Please be advised this transcript is AI-generated and may not be word for word. Time codes refer to the approximate times in the ad-supported version of the show.

00:00 - Leo Laporte (Host)
It's time for Twit. This Week in Tech, mike Elgin's here, brian McCullough and Richard Campbell. We're going to have a lot of fun talking about AI, a big jury award for a victim of a Tesla crash, more than $240 million, apple results, google results and Australia banning YouTube for people under 16. How's that going to go? All of that and more coming up next on TWIT Podcasts you love. From people you trust.

00:34 - Richard Campbell (Guest)
This is. Twit.

00:41 - Leo Laporte (Host)
This is TWIT this Week in Tech, episode 1043, recorded Sunday, august 3rd 2025. It all starts with Baby Shark. It's time for TWIT this Week in Tech, the show. We cover the latest tech news with a rotating panel of fine, industrious, smart and connected journalists, starting with Mike Elgin in oaxaca.

01:06 - Leo Laporte (Host)
Two out of three ain't bad which one aren't you, I didn't add good looking so we'll add you.

01:13 - Leo Laporte (Host)
You put you in the good looking box. Uh, his newsletters at machine societyai.

01:18 - Mike Elgan (Guest)
He does podcasts, he does everything I do, dishes I do, I do it all. Yes.

01:24 - Leo Laporte (Host)
An excellent pizza cook too, I might add. Thank you for being here, mike.

01:29 - Mike Elgan (Guest)
It's nice to see you, glad to be here.

01:30 - Leo Laporte (Host)
Thank you, a beautiful evening in Oaxaca, I'm sure. Yes, it is Also joining us from. Well, we could have changed the name. It used to be the Tech Meme Ride Home. Brian McCullough is here, the internet's historian now the tech brew ride home. Is it a beer show?

01:54 - Brian McCullough (Guest)
Yeah, that was the original pitch. Somehow the name got through all of the various editorial meetings and things like that. No, it's the exact same show, literally the only thing that even the music is the same. The only thing that is changing is at the beginning of every show, instead of saying tech meme, ride home, I'm saying tech brew, ride home.

02:07 - Leo Laporte (Host)
Now part of the morning brew crew, which is fantastic. Congratulations, thank you. That is a excellent network of shows, growing apparently, so that's good. Also with us. One of our own, he is Richard Campbell. He doesn't do the beer, he does the whiskey own. He is richard campbell. He doesn't do the beer, he does the whiskey. Host of run as radio and dot net rocks and, of course, regular every wednesday on windows, weekly on this network, hi, richard hi, just back off a boat story at the back at the end of this, if you like okay, we can add some whiskey.

02:40
you know, that's the one thing this show's been missing all this time. You were on a cruise over the week. In fact we missed you on Windows Weekly. That's why I said hey, but you know, maybe we could get him off the boat and into the show on Sunday and you very kindly left early, I think.

02:56 - Richard Campbell (Guest)
I just took the early exit. We had options, but you know, after a week on the boat you're ready to get off anyway. It was an Alaska cruise, yeah, this time out of Vancouver, which is nice. We did Sitka and Juneau and Ketchikan, and we also stopped at a new stop called Icy Strait, which is the local Tlingit band in Hoonah, which is 800 people. They've built out quite a little facility there now Did you get to see the Mendenhall Glacier or any.

03:25
We didn't. I could see it from a distance from juno um, they have warnings on that glacier today, a potential glacial outbursts which you know sound that doesn't sound good we were supposed to go up the tracy fjord but it was all fogged in and the glaciers are calving and, let's face it, ice and cruise ships not good. I think there's a there's a history about that.

03:46 - Leo Laporte (Host)
Made a whole movie about that sort of yeah although I wouldn't mind meeting kate winslet, but perhaps not under those circumstances well it's good to have you back on dry land and on the show today big story in fact.

03:59
I was having a dinner with samabool samad, our car guy, who was in town, uh driving some hyundais around uh, when this story broke, a jury has ordered tesla to pay more than 240 million dollars in an autopilot crash. Now I'm sure tesla will appeal this, but this is in Florida. They didn't find the Tesla. Interestingly, 100 responsible uh, a driver was using full self-driving in a Tesla, so a couple years back he dropped his phone a couple years back, it's well back six years back.

04:39
Yeah, uh, this case been going on for four years. He dropped his phone distracted, reached down to get it, thought he was cool because full self-driving was engaged. The car continued and ran into another car and hit and killed a pedestrian, hit the car, the car hit the, a young couple who are stargazing, and one of them was killed. Uh, so the lawsuit came not from the driver. That's one thing that makes this unusual in tesla lawsuits usually it's the driver of the driver's family if the driver's been killed in the accident. In this case, it was the family of the pedestrians who were killed. Um, the driver was found partly responsible because he trusted full self-driving. Uh, the lawyers for the family of the deceased claimed that tesla either hid or lost key evidence, including data and video recorded seconds before the accident. Tesla said we made a mistake. After being shown the evidence, they said no, honest, your honor.

05:52
We thought it was gone we tried really hard we thought we deleted I mean we thought it was the sister of the deceased says we finally learned what happened that night the car was actually defective. Tesla said, quote this verdict is wrong. It only works to set back automotive safety. I'm sure, right, you get there wrong and jeopardize tesla's and the entire industry's effort to develop and implement life-saving technology. They said the plaintiffs concocted a story blaming the car when the driver from day one admitted and accepted responsibility.

06:30 - Mike Elgan (Guest)
Tesla has to pay 43 million of a total 129 million compensatory damages for the crash, plus a punitive award of 200 million dollars interestingly, they're going after the branding of this and um pointed out that you know other car companies don't use things like autopilot or you know any sort of implication that it, in fact, is designed, you know, to be full self-driving, um, and so that that's part of the blame. You know the idea that he was misled into thinking the car would handle it just based on what they call the feature the, the feature of full self-driving.

07:14 - Leo Laporte (Host)
Uh, this isn't. I mean, I think, uh, was it australia that says you can't call it full self-driving, right, right, um, so there's always been this issue. It isn't full self-driving, it's driver assist, no, but it's called autopilot, though right yeah, uh, and so literally you, you call.

07:36 - Richard Campbell (Guest)
You use the term fsd for full self-driving, like they have been misnaming that product all along.

07:41 - Leo Laporte (Host)
Yeah, and using the word autopilot, which is not um. Tesla was partly at fault. It allowed the driver to act recklessly by not disengaging the autopilot as soon as the driver began to show signs of distraction. So that's why they were partially responsible. The driver said I trusted the technology too much. I believe that if the car saw something in front of it, it would provide a warning and apply the brakes. No, it just plowed right into the part. It was a parked vehicle.

08:11 - Mike Elgan (Guest)
It hit by the right which spun around, and then the spinning parked vehicles. What hit the victims?

08:18 - Richard Campbell (Guest)
yeah, uh, let's face it, any other um cruise control would have stopped. Yeah, my toyota would have stopped. Would have stopped, right like.

08:30 - Leo Laporte (Host)
So tesla says look, we tell drivers you got to keep your eyes on the road and your hands on the wheel. We even check for that, and if you remove your hands from the wheel, we go put your hands back. Uh, and they said the driver chose not to do that when he looked for the dropped cell phone and added to the danger because he was speeding. Tesla said look, he's gone through the same intersection 30 or 40 times previously and hadn't crashed not once, until the final crash. He said the cause is not tesla, but the driver dropped his cell phone and stopped paying attention. Uh, the juror. This is now.

09:08
I asked sam about this, uh, because the key in this is that the jury found some not all, not 100, but some liability for tesla, despite the driver saying yeah, I, I, I wasn't looking. And he said that sets a precedent. And in fact, uh, that's what this uh npr story says, it it, despite the driver's admission of reckless behavior, poses significant legal risks for every company's developed cars that increasingly uh drive themselves yeah, this is I was going to say.

09:43 - Brian McCullough (Guest)
There's a, there's a quote in the piece that says from a lawyer that says this will open the floodgates. Um, and there's earlier in the piece suggestions that they're surprised that it came to trial because a lot of other cases haven't. Like I thought that's what tesla usually does right, and the reason they can do that is because the people who are suing are usually the drivers of the car.

10:02 - Leo Laporte (Host)
So tesla says no, no, let's settle. But in this case it wasn't. It was a innocent pedestrian victim not driving the car, and so apparently they weren't anxious to settle. So tesla was not able to shut this down before discovery, which turned out not so good for tesla and the verdict.

10:21 - Mike Elgan (Guest)
So a precedent and another interesting thing is uh, as I think we mentioned, this was six years ago, this accident, right, so so the uh, the, the autopilot feature has gotten way better. We presume. Right, where are we at with that? Is it? Is it still? I mean, elon musk had been promising full self-driving mode for years and years, always saying, oh yeah, next year we'll have it. Um, are they getting close to full self-driving mode? I, I have no idea sam uh.

10:52 - Leo Laporte (Host)
Again, my car guy, samuel samud uh, has told me he does not believe cars will ever get to full autonomy because while they can navigate 90 maybe you know soon 95, 99 of the situations, they can never be ready to handle every single weird, unexpected incident that happens but they just have to be safer than people, right, because people, that that's.

11:18 - Mike Elgan (Guest)
I think they already are for people I think they already are.

11:21 - Leo Laporte (Host)
I mean, don't tell the pedestrian who has killed that, but well, yeah, that happens all the time, doesn't it? I hate to say it, but that happens all the time with cars driven by humans the dumb part with tesla is they're insisting on using only cameras.

11:34 - Richard Campbell (Guest)
Exactly no radar. Right, like just a forward-looking radar that would have stopped this. Like it's just that elon gets these hang-ups and nobody can stop him's like we can do this with cameras and he's killing people because of it.

11:46 - Mike Elgan (Guest)
I'm sure we've all been in a Waymo in San Francisco.

11:50 - Leo Laporte (Host)
It's amazing, I mean they do use radar and cameras Exactly.

11:55 - Mike Elgan (Guest)
They have a big rig on the top, you know they look weird because of it. Yeah. Yeah, they look weird and and. Yeah, they look weird and and and, but it works. It works really well. I'm I'm guessing that it's significantly safer than than human drivers.

12:09 - Benito (Announcement)
Hi this is benito. So I was driving down the city in san francisco yesterday, and I was so for a while. I would have to take this route through the city, where I would always be behind a waymo, and it was like the worst thing ever if you're downtown, every third car is a waymo the first time I saw one, I said, oh, look a waymo.

12:25
And then, oh, there's another, oh there's another, they're everywhere but then I all over then yesterday while I was in the city I noticed driving around way most. They drive a lot better now like they.

12:34 - Leo Laporte (Host)
They seem more more like people, like I'm not as pissed at them on the road anymore as I used to be, so I think that's an improvement well, I mean, maybe that's kind of what Tesla was saying is yeah, you know, you're gonna, we're gonna have these judgments, but it slows down development because we got to get these things on the road so that they can get better. They can't get better in a test tube, so that they have I don't know how many people to die.

13:00 - Richard Campbell (Guest)
You know well people.

13:01 - Leo Laporte (Host)
Okay, 50 000 people died in auto accidents. Uh, typically every year in the United States, a million globally. That's not even newsworthy.

13:12 - Benito (Announcement)
It happens so often right, but people are liable for those, for those accidents they're like this is the issue, or kind of driving safer that's.

13:24 - Leo Laporte (Host)
I mean, that's really the ultimate issues. Can we make people get to sign?

13:27 - Richard Campbell (Guest)
up for this. They're being done, and this has been happening to them against their will we're in a beta test that we never agreed to?

13:32 - Mike Elgan (Guest)
but plus I don't really follow the argument. How is this slowing anything down?

13:37 - Richard Campbell (Guest)
well, you can collect the data regardless yeah yeah, but if okay, I guess.

13:44 - Leo Laporte (Host)
But if you're liable for every accident you cause, then maybe you're going to be uh, less likely to. That's what they're.

13:50 - Richard Campbell (Guest)
That's I think that's what you're likely to be more careful yeah, maybe, okay, maybe, maybe that's, I mean, you know, 200 million dollars, or whatever it is.

13:58 - Mike Elgan (Guest)
Three, three, three, four hundred million dollars, it's nothing for tes.

14:02 - Leo Laporte (Host)
Tesla is going all in on the robo taxi. Here's a story from Ford. Forget, I'm sorry, forbes, forget cars. Tesla's future is a robo taxi Empire. Uh, he doesn't want to sell cars. Apparently Elon wants to sell robo taxis, but if they can't drive. Right now these robo taxis are still unlimited tests in just a handful of cities, although I guess they're coming to San Francisco kind of using a loophole with DMV, because DMV is saying you can't test in San Francisco, california Department of Motor Vehicles, but because they have a safety driver, somehow they're getting away with it safety driver.

14:49 - Brian McCullough (Guest)
Somehow they're getting away with it, um, but the, the promise that waymo waymo, specifically always used to make was once we can prove this out. It's sort of like the restaurant franchise model. We can just copy the playbook city by city by city, and and they, they seemingly are like atlanta, was it? D Dallas was the most recent one announced. So, yeah, I mean like putting the stupid history hat on for a second. If you go back and look historically at when automobiles first hit the scene, especially in the United States, the amount of carnage and I'm talking running over pedestrians, running over horses, running, um, like you're. Someone said that we're running a beta test on new technology that none of us signed up for, but that had that happened a hundred years ago.

15:33 - Leo Laporte (Host)
That's always what happens, yeah.

15:35 - Brian McCullough (Guest)
Um, now, part of the mitigating point of that is well, there were no laws, like you could run over a pedestrian, and there was no law on the books that said, hey, by the way, you could be liable for that.

15:47 - Leo Laporte (Host)
well, in fact and you probably know this internet historian the whole concept of jaywalking was created by auto manufacturers who wanted to make sure that pedestrians were held liable if they crossed the street at somewhere besides a crosswalk.

16:03 - Mike Elgan (Guest)
There was no law against jaywalking until the auto industry pushed it if you look at those old-timey videos that people remaster and stuff on youtube and you look at a city street before automobiles were huge, people used to go every which way all the time because everything was moving pretty slow and people were cutting in front of cable car type things and horses and stuff like that.

16:28 - Leo Laporte (Host)
It's kind of like the roundabout in the Place de la Concorde in Paris.

16:31 - Richard Campbell (Guest)
Exactly. It's just insane People got killed by horses too.

16:35 - Mike Elgan (Guest)
Exactly, but over the decades the concept of the road belonging to the cars took hold, but it took decades for that psychological reality.

16:46 - Brian McCullough (Guest)
Jaywalking laws were part of that, but nowadays you basically, yeah, you'll cross, you'll jaywalk or whatever rarely, but back in the day that was just how you got around town, you just go across the streets Going even further back and correct me if I'm wrong chat room but the first fatality from a railroad running over somebody, a train running over somebody, was literally because trains were so new, moving at a speed that was faster than any humans had any context.

17:18
For the reason the person got run over is because they couldn't imagine that the train would reach them as quickly as it did and didn't get out of the way.

17:26
But I'm making that point to say that one of the things that I find compelling about the idea of again to use the term running a beta test on self-driving is this is I've just given two examples of new technology that was run out in the open as a beta test. Nobody signed up for it but could get run over by it. This is a beta test that is being run with the explicit idea that this is going to be safer on the aggregate, on average, and so, as opposed to those previous times in history, again, if the amount of deaths that happened for the miles driven on American roads versus miles flown on airlines like airlines would be, no one would fly, you know. So, like we have someone made the point, we accept Leo made the point. We accept the hundreds of thousands of deaths and maimings and things like that by the car being central to our society, and so the point of this technology is to improve that sort of horrific statistic well yep, early flying was very dangerous.

