

Leo Laporte (171:40)

Oh, they made them up though, because. Yeah, they have different races. Thank you, Paris. Thank you, Jeff. Thank you, Benito Gonzalez, our esteemed producer and technical director. Thank you, Anthony Nielsen, who does a lot of our booking. Yeah. Including Ian. Our wonderful guest, Ian Kreitzberg. Subscribe to his newsletter at Puck News. Thanks to all of you. Next week on the show we will have the CEO of my favorite search engine. No, not Sundar Pichai. No, not Satya Nadella. It's going to be Vlad Prelovac, who is the CEO of Kagi. It's so good. I pay for it. It. And it is the future of search. So we will be. That'll be very interesting. They use a lot of AI as well. Thank you for being here. Thanks especially to our club members. We do intelligent machines every Wednesday right after the Whiskey Windows Weekly's Whiskey segment. That's 2pm Pacific, 5pm Eastern, 2100 UTC. You can watch live in the club. Twit, Discord if you're a club member. Otherwise on YouTube, tick tock, twitch, x dot com, Facebook, LinkedIn and Kik. Did I get them all? You're counting. Okay. Am I missing one?

Paris Martineau (172:56)

Probably. Or maybe I just didn't count right. I'm not a mathlete. I'm not. I'm not.

Leo Laporte (173:01)

No gold medal for you. That's right after the fact. On demand versions of the show at our website, Twitter, TV iam. You'll also find a link there to the YouTube channel. Great place to clip parts of the show. If you want to show, share it with friends and family. Do please, because that'll help us. And of course you could subscribe in your favorite podcast client and get it automatically. You don't have to think about it audio or video. Paris says if you leave us an interesting five star review, she will read it on the air. Apparently there were none last week. Or she was.

Paris Martineau (173:29)

There was. I was busy. I can read one right here.

Leo Laporte (173:33)

Okay.

Paris Martineau (173:34)

Paul VC says I love Twit and as a member I follow most the shows. But this Windows Weekly and Mac Break are my favorite.

Leo Laporte (173:42)

Favorite.

Paris Martineau (173:43)

Paris. It's nearly autumn.

Leo Laporte (173:50)

He loves Paris's amazing spook sound. Thank you, Paul. VC Two thumbs up for five star. We appreciate that. We appreciate all of you. Thank you for watching and listening. We'll see you next week on Intelligent Machines. One more time, Paris. I'm not a human being. Not into this animal scene.

Paris Martineau (174:14)

I'm an intelligent machine.