18:32 - Richard Campbell (Guest)
That's true also only for the wealthy right.

18:35 - Leo Laporte (Host)
So it's only if rich people died, if we could just work on that, that that wouldn't be so bad. I could, no, that's so mean. Uh, anyway it's. You see, this is why it's complicated. It's a complicated.

18:48 - Richard Campbell (Guest)
I'm still surprised that they didn't settle out like there's a, I'm sure they tried. I'm sure they tried, and I also wonder if elon was just too distracted like look at the timing of this maybe.

18:59 - Leo Laporte (Host)
Yeah, good point. Uh, apple's uh biggest revenue growth since December 2021. This is not typically a big quarter for Apple. This is not quite the worst quarter. Next quarter will be the slowest quarter but it's far enough from the release of a new iPhone that it's not typically a big quarter but it is a big quarter for Apple. Sales of the iPhone grew grew 13 year over year. Overall revenue grew 10 percent, its largest quarterly revenue growth since uh 2021. Of course, some of that is services. The service revenue might be in jeopardy depending on what uh the courts do with google's 20 billion a year payment to Apple for for making a Google search the default on the iPhone. Uh, apple shares, of course, went up. Apple, I think, is now four trillion dollar company.

19:57 - Richard Campbell (Guest)
I haven't checked uh, lately the services bump was tiny compared.

20:01 - Leo Laporte (Host)
It was mostly iPhone sales which is interesting because it we're in off and off quarter five. No, but what was?

20:06 - Mike Elgan (Guest)
happening is all the talk about tariffs and the tariff bump raising the prices. People were getting one while they were still.

20:12 - Richard Campbell (Guest)
That makes sense prices and also or a Mac or an iPad, like all of them, took unusual bumps and I think it was just hey, if you were going to buy one this year, you should buy it now well apple did say the tariffs in the june quarter cost it 800 million dollars.

20:29 - Leo Laporte (Host)
That was a little lower than the 900 million they were saying they might have to pay but they're expecting more next quarter yeah, 1.1 billion, uh, next quarter.

20:38
Interesting that they're going to absorb some of those costs. Uh, this is what the president wants companies to do. He doesn't want it to, doesn't want them to pass the tariff costs along to customers. But there's only a limited amount of profit any company's willing to give up, and Apple, even with its massive margins I'm sure, still passes along more than half of the cost of tariffs just to focus on the negative for a second.

21:02 - Mike Elgan (Guest)
Do we have any sense of why iPad revenue is down 8%?

21:07 - Leo Laporte (Host)
The analysts say, because they didn't have an iPad Pro to sell this quarter. And you know, typically this is a slow quarter because people are aware that in September Apple will announce new iPhones and possibly in October new iPads. So it is typically a fairly slow quarter, as next quarter is as well. I think you nailed it, though there was a. The expectation of tariffs brought people to the store. Yeah, um cook said about one of the company's 10 percentage points of revenue growth could be attributed to that.

21:38 - Mike Elgan (Guest)
Yes, that's 10 of 10 percent, um it's rare when, when politics um you know, political conversation, of course the idea was that there were going to be material um you know uh tariffs, but usually affects gun sales, not phone sales. Right, every time there's an election, uh, one side always says the other side is going to take away all your guns.

22:02 - Leo Laporte (Host)
People go buy a gun, uh, before it gets taken away or whatever and so, as a result, we have a per capita average of four or five guns although it's only 20 that own a gun right like it's

22:13
really right the people who own them own a lot of them because they keep buying one every election um. Iphone revenue was strong, said Tim Cook, because the his thought is because the iphone 16 is more popular than the 15 was. Iphone 16 sales up strong double digits versus the predecessors. Mac business grew fastest of any of apple's units 15. Why do you think that is, mr pc?

22:40 - Richard Campbell (Guest)
richard I think, I think definitely was the tariff pump. But let's face it, that Mac is a gorgeous machine. It's the best computer on the planet right now. Nothing even comes close. The MacBook Air is its best selling Mac, yeah. So if you're digging for a reason to buy a machine, tariffs is a good excuse, whether they're real or not. And that's the best computer on the planet, why wouldn't you buy it?

23:00 - Brian McCullough (Guest)
By the way, leo Apple is not even close to $4 trillion. It's down at $3 trillion. It's down significantly for a year to date.

23:09 - Leo Laporte (Host)
Oh, I'm sorry.

23:11 - Brian McCullough (Guest)
Maybe it was.

23:11 - Leo Laporte (Host)
NVIDIA Somebody went to Microsoft.

23:13 - Brian McCullough (Guest)
No, it was Microsoft that went to $4 trillion.

23:15 - Leo Laporte (Host)
Microsoft is beating Apple Yep by a trillion. Wow, I apologize. Yeah, that's right, microsoft did the 44t. Uh apple services business was up 13 percent 27.42 billion. Uh cook said its growth in the iCloud subscriptions and app store revenue, despite the fact that the app store is under assault in the eu. Um, and wearables didn't do so hot, I'm not sure why that is. Apple's wearable unit declined 8.64 percent um.

23:52 - Mike Elgan (Guest)
China slightly up, which is important to Apple, um is, is is iPad, as, I'm sorry, is the Vision Pro considered a wearable? Yeah, I think the Vision Pro is a wearable. They probably had an artificial bump the previous year, which they didn't have this year.

24:10 - Leo Laporte (Host)
Yeah, because that came out last year. Yeah, yeah, maybe that's it yeah.

24:14 - Richard Campbell (Guest)
They stopped talking about the return numbers on the Pro, on the Vision Pro too.

24:18 - Leo Laporte (Host)
so oh, that's a bad sign. Yeah, we knew there were a lot, at least in previous uh reports. Revenue 94 billion dollars uh, it's a. It was a. Now, what was? Microsoft had a good quarter too, didn't they? And google had a spectacular quarter. So, uh, despite the fact that the stock market has been taking a hit in the tech sector, uh, these companies are doing okay. Iphone no longer have more than more than half of apple's revenue, surprisingly in.

24:52 - Brian McCullough (Guest)
In both the cases of google and microsoft, it's because it's the ai stuff, it's their cloud.

24:58 - Leo Laporte (Host)
Yeah, it's all about ai, isn't it?

25:00 - Brian McCullough (Guest)
yeah, I mean basically basically killing it in terms of like it actually, weirdly, amazon was the one that AWS was not reporting as good numbers as Microsoft and Google did for their cloud AI stuff. But you know the thing that Tim Cook said on the earnings call that'll get them back to a four trillion, or get them to four trillion dollars would be. He said, um on ai. Apple is quote very open to m a. That accelerates our roadmap vis-a-vis ai. Um, so I mean, look, if they come out tomorrow and say, hey, we just acquired perplexity, hey, we just acquired god forbid anthropic, which they could do, they've got the cash yeah, and anthropic's raising right now, so this would be the time to go ahead.

25:48
I don't think anthropics for sale do you? I don't? Complexity for sure, complexity for sure, and and actually don't they? Doesn't apple have a deeper relationship with perplexity at the?

25:59 - Leo Laporte (Host)
moment. So eddie q said in the testimony in the epic trial that in fact they were looking at acquiring perplexity. They do have a relationship with perplexity and it would make sense because what perplexity primarily is and I love it, by the way, I use it is a search tool and that's what apple would really like to have as a kind of an ai search tool.

26:17 - Richard Campbell (Guest)
I think everybody google and their principal interface is the phone like it's the best fit you could find yeah it's a natural, isn't it?

26:24 - Mike Elgan (Guest)
yeah, they do things, uh, in hooking into the phone that siri doesn't even do in terms of agency and stuff like that through perplexity. That's a great app this is.

26:36 - Richard Campbell (Guest)
This is just a number. This is what is the number the investors will accept right from apple to make this happen.

26:41 - Leo Laporte (Host)
It's probably too high according to uh, mark German, who every Sunday writes his rumor newsletter, the power on newsletter for Bloomberg, uh, apple has a new team called the answers team. This is where John gandrea apparently went. He was the big acquisition from Google who turned out not to be, uh, such a star at Apple and has been kind of pushed aside. He's still doing research, so this is a new team. The Answers team that will Answers, knowledge and Information, or short is AKI Gurman says. This group, I'm told, is exploring a number of in-house ai services with the goal of creating a new chat gpt like search experience, and this is the right thing for apple to do to get the perplexity price down right.

27:38 - Richard Campbell (Guest)
Like you show, you've got x many months to make this deal before. I have a competitor and I don't need you anymore.

27:48 - Leo Laporte (Host)
Number of people from the formerly of the Siri team. The Siri is no longer a candidate to become a chat GPT like anything, including Robbie Walker. He's now the senior director of the AKI team, reporting to John G Andrea. In fact, I think a number of the Siri people have ended up on this AKI team. You're right. I might guess it's more like a skunkworks where they're doing something, maybe with the eye towards bringing the price down on perplexity. That makes sense.

28:20 - Brian McCullough (Guest)
All the other reporting is about how, I mean, we know that Zuck is hiring everybody he can in AI, even offering a billion and a half dollars to-. Yeah, we'll get to that story? Yes, no kidding the?

28:32
it's the. What is it? The foundation model team at Apple that apparently has lost so many people in recent times, and the reporting from the information that I did this week on the show was sort of that. The talent problem is not just that people would leave Apple, it's that the people that are still, since no one really knows what Apple's strategy is. That's kind of why they should just pull the trigger and buy somebody. Like, if you know this is the direction we're going in, then you could probably retain talent, even if you don't, you know, increase their salaries and things like that. If you just tell them, ok, this is where we're going, this is what we're going to do, and people internally don't have that sense yet, apparently on Friday.

29:15 - Richard Campbell (Guest)
people want to make something and if you don't give them a path to make something, they're going to go somewhere. They can make something.

29:20 - Leo Laporte (Host)
Well, they're trying to keep people in the fold. Information reported that on Friday, tim Cook, at an all-hands meeting, told employees Apple has exciting plans I can't discuss With his own people.

29:36 - Brian McCullough (Guest)
You need to read in some of those people really quickly, because $300 million is very enticing.

29:43 - Mike Elgan (Guest)
Yeah, all of this has to be quite disturbing to Apple. I mean, I'm sorry, to Google, because Google has always had a privileged relationship with Apple and iPhones, and if Apple is maybe buying perplexity, maybe reengineering Siri to become a search engine and so on, they're going to have no use.

30:02 - Brian McCullough (Guest)
But this is also the time because they might lose the contract like you know, build your own.

30:08 - Richard Campbell (Guest)
Uh search, now's the time yeah, because the government won't let you do that anymore, right?

30:14 - Leo Laporte (Host)
right uh. This is, by the way, the first time they've had an alts hands, apparently in years, according to the information, probably because of leaks, uh, which is why tim's not completely candid with his own people. He tells the staff ai is quote hours to grab it's hours to grab hours o-u-r-s, not h-o-u-r-s. Uh, important uh distinction.

30:43
Uh, now you're trying to grab a train that's whizzing by he says the ai revolution is as big or bigger as the internet, smartphones, cloud computing and apps. I think we'd all agree that. He says, quote and this is again a leak from the all hands bloomberg got this one. Apple must do this, apple will do this. This is very Churchillian. This is sort of oh, he just blew it.

31:06 - Mike Elgan (Guest)
Sort of we will grab them on the beaches. We will grab them on the beaches.

31:11 - Leo Laporte (Host)
This is sort of See, I don't think Churchill would have thrown the sort of in Ours to grab. We will make the investment to do it. So it'll be interesting. We've rarely been first. He told staffers there was a PC before the Mac, there was a smartphone before the iPhone, there were many tablets before the iPad, there was an MP3 player before the iPod. He's basically saying, yeah, we all always have the winning horse, at least not the first horse over the finish line, but we'll win in the long run. He says we invented the modern versions of these products. And this is how I feel about ai. This is a pep talk. This is not a. This is not a planning meeting. This is him trying to keep everybody happy before they go to work for zuckerberg but also if, if, uh, if.

31:57 - Mike Elgan (Guest)
I know apple and of course we all know apple very well and how they think about things. To date, people think of ai chatbots. It's okay, here's a chatbot and here's a search engine. We're going to use the chatbot like it's a search engine or we're going to use a rag system like perplexity. That's kind of both the search engine and a chap apples, you know, with their name answers, right. They're going to say they're not going to have this thing.

32:20
That's like, okay, here's a chatbot, here's to have this thing. That's like, okay, here's a chat bot, here's a personal assistant. It's, it's basically going to be whatever you know all the above, but it's just going to be a person answering your questions and hopefully they'll make it personalized, right, keep all the personal data and they're on the phone and so on. And this falls really neatly in with the glasses they've been working on forever and probably won't see till the end of next year. So this is really the killer app, and I think they understand that the killer app is going to be a personal assistant with omniscience and that knows all about you, that lives on your glasses and that.

32:58 - Leo Laporte (Host)
Let me ask you guys this, because I think part of the reason for this all hands is that I feel like maybe and I want to know what you guys think we are on the cusp of a big breakthrough in AI, like I don't know if it's chat, gpt-5, I don't you know, I don't know if it's it's Google, but it feels like we are really. This is the Eve, almost the Eve of, dare I say, at agi, is that no?

33:27 - Richard Campbell (Guest)
you disagree. I think we're hitting the bottom of the trough of disillusionment, is what you're feeling? Is that we're actually we're actually as bad as it can get. Is that what you're saying? Well, we're sort of getting to the point where it's like okay, we're see the limits now, these things are good at some things and they're not good at other things.

33:43 - Leo Laporte (Host)
And get real, but don't you I mean, look at people who are doing vibe coding at this point are saying, hey, I'm not going to have a job in senior software engineers, I'm not going to have a job in a couple of years. This stuff is so good uh. Google is clearly terrified it's going to lose its hegemony in search. That's why they added these kind of not so good ai, uh, whatever it is garbage to its search results. I think there's a lot of fear in silicon valley that chat, gpt or anthropic or somebody is, or maybe the chinese, because we're seeing some really strong models from china right now, small and interestingly open weight models from china right now that are in some respects, beating the best models we have, the best closed source models we have. We're seeing both chat, gpt and uh gemini, you know, score gold in the international math olympiad. We're seeing better and better benchmarks all the time it.

34:43
You think we're at the bottom of the trough? Yeah, I feel like we're. Just. This could be the crest of the wave, where maybe that's the same thing. I don't know. I I disagree, brian.

34:54 - Brian McCullough (Guest)
no, I agree with richard in the sense that I mean, again, this is all definitions of buzzwords and stuff, so we can debate these things. But if it's agi, that's like words and stuff, so we can debate these things. But if it's AGI, that's like gonna cure cancer and give us nuclear fusion and maybe kill us all versus. Well, if it's the bottom of the trough of disillusionment, you don't necessarily. I'm not one of those people that thinks that AGI is right around the corner, but it could be sort of like a 1996 period when people are like you know, there's this thing called broadband and three or four years down the road, imagine the things you can do on the web. So the bottom of the trough of disillusionment is the beginning of reality traction in actual mainstream ways. It's not just people dialing into AOL, it is people saying oh, you know, you can run entire businesses on the web. You can do like look at all the things that we can do with telecommunications. It's actually the rubber hitting the road, even without GPT-5 becoming self-aware.

35:59 - Richard Campbell (Guest)
Yeah, I don't know if GPT-5 is going to be all that. And let's face it, the rate of improvement is slowing right like they're not getting absolutely.

36:05 - Leo Laporte (Host)
I feel like we're also seeing some hints of self-improving ai, I'm hearing that phrase as well which would be a hockey stick if that happens, mike what do? You cover this in machine in your machine intelligence newsletter. What do you say?

36:20 - Mike Elgan (Guest)
I mean I? I think we're on the brink of a cultural transformation that results from the ubiquitous application of AI.

36:29 - Leo Laporte (Host)
More like what Brian was talking about. That happened with the internet, where it's becoming ubiquitous now.

36:33 - Mike Elgan (Guest)
Another analogy would be electricity. Right, so you had electrical systems being developed in the 1830s and 40s, and for a long time it was the telegraph, and then, eventually, cities started getting lights that didn't have, you know, dangerous gas lines going under the city to power the lights. And you know, by the time the revolution was happening, it was something of a banality. The revolution was a toaster and uh, you know, uh, an alarm clock that you know that you didn't have to wind okay, like that.

37:08 - Benito (Announcement)
But if you were, uh, if you were a woman in the appalachias or in the texas hill country.

37:15 - Leo Laporte (Host)
If you were a woman in the texas hill country, the appalachias, who was going three or four times a day to drag gallons of water back to the house for your family or hand washing your clothes in a stream, electricity was a massive revolution.

37:28 - Mike Elgan (Guest)
It changed your life and your future well, and then I think that's going to be much more than a lamp, right? But I mean, if you, if you think about the potential for what ai can do for an entrepreneur, right, a one, a one or two person business, right, right, it's, that is a real, like a major thing. They can do marketing, they can do they. They can do marketing in a hundred languages, they can. You know, there's so many applications for AI for small businesses that that they are the Appalachian housewives of our era, right, because it's their whole lives are going to be.

38:00 - Leo Laporte (Host)
They're going to be able to do things that previously couldn't even have imagined well, put me down, for I think we're about to see a phase change and a sudden shift, a paradigm shift, something hockey stick. Like you guys all think. It's going to be incremental and and it's going to become more ubiquitous. I feel like we're on the edge of a big, unpredictable phase change. That's just my. It's just a feeling probably.

38:31
I mean if, if we do, it'll probably have something to do with agentic ai sort of operating and self-improving ai, yeah, and, and and models and tools beyond large language models, um, I think that their people working very hard, and in many cases not publicly, on new developments that could very well change everything, kind of like the discovery of have been moving so fast just in the last six months have been like around video, where now we're used to oh, you can create five seconds of really impressive video around anything and then you stitch it together and you can make 30 seconds and then three minutes and then like, think, like things, like video.

39:17 - Brian McCullough (Guest)
I feel like in 18 months it's over, like you can do well yeah, look what vl3 can do is that not?

39:24 - Leo Laporte (Host)
almost a phase change. I mean that's. I think we're getting a little complacent, because it was a couple of years ago. It everybody had seven fingers and Will Smith's spaghetti was all over the place and in just a couple of years, massive improvements my feeling is we're actually about to shift even in into a whole new gear.

39:46 - Brian McCullough (Guest)
But maybe that's I have a question. You know is skeptical.

39:50 - Benito (Announcement)
I have a question about what are you calling? What are you calling it? Vibe of my futurism, vibe futures, vibe futures. Uh, my question is do you think vo is going to be free forever, because how much do you think that's actually going to cost when they're going to actually start charging people for that? Do you think it's really gonna to be free forever, because how much do you think that's actually going to cost when they're going to actually start charging people for that? Do you think it's really gonna like be free forever?

40:08 - Leo Laporte (Host)
well, I think another thing we don't know. First of all, the real cost, as we know, of ai is in training, not in use.

40:16 - Brian McCullough (Guest)
Okay not anymore, not anymore. Inference is now the. It has become the most expensive part of that.

40:22 - Richard Campbell (Guest)
yeah, really Interesting, that's a big shift and the best programmers I know utilize these tools. The guys that are doing Are running out of credits. They're doing a six-week sprint, in three days, like they're doing tremendous levels of productivity and they are spending thousands a month and you know what it's worth it? Yes, also, not enough, right? They're still getting. Even with as much as they're spending, they're still getting cut off because it's still costing the company more than what the company is charging them.

40:51 - Benito (Announcement)
See this right there. This is the problem. I don't think this is never going to be free forever.

40:55 - Leo Laporte (Host)
So how is this going to actually make economic sense? First of all, I don't think you could predict this, because there's a huge financial incentive for these companies to come up with more efficient ways to do this, and I think the Chinese companies actually have found small models, have found ways to do this with less.

41:12 - Richard Campbell (Guest)
To be clear, the Chinese have never done anything efficiently. They don't actually.

41:16 - Leo Laporte (Host)
They don't need to.

41:18 - Richard Campbell (Guest)
There's no incentive for them, but also.

41:20 - Brian McCullough (Guest)
Benito, it can be free for can be free for a very long time If the people offering them for free believe that this is a generational land grab where I need to train the filmmakers, the coders, the whomever. That's the case. Use my tools. They'll be free for as long as it takes before they have to. Once they have their moat, then then they'll charge you.

41:41 - Leo Laporte (Host)
Uber rides were way below cost for almost a decade. You know, sometimes to acquire a market and the upside, look, maybe this is all going to crash and burn, but the upside, if, if agi, if asi, if really capable ais come down. The road is so big that a lot of vcs, a lot of companies, look how much money so it's the drug dealer model.

42:06
Your first first hits free well, you're such a skeptic it's good I surround myself with skeptics because I I have a fee, I've got a feeling, feeling I can't hide you know what benito here's?

42:20 - Brian McCullough (Guest)
here's the best test is the mom test for things like well, my mom, my mom, will never have an iPhone or a smartphone. And then one day your mom calls you from the iPhone or my accountant will never, you know, bill me from the web or have a website. And then you know, now your accountant does everything virtually Like. So the day that your mom has a Claude subscription, a chat GPT subscription of twenty $50 a month, that's when you know that the jig is up.

42:47 - Leo Laporte (Host)
Oh, I don't think that'll happen and I don't think that that's. I think that's an old way of looking at things Mass, mass adoption. I don't know if this is a mass adoption product at all, I think the answers.

42:57 - Brian McCullough (Guest)
$20 a month for answers? Oh yeah, I mean yeah.

43:02 - Leo Laporte (Host)
I pay 25 bucks a month for for a better Google search for Kagi.

43:06 - Richard Campbell (Guest)
Exactly the interesting is that the Internet free model has been unraveling for a while.

43:11 - Leo Laporte (Host)
That's right, this is an amplifier of that.

43:14
Yeah, all right, we have to take a break because this Internet is not free. Your attention is all you have to pay. It's great to have you guys talking me off the ledge. Mike Elgin is machine Societyai talks about this every single day. It's fantastic. And you do a show with Emily forlini we love, uh. And of course, he's also in Oaxaca getting ready for another gastro nomad Adventure. This is true. We'll talk about that in just a little bit. Great, have you Brian McCullough with a new home? He's part of the morning brew crew at the Tech Brew Ride Home. Now you don't have to change your subscription if you subscribe to the tech, subscribe to the tech?

43:54 - Brian McCullough (Guest)
Absolutely not. All that happened was the art changed. The name that I say at the beginning of this show changes. The music's the same, the advertiser should be getting better, but yeah, it's just the same thing. I've done for almost eight years now, over 2,100 episodes, wow, wow, just tech brew ride home instead of tech meme ride home. And, by the way, love tech meme, love Gabe, love all the people there, just potentially a new audience for me to reach.

44:23 - Leo Laporte (Host)
That's it now you like this guy named axel and I ain't talking axel rose. I'll tell you that right now it's great to great to hear this brian. You deserve the success. Also our dear friend richard campbell, who was just off the boat and staring.

44:38 - Richard Campbell (Guest)
Episode 1000 a run as in the face too.

44:41 - Leo Laporte (Host)
Hey, that's exciting, yeah well, you know, podcasting, I think, is going to be big pretty soon. So one of these days we're in the face too. Hey, that's exciting. Yeah, well, you know, podcasting, I think, is going to be big pretty soon. So one of these days we're in the right position here. Actually, that's one of our, one of our stories today. We'll talk about the new kinds of podcasting. But first a word from our sponsor.

44:56
It's my little thinks canary. See this little doohickey there it's sitting on my desk. It's about the size of a external usb drive. It's connected. In fact, if you saw this, you'd say, well, you know, that just looks like you got something connected to the internet. There. It's got a ethernet connection and power, but oh no, to you it may look like a little box, but to the bad guy that looks like a windows share. Oh, you know what I'm going to make this one SharePoint. Let's invite the bad guys in.

45:29
This is the interface to the Thinks Canary. This allows you to take this it's a honeypot and make it anything you want. Right now, windows Server 2019 Office File Share. But if I want to configure it, I can change it to almost anything I want, including let's make it something really, really vulnerable MS SharePoint 2010. Now, what's good about this? It has a Microsoft Mac address. Everything changes. Actually, this is a Hewlett-Packard Mac address because it's Hewlett-Packard hardware, but it's not. It's not, it just looks that way. To the bad guy it looks like something delicious, something valuable, something not vulnerable, not a honeypot, but something they might want to attack.

46:16
You can also create something called Canary tokens little trip wires. You think Canary will phone home if anyone opens any of these files and, by the way, they can be anything from an API key to a credit card, to an Azure login certificate. They could be a spreadsheet, a Word document, but the minute a bad guy see so I have, I can tell you because you're friends, you're not bad guys. I have a few Excel documents spread around, including on my Google Drive and my OneDrive, that say things like employee payroll information. No hacker could resist that. Why do you need these? Because, frankly, you can have all the security in the world, all the perimeter defenses, but how do you know if a bad guy gets into your network? These things, canariesaries, are honeypots. They can be deployed, you saw in minutes, but if someone is accessing your lure files or brute forcing that fake SharePoint server.

47:17
Your Thinks Canary will immediately tell you you have a problem. There are no false alerts. If you get an alert from your Thinks Canary, there's somebody in your network, a malicious insider is probing your reports, or you got a bad guy who got through your perimeter defenses. Here's the bad news. On average, it takes companies 91 days to know they've been breached. Three months for a bad guy to wander around in your network. Look at everything. Look for places to place ransomware, time bombs to exfiltrate emails or customer records. How embarrassing is that.

47:53
These Things Canaries are so easy. You choose a profile. You saw how easy that was. Then you register it with a hosted console for monitoring and notifications, and the notifications can be any way you want. Your Things Canary can send you a text message email. It can use syslog it. You webhook slack. There's an api. You can even create your own interfaces. You can have them all. Then you just sit back and relax.

48:16
Attackers who breached your network or malicious insiders any other adversaries cannot help but make themselves known by accessing your thinks canary or those things canary lure files. I love this thing. Big banks might have dozens of them. You should probably have one for every segment for sure. If you. I'll give you an example visit canarytools slash twit.

48:38
Canarytools slash twit. Get five of them for 7 dollars a year. That includes your own hosted console, your upgrades, your support, your maintenance, all the alerts in any way you want. And if you use the code TWIT in the how did you hear about us box, you're going to get 10% off for life. But if you say, you know I want to try before I buy you, can they have a very generous two-month money-back guarantee for a full refund 60 days. I have to tell you, though, we've been doing these ads for eight years and all that time that refund has never been claimed. Nobody's ever asked for the money back, because once you get a Thinks Canary, you go God, how do I live without it? Visit canarytools slash twit. Enter the code twit in the how did you hear about us box. Canarytools slash twit. Don't forget that offer code twit. You need this thing. You really do. Thank you Thinks for long-term support.

49:36
By the way, uh, the doj is still going after apple. Not one more apple story. Uh, they sued apple last year saying you illegally monopolize the smartphone market. Apple says what do you what apple? Apple says this threatens. Uh, it's a dangerous precedent and allowing the government to take a heavy hand in designing people's technology. The filing this week uh by apple responded to every point the doj made in its complaint. The doj says apple stifles the success of super apps or apps that offer multiple services on one platform, maybe like Epic or Microsoft wanting to do a game store, which Apple won't let them do. The company says our rules allow and support such apps. Is that true, richard?

50:27 - Richard Campbell (Guest)
I don't think so I don't think so.

50:33 - Leo Laporte (Host)
They say a multitude are available right now in the app store. I don't think not. Not the microsoft xbox game store, not the steam store, not the epic store. Apple also denied the doj's allegation that it blocks cloud streaming games, says oh no, we allow game streaming over the web and in the app store. Anyway. I'm not going to go through all the points. Apple is simply not a monopolist. Apple said it also. Uh, it says we don't. How can you even say we have a monopoly? The only way the doj could do this is by creating a separate category from smartphones called performance smartphones, which apple says does not correspond to economic reality. I think they've got a good point there in order in order to make a monopoly.

51:18
How you go after apple honestly yeah, it's like you run out of good ideas here, working your way through the bad ideas the lawsuit heads to the point that I'm sure apple would like it or prefer it not to, which is the discovery phase. Uh, it is in discovery that we'll learned so much about how Apple works behind the scenes in the Apple Epic lawsuit. If ever Apple could get out of that, I'm sure they would. But here that, here it comes, and that'll be good for us. We'll have stuff.

51:49 - Brian McCullough (Guest)
I was going to say that's only good for us.

51:51 - Leo Laporte (Host)
Gives us things to say yeah, we learned so much stuff last time. All right, we talked about apple's earnings. Alphabet also had a good quarter.

52:02 - Richard Campbell (Guest)
Actually, alphabet had an exceptional quarter um, when you can't attribute to tariffs no, that's right.

52:09 - Leo Laporte (Host)
Their service, their services company, their overall revenue 14 growth which beat wall street's, uh guess by quite a bit. Revenue 96.4 billion dollars. Advertising revenue was 9.8 billion a big jump. Cloud revenue up to 13.6 billion, very, very successful. What was the profit? It was something 28.2 billion 28. That's just for 12 weeks. Yeah, that's more than $2 billion a week in profit After tax Profit.

52:54 - Mike Elgan (Guest)
They made $10 billion in YouTube advertising. Uh advertising up 13 percent.

53:01 - Leo Laporte (Host)
That's amazing yeah, but YouTube is really, I think what's safe to say at this point has become the dominant way people consume video.

53:09 - Brian McCullough (Guest)
Yes, over TV, over cable, over everything I mean the all of the numbers bear that out. Like you have a story on the on the um spreadsheet here about like, oh, this is how young people uh watch video now or present, but that that doesn't even matter. Like they beat. Remember when, when net netflix used to say our real competition is not hbo, it's video games, it's your real life, it's your spouse. Like, youtube is just killing everybody in that metric of like, if you're gonna spend time like it's now YouTube and basically TikTok, you know Like in terms of time spent not doing something functional, just sitting back and consuming from the financial times.

53:58 - Leo Laporte (Host)
Younger people now turn on youtube in the uk above any channel on their television sets. This is the uk's media watchdog reporting. Not even uh youtube. One in five of those born after 2010 gen alpha turned to youtube first on their TV. Us old folks people over 55, doubled their time on the service.

54:20 - Brian McCullough (Guest)
That's their growing demo Old people.

54:22 - Leo Laporte (Host)
Oh, 100% yeah, because we didn't watch YouTube, because we didn't have YouTube.

54:28 - Brian McCullough (Guest)
Last time my parents were in our house, I realized my dad was a full-on YouTube convert. No kidding, I realized my dad was a full-on YouTube convert For the reason that he became an eBay convert 20 years ago, because he loves 1970s era stereos. And there are multiple creators on YouTube with millions of followers that have channels devoted just to 1970s era audio equipment.

54:50 - Leo Laporte (Host)
That's the real secret of YouTube. The way they beat the networks is in niche programming, I mean they are able to touch everybody's particular interest remember when the complaint with cable was consuming?

55:03 - Richard Campbell (Guest)
you know 500, 500 channels on try 5 billion.

55:07 - Leo Laporte (Host)
Yeah, yeah, you know. Uh, across all age groups in the UK now the second most watched TV service behind the bbc, ahead of itv. Uh, part of the reason ofcom, the british uh regulator, is looking at this is because they, of course, want to write, they want to regulate it. Uh, australia has now and this has got to be controversial they have a teen social media ban. Youtube was exempted no longer In Australia. If you are under what is it 16? You can't watch YouTube.

55:47 - Richard Campbell (Guest)
How's that going to work? Good luck with that.

55:51 - Brian McCullough (Guest)
Well, they also YouTube and Google generally. They're're using ai now to, even if you lie about your age, um, they're going to try to catch you and um change the content you could view yeah.

56:06 - Leo Laporte (Host)
According to the australian uh government, the internet regulator uh, their survey found 37 percent of minors reported harmful content on youtube is. Is youtube a threat to young people?

56:18 - Brian McCullough (Guest)
oh, hell yeah listen with a, an 11 year old and a nine year old. A thousand percent. What is so what?

56:26 - Leo Laporte (Host)
is it they're not just watching coco melon?

56:30 - Brian McCullough (Guest)
listen, man, they might start there.

56:32 - Leo Laporte (Host)
You know how it works yeah, is it the recommendation engine? That's a problem.

56:36 - Brian McCullough (Guest)
Yeah, and even there's the YouTube Kids app Filter. Yeah, but I think I've told this story on this pod before. But one day I'm walking through the room and I can't remember what they were into Bluey or whatever it was at the time, into, you know, um, bluey or whatever it was at the time and the bluey family's in a car going through a mcdonald's drive-thru and the drive-thru is on fire and there's gunshots and, oh god, yeah, you mentioned that, that's right, yeah like that's ai slop right well, but this was like two or three years ago yeah like they've run out of bluey.

57:10
Yeah, no, but that's the point it all starts with baby shark and it just goes downhill from there well, or you know, they're all into watching people play minecraft or whatever video game play along stuff and then you get the well, the next influencer, the next influencer that is cursing, and then the next one that is, you know, like, yeah, there's no way youtube used to demonetize you.

57:32 - Leo Laporte (Host)
If you had f words within the first few minutes, they backed down even on that.

57:38 - Richard Campbell (Guest)
Maybe they figure we're just going to. What's happening is the ad revenues come back like the bigger thing with youtube? Is the adpocalypse is finally over, is it?

57:45 - Leo Laporte (Host)
yeah, obviously that number you told me, uh yeah, proves.

57:48 - Richard Campbell (Guest)
A few years ago there was this there was a couple of events where the most influential podcast, most influential youtubers, did inappropriate things for views and the ads ran screaming because their their customers freaked out and uh, and ads changed utterly for a while and a lot of youtubers dropped out. The stress went way up, the incomes went way down, and that's winding down. The youtubers have changed, the formats have changed, the number of products is is diversified, and the new filtering tools work in a different way, and so the ad revenue is flowing back in, and so YouTube's reducing the constraints and encouraging more things, because they need to place more ads.

58:30 - Leo Laporte (Host)
Jay, who is ironically watching us on YouTube right now. Jay, who is ironically watching us on YouTube right now, says as a parent, I can definitely say YouTube shorts are 100% not regulated or monitored and that's how kids get around the age restrictions on videos. So that's an issue Prime Minister Anthony Albanese said I'm calling time on it. I want Australian parents to know. We have their backs. Youtube responds, by the way, the ban will take effect in December. Youtube responds by saying we're watched by nearly three quarters of Australians in the age 13 to 15 bracket 75%. We should not be classified as social media because our main activity is hosting videos. They were exempt as a video hosting company, not a social.

59:22 - Richard Campbell (Guest)
So make them liable for the content they show. If you're a video hoster, you know television stations can't show this stuff, so give them the same rules. If that's what you're going to be, then follow the rules. You've been using social media as the cover to not have to follow the rules. Now you're calling yourself broadcaster youtube says.

59:40 - Leo Laporte (Host)
Our position remains clear. Youtube is a video sharing platform with a library of free, high quality content increasingly viewed on tv screens. It's not social media. It is true that a lot of the content is consumed on the, in the living room, on the big screen TV yeah, I love the idea of the being regulated like a broadcaster. Let's see how they like that well, the reason broadcast, at least in the US I don't know what it is in Australia.

01:00:04 - Mike Elgan (Guest)
The reason broadcasters are regulated in the US is because they use the public's airwaves yeah, but the internet has historically never been regulated for content, and that's this is where, in the, our problem lies, right the elephant in the living room, of course, is tiktok, which uh, I assume is is uh treated as a social network and that's really the alternative. You mentioned that. What did you say? Three quarters of australian youths are, uh are using youtube, but the other quarter are probably on tiktok, uh, primarily. So you know, tiktok is uh vastly uh worse for kids, because it's just such crack cocaine, you know, just go swiping through it's, yeah, it's crack for all of us on youtube at least you can go and you can learn how to fix the sync.

01:00:48
And people use it. You know they search for, for content they're looking for, whereas on tiktok they don't do that so much, right?

01:00:55 - Leo Laporte (Host)
so I think the main use of tiktok is just let the algorithm control my brain for the next two hours, um, whereas youtube is only like that for some people some of the time uh, the law passed uh last year only requires quote reasonable steps by social media platforms to keep out Australians younger than 16 or face a fine of 40 49.5 million Australian dollars, which Google could easily pay, given they're making two billion dollars a week.

01:01:28 - Richard Campbell (Guest)
Yeah, a per is it per person and it's per offense, so that could be per person that could that could add up.

01:01:35 - Leo Laporte (Host)
Yeah, the problem is, of course, we don't have a good way to do age verification. Google says it's working on a way to do it just by looking at your head now there was a whole bunch of other signals that they were.

01:01:51 - Brian McCullough (Guest)
Again they're throwing it to AI to say, okay, sure, you gave us a birth date, but then, like what you're watching, like what else your account is doing, we're going to but it might be sufficient to satisfy this australian rule, which is well, we're, we're doing what we can.

01:02:07 - Leo Laporte (Host)
Youtube said it's beginning to roll out age estimation technology in the us to identify teen users.

01:02:15 - Mike Elgan (Guest)
Um, let me see if this is they're also offering up the ability to verify your age showing our.

01:02:23 - Leo Laporte (Host)
So it says if the new system incorrectly identifies you as under 18, you'll be given the option to verify your age with a credit card, government id or selfie. So they're not asking kids to verify no, they're only asking people to say they're finding a way to get.

01:02:38 - Richard Campbell (Guest)
Regulators should not require them to check everybody's id right, which should be better than the mechanism to say we're going to do it automatically. It doesn't work, then we'll check their id. Yeah, that way we don't check everybody's id, which is what they really don't want to do. This is expensive so how?

01:02:53 - Leo Laporte (Host)
so? It says in this TechCrunch article YouTube isn't sharing specifics about the signal it's using to infer a user's age. Notes that it will look at some data like YouTube activity and the longevity of a user's account. Yeah, if your account's more than 20 years old, you're definitely over it Probably good. The new system. Well, this is interesting. The new system will only apply to signed in users. Oh, I didn't know this. Signed out users already can't access age-restricted content. Brian, do you have any insight into how they're going to do this?

01:03:26 - Brian McCullough (Guest)
no, because I mean, isn't there's? It's, it's been like a wave around the world of this age-restricted stuff. You know, in Britain the porn stuff and the you gotta use VPNs and state by state in the United States or whatever. It is weird or not weird, but I guess it's just sort of the accumulation of time. But the fact that all around the world this sort of hey, let's go ahead and start age verifying and keeping the kids off of the socials or at least throwing sort of soft padded walls around the kids on the socials, it's really happened in the last six months.

01:04:09 - Leo Laporte (Host)
But of course nobody's better than a 16-year-old or a 14-year old at figuring out how to get around these restrictions. In fact, according to ours, technica vpn use soars in the uk after their age verification laws go into effect. Proton says our signups in the uk increased 1400 percent. Uh does. Is that sufficient to use a vpn? I guess you could say well, I'm not coming in from the uk, I'm coming in from rome or somewhere right, you pick a country that has the laws of you, that you want to be governed by.

01:04:46 - Mike Elgan (Guest)
Um. Italy actually requires parental consent for anyone, so rome would be bad.

01:04:50 - Richard Campbell (Guest)
Okay, yeah, rome would be bad, but uh you better make an english-speaking languages, even an ai could figure out. Hey, you're only watching. Oh, good point, you know right, it'll be one of those things like how tuvalu or whatever that, uh, that island nation is that has made all the money, yeah yeah, so they'll.

01:05:07 - Brian McCullough (Guest)
They'll become the vpn center of the world because they'll be like that.

01:05:10 - Leo Laporte (Host)
I'm watching from barbados man. Yeah, I'm. I'm a 21 year old barbaden. Uh. Winscribe. Vpn posted a screenshot on x climbing to show up a spike in new subscribers. Ad guard vpn claimed they've seen a two and a half times increase in install rates since the law went into effect. A week ago, friday, uh proton said, just a few minutes after the online safety act went into effect. A week ago, friday, uh proton said, just a few minutes after the online safety act went into effect, we are signed up.

01:05:39 - Richard Campbell (Guest)
Surged by 1400 when did they put age restrictions on vpns?

01:05:45 - Leo Laporte (Host)
well, that's what china? Well, no, china just bans them outright, don't they? As does russia?

01:05:52 - Mike Elgan (Guest)
nord says uh, there's been a 1000 increase in purchases, so yeah but it is an interesting problem there's a global uh demand for age verification and, um, you know you can do, you can do all kinds of things, but ultimately, you know, 12 year olds and 11 year olds and 10 and 10-year-olds don't have credit cards, driver's licenses or ID or anything like that. They have library cards maybe. So it's a difficult problem, but I think ultimately it's going to be AI. We're talking about how AI is just going to be everywhere. I think this is one of the things that AI is going to have to be a big part of. Just that'll look at you and decide how old you are ironically, given how much of an accelerationist I am.

01:06:40 - Leo Laporte (Host)
Based on my earlier comments, these are exactly the things ai should not be allowed to do, right? Oh well, yeah, let's take a little break. More to come. You're watching this week in tech with mike elgin, brian mccullough and, of course, our great friend richard campbell. If you haven't seen his three and a half hour pc build from from last week, you got to be in the club to see this stuff. By the way, we uh we thank our club members. We stream as all these uh club events live on eight different platforms youtube, twitch, tiktok, xcom, linkedin, facebook and kick. In fact, we're streaming right now on on those platforms. But after the uh, the pc build ended, then we put it into the private twit plus feed for club members. If you're not a club member, you can get it just by going to twittv slash club twit uh. And then, a month after that, 30 days hence, it goes public again on our twit youtube channel for all the 13 year olds to watch. So, um, thank you for doing that. Richard did it. Is it working?

01:07:45 - Richard Campbell (Guest)
oh, yeah, the machine's fine. It'll replace this machine. Actually, nice, so it's. It looked like a pretty nice pc you're building great. It was a great machine in the in that fractal north case is gorgeous, just a beautiful work of art. Is it around somewhere? Yeah, it's. Yeah, I'll dig it out after the night for the next time. Does it light up? Not really, no, I'm not that guy, right, I like that. I like the wood trim. I like, oh, I like wood. It has no external bays, because that's not a thing anymore. I'm not putting a DVD drive in it.

01:08:14 - Leo Laporte (Host)
Oh, it's all, it's a monolith, a wooden monolith.

01:08:18 - Richard Campbell (Guest)
It's a fan, it's all fans, and cooling to keep it quiet. Yeah, nice yeah very nice.

01:08:24 - Leo Laporte (Host)
Thank you for doing it, even though we had to lift the motherboard out at one point and put it back in again.

01:08:28 - Richard Campbell (Guest)
Yeah, how did the pasting go? Oh, the pasting was fine. That's the. That's an easy part, and I had a. I got the kit from, uh, from arctic, so I had a little board to. Oh, that makes it easier wipe it on and template. We had a great discussion in the in the chat about what's the right pasting technique. It's like not too thick, smooth it out.

01:08:46 - Leo Laporte (Host)
Did you baptize it with a little drop of uh, some jameson or something?

01:08:51 - Richard Campbell (Guest)
no no, you know, I I mean I'm not saying I'm not above threatening a machine with a squirt gun, like if you don't behave things are gonna get hard for you, but uh, I love it. I, I, I think I think they called me on the channel at one point I'm the bob ross of pc builds, because I just don't.

01:09:06 - Leo Laporte (Host)
There's only happy little accidents here, right like it was also calming and beautiful, just like micah's crafting corner. Some of the stuff we do in the club is very peaceful, some of it not yeah, I think I hit the style that the crowd, the crowd expects. So yeah, no, everybody loved it. I'm sorry I missed it. I got called away for yeah we.

01:09:24 - Richard Campbell (Guest)
It went, I think about an hour or so and I I said to anthony he's like kind of glad we didn't wait for leo on this one well, I told you.

01:09:31 - Leo Laporte (Host)
I told him not wait, not to wait, go right ahead and I would get there as soon as I can. Well, I told him not to wait, go right ahead and I would get there as soon as I can.

01:09:37 - Richard Campbell (Guest)
I thought you'd be done. Yeah, that's what I figured, so I didn't even check.

01:09:39 - Leo Laporte (Host)
I had no idea you were still there three hours later. Yeah, anyway, thank you for doing that, and if you're not a member of the club, you're missing some great content, some really fun stuff. Our show today brought to you by Zip Recruiter the best way to hire. It can be overwhelming to have too many options in life, such as well when you're trying to figure out what TV show or movie to stream, or you're on vacation and you've got all those sites to see. Well, the same applies if you're a business owner who's hiring. It can be overwhelming to have too many candidates to sort through. But you're in luck. Ziprecruiter now gives you the power to proactively find and connect with the best ones quickly. How? Through their innovative resume database. And right now you can try it for free at ZipRecruitercom. Slash twit.

01:10:37
Ziprecruiter's resume database uses advanced filtering to quickly hone in the top candidates for your roles. You can see a candidate you're really interested in and then unlock their contact info instantly. Did you know 320,000 new resumes are added to ZipRecruiter every single month? 120,000 new resumes are added to ZipRecruiter every single month. That means you can reach more potential hires and then sort through them and fill those roles faster. No wonder ZipRecruiter is the number one rated hiring site based on G2. Skip the candidate overload. Instead, streamline your hiring with ZipRecruiter. See why four out of five employers who post on ZipRecruiter get a quality candidate within the first day. Just go to our exclusive web address ZipRecruitercom slash twit right now to try it for free Again. That's ZipRecruitercom slash twit ZipRecruiter the smartest way to hire, and we thank them so much for their support.

01:11:36
Of this Week in Tech, a couple of AI stories. Amazon, it has been discovered by the information, has agreed to pay the New York Times $20 million a year for its content. I think these are good story. I think this is. I'm sure other AI companies may not agree, but this way content creators get somewhat compensated for their work. Right, it's only about 1% of the time's revenue, but every 1% counts.

01:12:11 - Brian McCullough (Guest)
That's actually not insignificant. No, that's a good point.

01:12:15 - Leo Laporte (Host)
Yeah, it's good, it's a good point, yeah it's good.

01:12:16 - Brian McCullough (Guest)
It's a good precedent yeah, if you're at a company and you bring in one percent of revenue on one deal, that's you and it's just one ai company.

01:12:24 - Leo Laporte (Host)
This is the first time amazon has made an agreement with a publisher. It's the times first ai licensing deal. I didn't realize that. I thought for sure they must have other deals. In fact, they are suing open ai. Right, they're the ones that are.

01:12:38
They're the ones that have put the flag down in terms of suing in specific cases, right right, your company, axel springer, your new company, has a deal with open ai, as does news corp. Reddit signed a deal with google right for, I think, 20 million dollars also, maybe that's the new going rate. I have to think it's valuable to these companies, these ai companies, and at this point they're writing checks. Well, hey, by the way, reddit reddit.

01:13:07 - Brian McCullough (Guest)
I mean, this is not I'm not an investing sort of dude, but like have you seen what reddit stock has been doing and their revenue? They have been, I think, doubling revenue in the last year since they started getting on that sort of AI licensing train.

01:13:23 - Leo Laporte (Host)
Yeah, their revenue increased 78% to $500 million. That's year over year. Daily active uniques increased 21%. I'm a big Reddit fan. I use Reddit like crazy. Daily active uniques increased 21%. I'm a big Reddit fan, I use Reddit like crazy. But one of the things they did that was really smart. They created their own search tool, and this is in a response to the fact that Reddit users have for a long time known that if you added site colon redditcom to a Google search, you often got better results than just a general search of the web, especially if you're looking for expertise like which cough syrup to buy.

01:13:58 - Brian McCullough (Guest)
It's not only that, because we sort of alighted over this when we were talking about Alphabet and Google and got onto the YouTube thing. But you know, so far Google slash, alphabet has survived the AI moment without it affecting their revenues. But it's kind of because they might be in the act of abandoning the web, because if you have a cloud side and you're just providing answers, you don't necessarily need the web to provide those answers anymore and you can still put advertising around it. But to Reddit's point, if you're Reddit, you're probably seeing your traffic from Google search and things like that declining, so you might as well build your own search product.

01:14:44 - Leo Laporte (Host)
Well, and here comes Dig right, which is going to be an AI kind of version of Reddit. Kevin Rose and, ironically, the creator of Reddit, alexis Ohanian, bought the Digg name back and they're starting to create a new Diggcom. Digg preceded Reddit. I think a lot of us think that Alexis and Steve Huffman copied Digg to some degree, creating Reddit. I don't know if Reddit feels threatened by Digg, but Digg will be AI first.

01:15:13 - Mike Elgan (Guest)
So maybe now is the time, fordit, to cash in on these ai deals and, by the way, uh, uh, you, you brought up amazon, so amazon's ceo is sort of getting in hot water for saying that, uh, he thinks there should be advertising in the conversations you have with alexa plus, oh boy, oh good, like the truman show, remember. Remember Truman's wife would do ads, like when she was talking to him, that sort of thing.

01:15:40 - Leo Laporte (Host)
Truman, you know, your clothes are brighter than ever, thanks to Tide.

01:15:44 - Mike Elgan (Guest)
Exactly, exactly, and also and and the other the other bit of so. So that was bizarre. I think that that's the best reason yet to never use alexa plus. This is awful that they.

01:15:56 - Leo Laporte (Host)
Well, I've used it and it is awful. It's not great. Uh, we got it somehow.

01:16:01 - Benito (Announcement)
I don't know how we got it but this is how they're going to keep your ais cheap. By the way, people, this is how they're going to keep it cheap.

01:16:06 - Leo Laporte (Host)
Uh, yeah, you know the head of perplexity said uh yeah, one of the reasons we created a comment browser is because now we're going to have lots of signals for advertising, right?

01:16:15 - Mike Elgan (Guest)
and uh, speaking of things you you do use. Uh, leo, did you hear that amazon is acquiring bai?

01:16:21 - Leo Laporte (Host)
yeah, how did I throw my b.

01:16:26
So this was that little device that I've been wearing since january. They announced at ces in january and immediately picked it up because it was only 50 bucks, no subscription, which should have been a red flag, right right, that that should have told me. Right there, they're angling for an acquisition. But the idea was this thing recorded everything it heard for six months. It recorded all my conversations, everybody, everything. Every day, at the end of the day I'd get a list of things I'd agreed to to-do list items, a kind of sycophantic but fun summary of how my day went.

01:17:00
Uh, you could ask it questions. I made it sound like, uh, jk simmons, so it was kind of a gruff ai. I loved it, but as soon as I saw that amazon bought it, I thought I don't want this six months of my life. I mean, this is alex on steroids, right, this is everything, everything I've said for six months, going into that, including, by the way, one of the things they do at the end of every day is say, hey, we learned 23 new facts about you, would you uh, look at them to make sure we got it right?

01:17:25
it had more than 2 500 facts about me which amazon just bought the facts about you well, I deleted it immediately, but you will be glad to know that it hasn't slowed me down. I am now wearing two new pins and I have a third on the way. So I'm I. So here's okay, ai accelerationist again. Yeah, I truly believe that the one of the most valid.

01:17:51
There'll be many, many uses of ai in my life. I already use it all the time. But one of the most. There'll be many, many uses of AI in my life. I already use it all the time. But one of the most useful would be an agent working for me that listens to everything that goes on here's everything I agree to, and so forth and helps me in my life, kind of like her in that movie that helps me negotiate, navigate my life, remember things that happened, and so that's why I got the B. Navigate my life, remember things that happened, and so that's why I got the B. When I got rid of the B, I got this thing from.

01:18:19
This is called the limitless AI pin kind of does the same thing records everything, bullet points, it, analyzes it. Unlike the B, it saves the audio. So they say you really should ask people before you wear this in their presence. They don't worry, I can't hear you guys because I'm wearing headphones, so it only hears me. This is called the omi. I kind of like this. One same idea, but this has an open um api, so there are literally more than a thousand apps for it. You can have steve jobs, be the ai assistant that advises you based on what this thing hears, or jobs be the ai assistant that advises you, based on what this thing hears, or machiavelli, or he sounds like, I don't know.

01:19:04 - Brian McCullough (Guest)
There's no way. That's a stereotypical italian voice.

01:19:05 - Leo Laporte (Host)
Right, hey, you decided that you wanted to run the world. I can't help you, so, uh, anyway, I think, and then I've got one more coming called fieldy, so this is a category that is now very crowded.

01:19:16 - Mike Elgan (Guest)
Okay, so. So, lee, are you going to want to throw all those away? Because have you heard about the Brilliant Labs, halo smart glasses?

01:19:24 - Leo Laporte (Host)
I have those over here. Wait a minute.

01:19:26 - Mike Elgan (Guest)
Oh, you have the old ones, but the new ones, the next ones, are going to work like the BAI Plus. They're going to have multimodal AI, so they're going to hoover up information through the camera, so these things, these things are going to cost 300 bucks only. They say they have an agentic memory system called narrative that collects data from the from the camera and the mics, and it creates a knowledge base, just like the BAI does.

01:19:50 - Leo Laporte (Host)
So these are the original ones that. I really didn't like it very much.

01:19:53 - Mike Elgan (Guest)
The original.

01:19:54 - Leo Laporte (Host)
Original ones were a monocle yeah, I, yeah, I didn't get the monocle which you could see the prism here that reflects the heads-up display. I don't like this. A heads-up display should not have a prism in it. It should just know somehow bounce laser light off the screen but that company is coming a long way fast.

01:20:09 - Mike Elgan (Guest)
So the yeah, and they have an open api as.

01:20:12 - Leo Laporte (Host)
So this is really a development platform like the OMI, so I think that that's probably pretty promising. I agree, this was Gordon Bell's idea. The digital equipment guru Right, he wore, when I interviewed him, a little device around his neck that was doing the same thing, was recording everything, and he came up with the idea. Same thing was recording everything. And he came up with the idea. His wife, gwen bell who was wonderful got alzheimer's and he was trying to find a way to help her kind of remember things. And I know I'm getting alzheimer's at some point if I live long enough, so both my parents have it, so I just I'm just trying to record everything now so I could say, hey, what did I do last week?

01:20:52 - Mike Elgan (Guest)
well, but that's the thing about the brilliant labs, uh glasses. Then the next version is that they have this noaa multimodal ai agent that will basically you have conversations with your glasses about and it will incorporate the stuff that it's hard I think that's very promising.

01:21:07 - Leo Laporte (Host)
I don't know what the form factor will be. Johnny ivan, yeah, this is not a good look, is it?

01:21:12 - Mike Elgan (Guest)
yeah, you look like, you look like. Uh, uh, harry, harry S Truman the Vox stops here.

01:21:20 - Leo Laporte (Host)
Um, I think there will be, though, a wearable. I don't think it'll be a pin ultimately, although apparently what Johnny Ives doing for, uh, open AIs might be something like this. We don't know, it's going to be glasses. It's going to be glasses, it should be glasses, yeah just due to the success of the uh ray-ban made array bands which I know you're a big fan of but check this out, leo.

01:21:41 - Mike Elgan (Guest)
Here's the craziest thing about the brilliant labs halo smart glasses. They're going to have something called vibe mode so you can vibe code apps for the glasses while wearing them, by talking oh, I love it.

01:21:54 - Leo Laporte (Host)
Yeah, maybe it'll be for me. Like those pins. I'm wearing a couple of those pins. I could just wear a few of these glasses at the same time.

01:22:02 - Brian McCullough (Guest)
I would recommend that you're like harry truman with flavor flave I got a big clock uh yeah, I don't know.

01:22:12 - Leo Laporte (Host)
I'm all in on this stuff, yeah, but I don't think it's hype, I don't think it's bs. I mean, I think it could be. I don't think it's known yet. It could flop completely. We've been through three or four ai winners.

01:22:23 - Mike Elgan (Guest)
It could be another one, but if it's, if it hits, this is gonna be incredible and I'm I want it, I want to be in on it, that's why I think that the, the, the earth changing, uh effect of ai won't be because we have super intelligence, it will be because we have ourselves plugged into it right, so completely.

01:22:49 - Leo Laporte (Host)
It needs to train on our lives. Right exactly, or maybe not trained, maybe that's more like a rag.

01:22:56 - Leo Laporte (Host)
That's context for its existing training, you know I don't know, I I don't know enough about it to know if it's going to take off, but I I want to believe.

01:23:05 - Brian McCullough (Guest)
I think that's the problem we sort of touch on this obliquely for a second. But like what's the ultimate business model for this? Are you paying? Are are in? Are? Is my mom paying 20 to 50 a month for a subscription? Is that? Enough is it free but ad supported, or third option, both. So they're getting you coming and going.

01:23:27 - Leo Laporte (Host)
Well, this is where there is an opportunity for apple because, believe it or not, apple's kind of made the privacy promise. There's certainly marketing on the privacy promise. If they could come up with something that is an Apple service I already pay Apple, I don't even know it's a lot of money for Apple one If they could come up with an Apple service that they could say look, it's private, we don't share this with anybody, we don't put ads in it I think that's worth a hundred bucks a month, maybe. Um, I think that's worth 100 bucks a month, maybe 500 bucks a month. It's worth a lot, yeah, and I think, uh, you know, the usefulness will have to be proven, but it there's the potentials there and I was it richard that said that he thinks the idea of everything has to be free.

01:24:09 - Brian McCullough (Guest)
Like I think we're through the looking glass, we know it's not like, we know it's not well, but what I'm saying is is like people are comfortable now. If it's, if it's giving me answers for everything in my life, I I'm willing to pay 50 bucks for that people pay for the wall street journal.

01:24:25 - Leo Laporte (Host)
Right, it's got an expensive wall, the the whole.

01:24:28 - Richard Campbell (Guest)
If you're not paying for the product you are. The product thing has finally sunk in. People get it, yeah, and so more and more people you know even young people look at a given product and go, oh and I'm not paying for this, so how are you exploiting me? Right, they're. They're at least looking at whether I'm going to spend money or not, but it's like we won't use a product that's free because they know they're being exploited and mentioning the vibe coders and them keeping it on, you know, 24 7 if it's, if it's making you money.

01:24:54 - Brian McCullough (Guest)
Leo was mentioning the wall street journal but any, any of those, you know wall street newsletters for years that cost you thousands of they make you money, they. If it makes you money, you can pay whatever is above, yeah but?

01:25:06 - Mike Elgan (Guest)
but if you look at the chat uh companies, uh, the ai chatbots, they're making money in multiple ways that people don't even know about. So let's say you're a paid user of perplexity, which, leo, I know you are, I am too, so they take our subscription revenue. They also have paid prioritization and and enhanced um, uh, uh, what do they call it? Branding for partners? So they partner with companies. Those companies results in your answers oh and that. Oh, it's kind of like advertising.

01:25:38
Yes, and then it's like OpenAI is actually trying to get micropayments when they suggest something, and then there's an ad at the top or there's a link at the top and they go and you buy the product.

01:25:50 - Leo Laporte (Host)
You have to be careful. They get a piece of that. This is what enshitified Google, though right. Yep, this be careful piece of that. This is what in shitified google, though right. This is why I paid 25 bucks for kagi right exactly because I didn't, it actually ruined google full disclosure.

01:26:01 - Mike Elgan (Guest)
My son works at kagi, uh and so, but but yeah, exactly, that's more disclosure his boss is going to be on intelligent machines on wednesday fantastic yeah, that'll be fantastic.

01:26:12 - Leo Laporte (Host)
I'm really curious. I was a fan of Neva before this, which was a right, similar paid ad free search. I accepted. Just as you said, richard. I wasn't going to get free search, that wouldn't be in shitified, so I thought I'm going to pay for it. Neva was one of the first to have AI results. It was, I thought, very good they they pivoted because they said we can't compete against google. But times have changed. They might have pivoted too soon. We'll talk to vlad on wednesday on intelligence machines.

01:26:41 - Mike Elgan (Guest)
That should be interesting I mean it's, it's part of uh, you know, just to speaking of certification, it's part of the problem with ai chatbot. So one of the things that they do is they go to open access scientific and other types of websites that provide ad-free content of very high quality, and they are hitting them so hard with their, with their spiders, right yeah, well, we're in that phase, aren't we now?

01:27:06
yes, so that actually makes the not only makes their results faster, it makes makes it slower to for users to use the original site.

01:27:14 - Leo Laporte (Host)
Wikipedia. We talked last week with Molly White, who's a Wikipedia editor. It's killing.

01:27:19 - Mike Elgan (Guest)
Wikipedia, right. And then they're also competing in search engines. People are using them as search engines, but they're doing something Google couldn't get away with, which is that they charge a subscription fee and then they enshitify it through advertising and all these other things. Right, they're doing both Nice job. If Google did that, you know, and also if Google provided prioritization in search results without disclosing it, without labeling it, people would be on fire about that, right, but that's exactly what OpenAI and Perplexity do.

01:27:51 - Leo Laporte (Host)
We need to take a little break. I do want to talk about. There is a lot of money flowing. You mentioned it, uh, earlier brian one and a half billion dollar pay package for one guy. We'll talk with wall street journal. Thinks they knew who that person is. We'll talk about where you can get it.

01:28:07
Yeah, and he turned it down. Yes, he, I don't. Anyway, we'll talk about that in just a little bit. Mike elgin is here. Brian mccullough uh, richard camel it's so nice to have. I always I look forward to it because it's I get the smartest people and we talk about the most important things going on in technology, and it is always an education for me. I hope you feel the same. We just really love having you here every Sunday.

01:28:35
We do the show from 2 to 5 pm Pacific, that is, 5 to 8 pm Eastern, 2100 UTC. As I mentioned, you can watch it live. You can get it after the fact. Download shows All 1,043 shows are available on our website, twittv. There's a YouTube channel with most of the shows. We've been doing video longer than almost anybody has. Uh, of our shows and uh, of course, uh, you can subscribe in a podcast client. Leave us a nice review, if you would. One of the problems with having a show that's more than 20 years old. Now, we're not the flavor of the month, we're not the flavor of the year, we're not even the flavor of the decade anymore. So, uh, you know, publicity is dependent on you, our viewers. So if you like what you hear. Share it with friends. Tell them about Twit. We would appreciate that Our show today, brought to you by Zscaler, the leader in cloud security.

01:29:29
They are experts in the understanding that AI cuts both ways right. It's both amazing and problematic. Problematic hackers are using AI to breach your organization faster and more effectively than ever before. On the other hand, ai powers innovation. It drives efficiency. On the hand, it also helps bad actors deliver more relentless and effective attacks. That's why you need Zscaler to help you on both sides. Phishing attacks over encrypted channels here's a bad statistic increased last year alone by 34.1%, fueled by the growing use of generative AI tools and phishing as a service kits right?

01:30:16
Meanwhile, organizations in all industries ours, little ones like me, big ones like the biggest companies in the world are leveraging AI to increase employee productivity. They're using public AI for engineers with coding assistance all that vibe coding. Marketers are using AI with the writing tools, helping them craft new campaigns. Finance is using AI to create spreadsheet formulas. Companies are automating workflows for operational efficiency across individual and teams. They're embedding AI in applications and services that are both customer and partner facing. Ultimately, ai is helping enterprise move faster in the market and gain competitive advantage.

01:30:58
All of that is great, but if you're using AI, you need to think about how you protect your company's private and public use of AI. You also, of course, need to think about how to defend against AI-powered attacks. That's what Jeff Simon did. He's senior vice president and chief security officer at T-Mobile. T-mobile uses Zscaler. He said Zscaler's fundamental difference in the technologies and SaaS space is that it was built from the ground up to be a zero-trust network access solution, which is exactly the main outcome we were looking to drive.

01:31:35
Up to now, we've thought of perimeter defenses as being the way to protect yourself firewalls, for instance and then you need a VPN to get through the firewall so you can get to work, which means you've got public-facing IP addresses which are exposing your attack surface, and nowadays they're no match for hackers using AI. It's time for a modern approach. Zscaler's comprehensive zero-trust architecture in AI ensures both safe public AI productivity, protects the integrity of private AI and stops AI-powered attacks. Zero trust means deny by default. It means bad guys can't do anything until you've explicitly authorized it. It really works. It lets you thrive in the AI era.

01:32:24
With zscaler, zero trust plus ai, stay ahead of the competition and remain resilient even as threats and risks evolve. Learn more at zscalercom security. That's zscalercom security. We thank them so much for supporting this week in tech. Very happy to know that they're coming back uh, in 2026. Really, really appreciate your support. Thank you, z scaler. So new york times headline I like it and and a beautiful graphic to to go with it. It says ai researchers are negotiating $250 million pay packages, just like NBA stars, ai technologists are approaching the job market as if they were Steph Curry or LeBron James.

01:33:14 - Mike Elgan (Guest)
I love that framing too. How on earth did we get to a point where people who are changing human culture and determining the fate of nations are making as much money as men who run around in shorts throwing balls into little hoops? That's a good point.

01:33:32 - Leo Laporte (Host)
But think about it the reason basketball players and football players get so much money is because they make that much money for those teams right, they're worth every penny and they're still underpaid.

01:33:44
Yeah right, the teams are still making incredible profits. I think it's the same premise. It's why Mark Zuckerberg who, by the way, has even said look, I've got a cash cow here, I got a fountain of cash with Facebook I can afford to do two things to hire and pay people the best people, the best paycheck they could ever get anywhere and invest in infrastructure and GPUs and buy more of them, because the people coming here don't just want money, they want access to the hardware. And I can do both. And that's his super intelligence team. The Times talks about a 24 year old mike. What were you making at the age of 24?

01:34:31 - Mike Elgan (Guest)
uh, let's see, I was a room service waiter, yeah you were living on tips.

01:34:38 - Leo Laporte (Host)
Exactly. I'm trying to think. I think I was a bus boy at denny's. Uh, this guy uh, he's a matt ditke, he's an ai researcher was offered 125 million dollars in stock and cash. It was over four years. So, let's be honest, it's a mere 30 million, 31 million a year. How will he get by? I?

01:35:03 - Mike Elgan (Guest)
know in silicon valley you haven't seen home prices dickie said it wasn't enough.

01:35:08 - Leo Laporte (Host)
I wanted to stick with my startup. So mark came over, he drove his lamborghini up his kuntosh I'm here, my kuntosh and uh revised the offer to $250 million, with $100 million of that paid in the first year. Ditke said I was so surprised. I asked my peers what to do. Some said take the deal. He did. It's kind of hard to turn that down. Good for him. Yeah, I mean you might as well cash in because who knows, next year that money may not be uh there for one year.

01:35:48 - Brian McCullough (Guest)
Yes, please one year listen. The lesson from this show is if someone comes to you, and offers you 150 million dollars. Take it, I mean, depending on what they're asking you to do but wait, because here's a guy.

01:36:04 - Leo Laporte (Host)
He, uh. He followed amira marati, who was former uh interim ceo at open ai started a company called thinking machines lab. No, by the way, no product, they don't. We don't know what they're doing, but the venture capital market has thrown money at them uh yeah, largest, largest seed uh round raise in the history of venture capital and we don't even know what they're doing. But it's more.

01:36:31 - Brian McCullough (Guest)
They're more investing in miramarati, right saying well, but and also point out all valley money right like oh yeah, yeah, yeah, just it's all weird, yeah, but that's worth pointing out is that what zuck is sort of disrupting here is what would have been pouring in from vc funds. Yeah, into these startups. Continue a little bit, that's his own fund.

01:36:54 - Leo Laporte (Host)
He has so much money. So he first thing he did is he went to mira and said I'll buy it, I'll buy the whole company. She said no. So what do you do then? You go and you talk to each employee. According to the wall street journal, in the following weeks he approached more than a dozen of her 50 employees to say what would it take? His number one target, a guy named andrew tullock, co-founder. He offered tullock a billion dollars, a billion dollars with bonuses that could have been worth as much as one and a half billion over six years. Tullock said no, amazing. Tullock said no, doesn't. Tulloch said no, I'm going to make you a billionaire dude Wait.

01:37:47 - Brian McCullough (Guest)
Mark Zuckerberg did similar things in his life. He was offered a billion dollars.

01:37:56 - Leo Laporte (Host)
What's better than a million dollars A?

01:37:59 - Richard Campbell (Guest)
billion dollars. I saw the movie. He was offered that for his company.

01:38:02 - Brian McCullough (Guest)
They were other shareholders Right, but the point is right. As opposed to, this is a check that I will write to you in person, which is a different thing, but also-.

01:38:11 - Leo Laporte (Host)
By the way, meta denies this. Meta's spokesperson said it was inaccurate and ridiculous, but the journal's standing by it.

01:38:17 - Brian McCullough (Guest)
We're all sort of schadenfreude-ing Is that? Does schadenfreude fit here? Sort of schadenfreude-ing Is that?

01:38:24 - Leo Laporte (Host)
the schadenfreude fit here. No, it's the opposite of schadenfreude.

01:38:28 - Brian McCullough (Guest)
Schadenfreude is the shameful joy in the misfortune of others we are having pure joy in the fortune of others. Like so is. I don't know if the German, for that is. The question is it's so rare. Because some of the debate that I've heard internally from like the AI community is because some of the debate that I've heard internally from like the AI community is like it is Zuck coming on so hard and so gauche that it's turning them off.

01:38:55
Right, and then it is the yeah, but if you want me, so bad, then there is something here that is, you're proving to me that this is a generational opportunity, right, that I could make more money if I had the balls that you had and stuck it out, as opposed to taking Well all right, brian, how much did you make at the age of 24? Not that much money, but I had. I founded my first company at 19. So, oh, so you were already an entrepreneur.

01:39:28 - Leo Laporte (Host)
I was I was just a bus boy at Denny's.

01:39:32 - Brian McCullough (Guest)
Let's put it this way I had my friends as my. They lived with me and paid me rent because I could afford a.

01:39:39 - Leo Laporte (Host)
Oh, you've always been like a mocker. Huh, You've always had that.

01:39:43 - Brian McCullough (Guest)
But not to the tune of $100 million, that's for sure.

01:39:46 - Leo Laporte (Host)
Well, that's I mean, and honestly, no one needs a billion and a half, right? If you thought well, if I stay with thinking machines, I'll be worth half a billion in a year or two.

01:39:58 - Brian McCullough (Guest)
That would be enough. There's also a real evangelism in true ai circles that this is not just generational, this is um civilizational yeah.

01:40:10 - Leo Laporte (Host)
So if you thought that, whatever zuck is doing, however, my money, much money, is pouring into it isn't gonna be the thing. If you thought, no, no, we got a handle on it here at thinking machines, you would stay because you want to.

01:40:24 - Richard Campbell (Guest)
You want to be where the action is, especially since you're going to make money well, you remember that all of open ai was built on the idea of pulling those people out of google mind right, and paying them less right for the privilege because the mission was more important than the income.

01:40:40 - Leo Laporte (Host)
Yeah, uh, yeah, maybe is that it's Zuckerberg's just so gauche he's getting people. I mean, there's a billion and a half from somebody who's gauche is still a billion and a half. It doesn't matter you're.

01:40:53 - Richard Campbell (Guest)
You're a tech giant. You have a choice buy them at the beginning, before they've been successful, and be successful with them, or buy them afterwards, when they're much more expensive yeah I mean they should all get together.

01:41:04 - Benito (Announcement)
People got offers, take them all at the same time and then just hang out at the office and play ping pong. That's what they should do.

01:41:10 - Mike Elgan (Guest)
Yeah, right go up on the roof.

01:41:13 - Leo Laporte (Host)
Well, we should point out that marati took a bunch of people 20 people from open. Ai, I mean it's not like she's got the moral high ground here or anything.

01:41:22 - Richard Campbell (Guest)
Well, I think that she did it with money or with a better goal. Right, people get to work where they want to work, like if you're selling a culture and a goal, it's better than I'm just going to pay you a huge pile of money.

01:41:32 - Leo Laporte (Host)
It's probably the case that the really good people are not motivated by money. In this case, right.

01:41:40 - Richard Campbell (Guest)
To some degree. I mean, I've known lots of folks who joined microsoft because the signing bonuses were a quarter of a million dollars and a half a million dollars and so that, and they literally put in their year and walked it's hard to turn it down, especially if student loans and yeah that's, that's get yourself out of debt kind of money.

01:41:57
yeah, you know, and that was just right in the mill mid-tier software developers right back in time, because there was a period there where nobody in the Valley would work for Facebook at all and so they went to the Microsofts and the Amazons and people outside of the Valley to do their recruiting.

01:42:12 - Leo Laporte (Host)
What it looks like is there's a general belief that there will be one winner.

01:42:17 - Richard Campbell (Guest)
Well, in the Valley. That's the belief right, that there is going to be a winter mute. You know, using William Gibson's term, like this super AI moment, which I just completely every piece of evidence says that's not what's going to happen. It doesn't look like that right now it looks like nobody has that.

01:42:36 - Leo Laporte (Host)
There are some people like Amazon that are laggards or Apple that are laggards, but the top companies are all roughly equal, yeah, and then how do you measure that per se?

01:42:45 - Richard Campbell (Guest)
what do you mean by equal per se?

01:42:47 - Leo Laporte (Host)
well, deep mind, I would say deep mind, anthropic and open ai are in the race. I don't see one company pulling ahead, yeah, and then there's qn, there's, you know, alibaba, there's the Chinese companies, which also seem to be doing some really good stuff. I don't think there's a Monopoly. Let's say on no success in AI well.

01:43:07 - Richard Campbell (Guest)
So one of the things that's coming out of this is that llms are just not that hard to build, right well, not only are they not hard to build.

01:43:14 - Leo Laporte (Host)
The the secret sauce isn't secret.

01:43:16 - Richard Campbell (Guest)
Google published all of those papers? Sure, it is the ultimately, it's the weights and it's the source of data like. Let's back up a bit on the some of the stories we've gone over here. If you start paying for data, if you start making sure you're only training models on quality data that you've paid for, what does that model look like after the fact?

01:43:34 - Mike Elgan (Guest)
yeah, like the other. Yeah, the other thing is that that you know, right now people think in terms of chatbots, but I think think chatbots their days are numbered. In five years nobody's going to be using a chatbot Really, go to OpenAI, they're not going to go to chatbots. Oh, yeah, yeah, type in.

01:43:50 - Leo Laporte (Host)
It'll be a device.

01:43:53 - Mike Elgan (Guest)
Well, it'll be on all the devices. It'll be part of applications. One of the coolest things I learned about it on this network I don't remember who, I apologize, but um is is use of the word lex. Uh, the word processor lex, lex, writer, lex, writer, is amazing. It's like a it's. It's like a word processor, a cloud-based word processor, but the ai is sort of baked in but it doesn't write for you. It's like a partner. You'd have conversations with it.

01:44:25
It's pair pair writing kind of like pair coding right well, I mean it's, it's passive, it's in the back of you, it doesn't do anything until until you explicitly, uh, ask it to do something, but it'll you can just say, hey, what do you think of this? And it'll say, wow, this is, and it gives you really good feedback about your work and stuff like that. Right? So I'm not using a chatbot.

01:44:44 - Leo Laporte (Host)
So you're saying AI will be kind of mixed into the things you'll use. Absolutely yeah, it'll just be everywhere.

01:44:51 - Mike Elgan (Guest)
Everything will be running LLM, small language models, all kinds of different kinds of AI but you won't have to explicitly go and say I'm using a chat bot now.

01:45:11 - Leo Laporte (Host)
I don't think that's really gonna be a major use case in a few weeks. We should say is that, uh, lexpage, yes, and is, I think, 15 bucks a month right? Uh, it's a collaborative.

01:45:15 - Brian McCullough (Guest)
If I were, I kind of makes me wish I were a writer, maybe I should be a writer I've been screwing around with writing a sci-fi novel and I've been using it for this exact purpose, where it's been in my head, for years. So I know generally the plot that I want to do. But when it comes to like, okay, describe what the planet looks like, I don't have time to think about oh, is it a mushroom-based planet? Are the aliens, you know, amphibian or what Like? So you ask AI to either give me five options for the next paragraph or write a paragraph of what could come next, and then you don't take it, because usually what that paragraph is is garbage. But that is. That's the vibe of everything. That vibing is the thing, because it's like you can get in a flow where you're creating, you have the structure in place, you know where you want to go, but all of the intermediate steps where you lose the flow and you, you that that's what it can do for you and, uh, leo lost the flow.

01:46:33 - Richard Campbell (Guest)
No, that's, we lost the Leo. That's what happened. And then that happened. That happens occasionally. Uh, I do think you know AI is almost not the product, or LMS andMs are not the product, they're the interface. You're just going to see more and more places where they're going to appear.

01:46:47 - Mike Elgan (Guest)
And I think that's my point. I mean, I'm a big believer in glasses, that they're going to be the interface for us to be using, use it the way we now use chatbots. We just talk, talk and I just, you know, I think, these companies, if I was open AI, if I was you know some of these companies, I would be all over the glasses thing.

01:47:09 - Richard Campbell (Guest)
Sure that is where you look is the most useful piece of information.

01:47:13 - Mike Elgan (Guest)
Right, and then, and then the answer can be just whispered in your ear by the, by the hardware that's just right there next to your ear.

01:47:19 - Leo Laporte (Host)
So I should, I should, I should. I last week I should ask this question because we had three people on who had successful newsletters. I don't think, brian, you don't do a newsletter, right?

01:47:29 - Brian McCullough (Guest)
uh, not yet, but uh, morning brew does do a lot of newsletters yeah, but mike does, I'm, you know.

01:47:35 - Leo Laporte (Host)
I was thinking maybe this whole podcasting thing doesn't have legs, maybe I should do a newsletter instead and this. But this would be. But the thing is I, I can write, but I don't like to write well, maybe if I had a tool like this, it would help me yeah, what you should do a newsletter.

01:47:51 - Mike Elgan (Guest)
That's hilarious what you should do, leo, is you should. You should basically tell ai to take every episode and turn in a newsletter right, write it.

01:47:59 - Leo Laporte (Host)
I have content, yeah it exactly.

01:48:02 - Brian McCullough (Guest)
It just did that within five seconds. Just what he said would be a newsletter within five seconds.

01:48:09 - Leo Laporte (Host)
Should I do a newsletter, write it out for me and put it up on not on Substack anymore, because they invite people to join Nazi newsletters. Let's do Beehive. How about that? Where is yours, mike Substack? Okay, you saw the story.

01:48:23 - Mike Elgan (Guest)
I did. I felt that the criticism was somewhat disingenuous.

01:48:26 - Leo Laporte (Host)
I think it was an accident right, right, exactly.

01:48:29 - Mike Elgan (Guest)
They say. Oh, you know, the harshest critics say that oh, this really reflects the founders.

01:48:36 - Leo Laporte (Host)
Well, we know, there are Nazi newsletters on Substack.

01:48:39 - Mike Elgan (Guest)
Yes, the question is I mean they take a point of view that says, hey, if it's legal, uh, by the first amendment they can publish on our service, however reprehensible exactly, and that's not a crazy policy, no, that's common carrier.

01:48:55
Next you're going to blame telephones for communicating, transmitting that that's a little more than that but but my, my problem with the criticism is that they're saying all oh, they really reflect their thing. No, actually the truth is that Substack is the best source of sort of anti-Nazi, lefty kind of publications, like the number of powerful leftists who left their you know Wall Street Journal and the New York Times to go do their own publication and who are now liberated because they're not under an editorial you know mission of a newspaper. They're mostly on Substack. So if you look at the totality of what's on Substack, it's mostly a force for, you know, anti-nazi type stuff. But the other thing is, like you know, they apologized for the error. They said it was an error. They said they're going to make sure it doesn't happen again. They said that the content that was promoted was damaging and harmful.

01:49:54 - Leo Laporte (Host)
It was pretty bad.

01:49:56 - Mike Elgan (Guest)
This was yes, but this is not the policy of a that's exists in order to make sure that nazis have taylor taylor lawrence uh publicized it on her.

01:50:07 - Leo Laporte (Host)
Uh, yes, newsletter.

01:50:09 - Mike Elgan (Guest)
Of course, in the sixth paragraph she mentioned that it was an accident that should have been in the. In the second lead yeah yeah, but but it's um, you know, so it's, it's a bit disingenuous and I I hope that. I hope that the mob doesn't sort of tear them down and cause it.

01:50:25 - Leo Laporte (Host)
But there's also been some some talk that maybe substack isn't the best place to to try to make a success in a newsletter yeah, it's so good though you know you like it you still get promoted and and you get.

01:50:40 - Mike Elgan (Guest)
I've never been no, I've never been promoted by them at all. They've never thrown me a bone at all. Which, uh, which I wish they would, but uh, but uh, no, I just the formatting just defended them. Maybe they'll uh yeah, they're much bigger if I were going to start a newsletter.

01:50:58 - Leo Laporte (Host)
Now would be the time. While I still have a podcast, I could promote the newsletter on.

01:51:03 - Mike Elgan (Guest)
Yeah, you, you should absolutely have a news. I mean, you know what is a newsletter, you know it's, it's like it can be a video newsletter, so podcast news it's like so lazy, I just like to talk.

01:51:12 - Leo Laporte (Host)
It's so easy, you just turn on the microphone, you talk staff make benito do it.

01:51:16 - Brian McCullough (Guest)
I don't know, make benito, do it but also the ai outsource that to ai right, okay, that would get legs outsource it to Benito to outsource it to AI.

01:51:25 - Leo Laporte (Host)
Yeah, yes, um Anthropic did a study in which it found out why AIs become evil they're also in parenthetically. This is from the verge hiring for an ai psychiatry team. I don't know if that means a psychiatrist for the ai or for the users of the ai. Uh, jack lindsay, an anthropic researcher working on inter interpretability, who was also tapped to lead the company's fledgling ai psychiatry team, told the verge something's been cropping up a lot recently is that language models and maybe you've noticed this, I certainly have can slip into different modes where they seem to behave according to different personalities. Now, by the way, the flaw here is in thinking that it has any personality in the first place, right? Well, gosh, chatgpt is not acting like itself. Well, that makes no sense. There is no itself, right? Lindsay said this can happen during a conversation. Your conversation can lead the model to start behaving weirdly, like becoming overly sycophantic or turning evil. This can also happen over training.

01:52:37 - Richard Campbell (Guest)
Now I have a theory that is blame the victim model. It's your fault, you did this. You made the AI crazy.

01:52:43 - Leo Laporte (Host)
yeah, you talked well, I think what happens is it has a limited context window, right, if you get too many tokens in there we've all seen this ai start to slow down, they start to act weird, they start to hallucinate more.

01:52:56
It's just a limitation of ais sometimes the token block overflows yeah, friday's paper came out of an anthropic fellows program which had been funded to do safety researcher research. Researchers wanted to know what caused personality shifts. They found that, just as medical professionals can apply sensors to see which areas of the human brain light up in certain scenarios, they could also figure out which parts of the AI neural network correspond to which traits. And once they figured this out, it's almost like a functional MRI of the of the AI's weights. They could then see which type of data or content lit up those specific areas.

01:53:42
The most surprising part of the research to Lindsay was how much of the data influenced. How much the data influenced an AI's model qualities. One of its first responses, he says, was not just to update its writing style or knowledge base, but also its personality. He said if you coax the model to act evil, the evil vector lights. That portion of the ai brain will light up. They also found out if you train a model on wrong answers to math questions or wrong diagnoses for medical data, even if the data doesn't seem evil but just has some flaws in it, the model will turn evil yeah, yeah, the other thing.

01:54:20 - Mike Elgan (Guest)
Well, first of all, folks, um, you shouldn't be just, uh, accepting whatever personality comes out of things you should be doing, role prompting, and tell it what it's personal, tell it, um, so, always do that. But did you hear about, uh, or did you talk on any of the podcasts? Uh, leo, about jeff lewis, who, um, who basically went down this rabbit hole on chat gpt? He was a he's a tech investor, an early investor of open ai, and he started to go down this rabbit hole on chat dbt. He was a he's a tech investor, an early investor of open ai, and he started to go down this rabbit hole, and so, to feed his questions.

01:54:50
Apparently, as far as we can tell, uh, the um, the chat dbt was increasingly relying on a website called the scp foundation, which is a fiction site that began on 4chan, moved on to its own website and it basically is about this concept that human beings are being replaced by digital versions of themselves. Anyway, the guy would post this big video on on x? Uh, with this, this essentially ranting about this uh, conspiracy theory. Thinks he knows the truth about what's happening and all this kind of stuff. Very disturbing, but here's a but.

01:55:23
But the point is that because he was going so deep and spending so much time on certain subject matter, it it had to reach into fiction, right, and they talk about, uh, you know, chat, chatbots being evil and all that kind of stuff having weird personalities. Well, fiction is a great place to get that kind of stuff right, because they're words Like the chatbot, doesn't, you know, just is looking for words to throw at you, and so, anyway, this is an interesting case where this is a prominent person, an investor in OpenAI, who basically allowed, you know, he had what they're calling AI psychosis right, where it's because of its sycophantic nature, because of the other things that it does, it's making people kind of allowing personality traits to be reinforced exactly?

01:56:18 - Leo Laporte (Host)
well, exactly so. If you go to chat gpt believing in the global conspiracy and start prompting it that way, it's going to confirm your bias, your, your, your nonsense, right? That's important to know, right?

01:56:27
and I don't know I'm not a fan of ai safety uh guards, to be honest, because, first of all, it's an illusion that you can even do it right. Problem number one, number two uh, who's to say I? I think what we really need to do is train people to understand what you're getting from the ai and and and nobody's going to become a molotov cocktail thrower because they got the recipe on chat gpt.

01:56:56 - Mike Elgan (Guest)
That's not how it works, right but you know I I'm a huge fan of, of, of smart, prompt engineering and I got a whole bunch.

01:57:05
It's all about the prompt yeah so I have a great one that I use on perplexity a lot and it's basically it basically says you know, starts out with role prompting, as most of these things do, and it says you are a world-class, highly uh, particular fact checker and expert in the verifiable information.

01:57:28
Give it this whole paragraph and then say fact check this colon, and then I paste in whole articles, I can paste in a claim, I can paste in anything, and it does such a good job. It really, really does a good job, job. It will take every single sentence, every single claim in an entire article and it will basically do an entire treatment of that one statement about where it came from, what, what they're saying, and it's. It's very, has very high standards for like, well, this is a claim, it's not really verified, verifiable, it's more of an opinion, blah, blah, blah. It characterizes everything, and so people should be doing that all the time, because this, at some point you're going to be lied to, you're going to be the victim of hallucination, and so always be telling the chat bots tell me the truth, give me information is verifiable, right, that sort of language that's why I like rag.

01:58:19 - Leo Laporte (Host)
That's why notebook lm, google's uh uh rag that generates now video as well as audio summaries is great, because you could say look, you're, you've got a corpus of information I've provided you. Stick to that, don't make stuff up, don't listen.

01:58:33 - Benito (Announcement)
Yeah right yeah, but you realize, what you're saying is you're asking everybody to be media literate.

01:58:37 - Mike Elgan (Guest)
Then right, that's what you're asking and that's impossible. That's a fundamental into possibility's not going to happen. But I'm asking the sweet smelling audience of this particular network, who are way above average in terms of their willingness and ability to do this stuff. So, yeah, no, I wouldn't ask it of everybody, but certainly-.

01:58:54 - Leo Laporte (Host)
Well. The opposite, though, is also pretty patronizing saying, well, the rest of you couldn't possibly do this right, so we'll do this for you, or something?

01:59:04 - Mike Elgan (Guest)
Yeah, that's actually one of the other cool things about Lex. They actually have this whole. If you go into their help systems, a lot of the help is essentially training about how to be a really effective user of AI. And so you can I might have to pay for this. This looks good, but you can find really good prompt engineering help and just copy them and keep them in a, a note, you know, in some document somewhere and then okay but don't do the one thing I did.

01:59:31 - Leo Laporte (Host)
We had our ai user group on friday and as the thing was starting up, I thought you know I use obsidian. I have a pretty big vault over a period of years in obsidian and I found a prompt on reddit that would organize my obsidian and I just I don't copy and paste somebody else's prompt it completely trashed my obsidian, so I mean it made it look like this guy would want it to look, but it's not how I wanted it to look, right, so don't write your own prompts maybe right, write your own, but but study.

02:00:03 - Mike Elgan (Guest)
But if you don't feel like you know what you're doing, you can start with somebody else's prompt and then go in and tweak it and uh, you can just do magical things with not mad darren oakey, who is a vibe code in our club says it's kind of the opposite.

02:00:17 - Leo Laporte (Host)
It's irrelevant to what I do, because either the code that writes works or doesn't work. So it's not rag, it's not, it's not it.

02:00:24 - Richard Campbell (Guest)
It's going to either work or not work and not work in software is not that clear either.

02:00:28 - Leo Laporte (Host)
Like well, that's true, it can work for a little bit you know, the compiler is a great threshold like will it actually run right, but is it actually doing what you want it to do?

02:00:38 - Richard Campbell (Guest)
that's a good question, a little more testing.

02:00:40 - Leo Laporte (Host)
That's why you get to write the tests. All right, we're gonna take a break. There's so much to talk about. We're just getting through it. But uh, more to come in. Just a little bit. You're watching a special episode of this week in tech with very three very special people. We really love having them on. Brian mccullough from the brand new tech brew ride home, great to have you. Um, where do people go if you already subscribe to tech meme?

02:01:10 - Brian McCullough (Guest)
right home, just keep that's you don't have to do anything. New boss, same as the old boss. Um what?

02:01:16 - Leo Laporte (Host)
what website? Should they go to a website or just go to a podcast?

02:01:19 - Brian McCullough (Guest)
oh, I haven't worked all that out yet, just your, your podcast. Yeah, oh god, leo, the amount of things you got to shift over after eight years and blah, blah, blah. But um, the bottom line is, if you want the tech news every single day in the afternoon in about 15 minutes, uh, tech, brew, ride home.

02:01:37 - Leo Laporte (Host)
You could just always search ride home that yeah yeah oh yeah, and you can, yeah, ride home dotinfo.

02:01:43 - Brian McCullough (Guest)
Right, that's it, that's right, yes so that's still going to apply it's still going to apply, but if you're looking at it right now, you'll notice that tech memes probably still on there.

02:01:52 - Leo Laporte (Host)
It still says tech meme, but then the album art says tech brew. So I understand, I'm getting there, getting there you're, you're in a transition period, exactly, that's all. Yeah, you're pre-op, it's okay, you're transitioning. Yes, yeah, you'll get there. Uh, also mike elgin. Uh, he has not changed the name of his newsletter machine society dot ai. Where is your next gastro nomad adventure?

02:02:17 - Mike Elgan (Guest)
well, it's uh here in oaxaca in a couple weeks, I think, and uh, and so we've been doing, we've been having a lot of fun with, with all of our experiences. We're doing Sicily nowadays and Sicily is just really a popular one. But yeah, we do. We do wine countries plus Mexico and Central America. So that's what we do.

02:02:37 - Leo Laporte (Host)
And it's a small group. How many couples?

02:02:40 - Mike Elgan (Guest)
Well, it's typical. Well for for for Oaxaca, it's going to be five couples that's a tiny group yeah, yeah, a tiny group, the one you were at, uh, the one you joined in Oaxaca, leo, that was the biggest one we'd ever done, because it was the first one you made room for us.

02:02:54 - Leo Laporte (Host)
I knew that.

02:02:54 - Mike Elgan (Guest)
You, exactly, exactly, and so we didn't want to say no to to our dear friend, so it was so much fun.

02:03:01 - Leo Laporte (Host)
Oh my, my God, it was incredible. The food and the fun.

02:03:05 - Mike Elgan (Guest)
Yeah, just to explain what we do, we have a small group of people live together in some really cool house of some kind In Europe. It's usually a farmhouse in the middle of a vineyard or something like that and then we just do food stuff all day. It's all food and drink. We learn how to make things. We go visit farmers, we go to exquisite restaurants, we have chefs come to our house to make food for us, we do wine tasting, we do all that kind of stuff. In Oaxaca we do mezcal tasting.

02:03:37
So it's just a ton of fun, and people don't have to do first of all, we don't tell anybody what they're going to do. It's just a ton of fun and people don't have to do first of all, we don't tell anybody what they're going to do. It's a big surprise, right? So every day that we get up, we say make sure you're wearing a hat and sunglasses, we're going to go. And then we go somewhere and they find out what we're doing when we're doing it. So it's, it's really, really fun and uh, it's uh something. I.

02:04:08 - Leo Laporte (Host)
I think everybody deserves to do that kind of thing. It's real traveling, it's not touristing, it's traveling, it's great gastronomy in most cases. Highly recommend it, highly recommend it. And richard campbell, of course, host of windows weekly and run his radio show and dot net rocks with carl franklin. Of course we have made a youtube playlist of all his whiskey recommendations, uh, on our windows site so you can youtubecom slash windowsweeklyshow maybe, I don't know, I can't remember.

02:04:27 - Richard Campbell (Guest)
Yeah, they redirect somethingweirdfrommyclosetcom. Oh yeah, that's the best way.

02:04:34 - Leo Laporte (Host)
Go to somethingweirdfrommyclosetcom and you will see all of the whiskey recommendations. It's become a really popular part and a fun part of. I don't even drink whiskey, but it's a fun part of the show because you learn history. We had a pretty much a trip down memory lane with the history of England a couple of weeks ago. It was pretty incredible. So all of this is here and it. It's a wonderful thing.

02:05:03 - Richard Campbell (Guest)
It's a wonderful thing I think we've had the, the mcallen 18 cherry cast come up. Yep, that was the one where I ended up talking about that, the. When they were building the expanded distillery, they dug up a mesolithic barley farm.

02:05:16 - Leo Laporte (Host)
Same location, 6 000 years old so they really could say it's 6 000 year old whiskey you know?

02:05:24 - Richard Campbell (Guest)
yeah, well, they said 1824 for a reason they was. That's when they could get a license. Up until then they've been making whiskey illegally. They didn't want to talk about it.

02:05:33 - Leo Laporte (Host)
That's wild, that's wild.

02:05:35 - Richard Campbell (Guest)
This is a part of the fabulous whiskey creation process well, it turned out to be two and a half hours long, wow lots of videos if you like whiskey.

02:05:49 - Leo Laporte (Host)
Thank you, richard, our show this week brought to you by miro. I've been talking about miro for a couple of years now. Every day, new headlines speculate about how ai might be coming for your job, which just makes everybody anxious and fearful. But a recent survey from Miro might encourage you, tells a different story. 76% of the survey responders believe AI can benefit their role. Of course, 54% do struggle to know when to use it, and that's why you should try Miro's Innovation Workspace. It's an intelligent platform that brings people and AI together in a shared space to get great work done.

02:06:29
Miro has been empowering teams to transform bold ideas into the next big thing for over a decade. I know you've heard about Miro, but today they're at the forefront of bringing products to market even faster. By unleashing the combined power of AI and human potential, miro's innovation workspace will help your teams be faster, more productive and, ultimately, more effective. Here's how Teams can work with Miro AI to turn unstructured data I mean even like sticky notes, screenshots into usable diagrams, product briefs, data tables. Bring in that napkin from the bar. That's kind of a little soggy. Prototype it in minutes. It's more than just putting a bunch of ideas on a board too. You can rapidly iterate with your teammates to bring these ideas to life fast, quickly build on your ideas without needing the perfect question or prompt.

02:07:23
That's always challenging, as we were talking about. That's always been a thing. I think that's gotten in the way. When you're using these chat bots, wouldn't you love to just be able to throw everything you've thought and all your ideas up there and work iteratively with an ai. You don't have to be an ai master or or to yet another tool. The work you're already putting into Miro's Canvas is the prompt itself. Isn't that great Help your teams get great done with Miro.

02:07:52
Check out Mirocom to find out how that's. M-i-r-o dot com. M-i-r-o dot com. We thank Miro so much for supporting us for many years. We've been big fans of Miro. Thank you, miro. So I saw a couple of articles this weekend actually about AI expenditures now becoming the most important part of GDP. Here's one from Paulul kadrosky ai, capex capital expenditures is so big it's affecting economic statistics, boosting the economy, and this should maybe be a little warning beginning to approach the railroad boom. No, it's not that big. Well, no, it isn't in fact. Here's the graph. Uh, railroads in the 1880s ended up being six percent of the gdp. Ai data centers still only 1.2.

02:08:50 - Brian McCullough (Guest)
But it's growing fast. It is above the dot-com era, although noah noah smith says he's debating that a little bit.

02:08:59 - Leo Laporte (Host)
But the the. This is no opinioncom will data centers crash the economy?

02:09:05 - Brian McCullough (Guest)
The statistic that blew my mind was that the contributor to GDP growth in the last quarter from AI expenditures, data center expenditures was larger than the entire US consumer. Now, that's not that the spending on AI was more than the US consumer spends, but the contribution to the growth of GDP.

02:09:31 - Leo Laporte (Host)
It has a bigger impact than consumer spending which is good, because consumer spending is probably not so healthy right now.

02:09:39 - Brian McCullough (Guest)
Some of the reading between the lines of some of these pieces. There's another one from chris mims in the wall street journal from two days ago. That is suggesting that, like without essentially, what kudrowski and mims are saying is that there is a giant stimulus that is happening in the economy right now that is entirely from the people building out ai and it's both.

02:10:03 - Leo Laporte (Host)
It's a. It's good and bad because of course with the railroad boom there was a bust. Most of those railroads went bankrupt but the tracks didn't go away and we got an intercontinental railroad and that infrastructure powered a century of growth in the same thing happened in web 2.0. You know Facebook, all that underground fiber companies might have gone out of business, but the fiber stayed. So you think that that's what we're going to see with this.

02:10:33 - Richard Campbell (Guest)
Well, that's the nice thing about data centers, right? Is there still data?

02:10:36 - Brian McCullough (Guest)
centers, except there is a very strong and very short sort of turnover Moore's Law style turnover where a data center, as its provision today, 18 months from now, is basically worthless.

02:10:53 - Leo Laporte (Host)
No kidding Wow.

02:10:54 - Richard Campbell (Guest)
Except that Moore's Law is not maintaining its cadence because the chips aren't improving.

02:10:59 - Leo Laporte (Host)
Well, and you've built the data center, you've built the hardware. You have to upgrade it, but it's.

02:11:03 - Brian McCullough (Guest)
You've done a lot of the point is that, and so the the Noah Smith piece it's titled. Could this be a new problem that could blow up the global economy in the way that, like in 2008, blah blah, because people are taking out loans, that they're using the data centers to, um, uh, what do you call it? Reinforce those loans or guarantee those loans? And what they're saying is that if you're using debt to do it, if the data center is right, still there, it's not like it got blown up, but if you then have to spend the same amount of money to put new chips into it, put you know, maybe the infrastructure doesn't have to be retrofitted or whatever, but it is not the same asset that is sort of evergreen as fiber in the ground or railroads on the ground Well, also in the interim it could be really bad.

02:12:01 - Leo Laporte (Host)
You know there was a crash in 1893, thanks to the failure of these railroads, 500 banks went under. It was the worst thing until the Great Depression. So many people out of work, and that's what Noah is also saying here. A data center bust would mean the big tech shareholders would lose a lot of money. It would also slow the economy. But the scariest possibility is it would lose a lot of money, would also slow the economy.

02:12:27 - Brian McCullough (Guest)
But the scariest possibility is it would cause a financial crisis. Well, if, if, again, it the the the great recession, the housing crisis, was because all of the banks were overextended and lending to homeowners and so there was a run on liquidity that they couldn't keep up with, at least at this point. And I reached out to Noah maybe I'll talk to him this week. Like, from what he says, if you read his piece all the way through, there are around the margins, some of the debt stuff starting to happen, but it's nowhere. It's not like your bank or, you know, bank of America, JP Morgan or whatever, is in hock up to 190% of their balance sheet to these uh data centers yet but, um, he's raising the, the warning that Well, and the tech giants are using cash cause they've got it.

02:13:20 - Richard Campbell (Guest)
They've got it and I was just double checking some of these numbers and like less than 40 of the price of a data center are the servers and storage getting power put in pouring all that concrete hvac, yeah, yeah all ups, is that's? It's more than half, so that's not wasted completely last. Yeah, and the other thing is, the machines don't burst into flames after 18 months, like we do run older hardware. After five or six years you start to question the failure rates.

02:13:47 - Brian McCullough (Guest)
The two things, though that maybe okay. The, the scare headlines are oh, could this blow up the global economy again? Not yet, but maybe we're laying the groundwork for something potentially bad down the road. But the real issue today would be number one what does the economy look like without this spending? And you know I'm not going to get into politics or anything like that about jobs reports from last week or whatever, whatever but there are suggestions that, again, this is a huge, essentially non-governmental fiscal stimulus that is propping things up. And then number two, what we do know is that this era of tech build out is not extremely job intensive. Yeah.

02:14:35
And so, even if all of the spending is happening, it is not necessarily creating a ton of jobs.

02:14:42 - Leo Laporte (Host)
Careful you could get fired for saying that.

02:14:47 - Richard Campbell (Guest)
But there is so much uncertainty in the work market that most businesses I know are being very careful with their money.

02:14:53 - Leo Laporte (Host)
We notice and this is completely anecdotal, but there is definitely a drop off in advertising as companies get more and more nervous about the future. It's not that they're thinking it's going to be bad, they're just uncertainty. Business doesn't like uncertainty. No, they're holding under their cash. Yeah, they're saying, well, let's just wait and see. And what's the first thing that goes? Your podcast spending. That's why I'm starting a newsletter. Ladies and gentlemen, I'm very happy to say that's why I'm starting a newsletter.

02:15:23
Ladies and gentlemen, I'm very happy to say, yeah, you're getting into the lucrative newsletter, can you make?

02:15:26 - Mike Elgan (Guest)
any money.

02:15:27 - Leo Laporte (Host)
Well, some people are there that I think that there are people making a million bucks a year on the, on the oh there's one that's making five million a month or something like that.

02:15:35 - Mike Elgan (Guest)
Uh, they let her from an American. She's making bank and she's got a whole staff and everything, so she's really good. I wouldn't be that good and she's got a whole staff and everything, so she's really good.

02:15:46 - Leo Laporte (Host)
I wouldn't be that good. Your plan is to not be that good. My plan is not to be that good. Yes, I actually subscribed a letter from an american, but, yeah, I'm probably one of those millions that are funding uh, funding heather cox richardson's uh, great success, yeah, um, what content strategy looks like in the age of ai? This is from fast company. We are rapidly getting close to something called google zero. This is something uh content companies have feared for a long time, where google no longer sends you traffic. Yeah, what do you do now?

02:16:26 - Richard Campbell (Guest)
traffic. Yeah, what do you do now?

02:16:30 - Leo Laporte (Host)
Google successfully started their own business. I was starting newsletter. Start a newsletter. The article in a fast company, uh, written by a Pete pical pachal, uh, creator of media co-pilot newsletter and a podcast host, so you know you can trust him says panicking is never a good strategy, but pivoting can be, and we've seen this this week. Wired and the verge both announced a strong push into newsletters. Newsletters, uh, I don't know. I we're. I was talking with paris martineau on wednesday on intelligent machines and and jeff jarvis and we we all agreed that while we subscribe to many newsletters, we read very few.

02:17:11 - Richard Campbell (Guest)
They end up in a folder in our email yeah, what's a newsletter is really about collecting email addresses so you have direct contact to your customers, right?

02:17:20 - Leo Laporte (Host)
I guess, how do you monetize that?

02:17:21 - Richard Campbell (Guest)
though.

02:17:22 - Leo Laporte (Host)
Well, by having something to sell yeah, I'd like to announce a new range of ai pins that people can wear in their home and their office. Hawaiian shirts leo hawaiian shirts. It could be the next big thing. Yes, um, actually, because of tariffs, these Hawaiian shirts are no longer a good deal. They went from being 60 bucks, uh each to 75 bucks.

02:17:48 - Brian McCullough (Guest)
They're made in Mexico, because Hawaii has a 75 percent. They're not, they're Mexican, hawaiian shirts.

02:17:53 - Leo Laporte (Host)
They're from San Miguel de Allende, but I love them and I'm going to continue to buy them. Uh, yeah, well, we'll see. There are certainly a lot of, especially in our sphere the tech journal a lot of companies folding, going out of business, disappearing, and non-tech, which went out of business a little while ago, at least kept their content online for a while. Now they're just referring it to the forums. You can't get to the old non-techtech articles.

02:18:24 - Mike Elgan (Guest)
It's just depressing as hell, yeah that's why, um, that's why I embraced uh, what is it called? Um? What the heck is my, my, my blog, a site called uh post Haven, I believe it's called oh, you still use post Haven. I, I learned about it from you, leo, but yeah, because pa.

02:18:43 - Leo Laporte (Host)
You know, back in the day was a y combinator startup called posterous and the whole point of it. I used it in 2009 when we were in china and I couldn't tweet or anything, but I could still use email because I can't block email. So I would email it to posterous, which would then syndicate it everywhere else. That's right. Facebook and twitter yep that was great.

02:19:03 - Mike Elgan (Guest)
They got sold. Or new thing is that they will keep your content. They promise to keep your content online until the end of time, even if you stop subscribing, if you stop being a customer, and so that's why I use it.

02:19:16 - Leo Laporte (Host)
Their tagline is post haven is the blogging platform designed to outlive us nice wow, that's pretty cool.

02:19:26 - Mike Elgan (Guest)
I don't know how they can promise that, but yeah, I mean it's who knows if they can keep that promise, but but I've had so many publications that I've worked for. Just have it be washed away and it's gone.

02:19:36 - Leo Laporte (Host)
Most of this stuff that I've done in my career is available only at the internet archive yeah, me too, actually come to think of it, except for the podcast we should all be making a donation to the internet archive. Yeah, I do, I do. God bless them there is a.

02:19:54 - Brian McCullough (Guest)
There's a new article I just put in the chat about beehive which I haven't read, so I can't uh verify that.

02:19:59 - Leo Laporte (Host)
But it isn't beehive everybody's that's the new hot nude yeah boot newsletter platform. Uh, it looks pretty good. Uh, a ghost is another good choice because it's open source, but nobody's except postave is making the promise to keep it forever. Is that like you buy a grave and then they promise to put you know, keep it clean for the rest, forever, for eternity, is it like that?

02:20:25
I've never bought a grave before I'm asking for a friend where's the best place to buy a grave site? Yes, they call it. What do they call it? There's something like eternal. There's a name for it, like eternal something. Yeah, um, which? Which story is it the information story you're talking about?

02:20:47 - Brian McCullough (Guest)
uh, yeah, where is? Beehive the next billion dollars startup and uh, they have hit 30 million in arr. But, wow, their fees start at 43 a month if you just have a thousand subscribers, which would be a problematic, since sub stack only takes a 10 cut. So, like it's the, it's the ramping up in scale. I guess that is the difference that Beehive doesn't compete with substance yeah, they.

02:21:14 - Leo Laporte (Host)
They talk about Oliver Darcy, who left CNN to start a newsletter which, I don't know, on the face of it probably doesn't seem like a good career move. He's making a million dollars a year, yeah, yeah, 85 000 subscribers not all of them pay, but enough to, he says have a million dollars in annual recurring revenue. I don't know, maybe I should do a newsletter. I don't know, I don't. I, I, I, you say I have this weird thing that you should. If you do something like that, there should be some value in it for the reader, maybe that's.

02:21:50
Maybe that's a flaw in my thinking, I don't know I, but then and then I have to, then I have to come up with it, right? Well, what can I put in there?

02:21:58 - Brian McCullough (Guest)
no, leo, we've been telling you ai can do this let ai do it, if you don't care about the product just add the slop.

02:22:06 - Leo Laporte (Host)
Yes, I'm part of the problem. Uh, we will later in this week have scott wilkinson's home theater geek for our club members. I hope he will review the annual value electronics tv shootout and lg oled tv's last place finish. Nilai patel, who was one of the judges, writes about it at the verge. I won't tell you who won. I don't want to ruin scott's home theater geek. If you're a club member, watch for that. Uh, when is it? I have to check see when it's uh coming up, because I think sometime this week we're gonna do uh tomorrow at 12 noon. So, um, no, that that can't be right. Yeah, monday, it says every thursday, but tomorrow's a monday, so that's when they do the live recording.

02:22:59 - Benito (Announcement)
But then the, the, oh. He puts it out on Thursday, it publishes on Thursday.

02:23:03 - Leo Laporte (Host)
So if you're a club member, you can watch it and participate tomorrow at noon Pacific, 3 pm Eastern. If you're not a club member, the audio will come out. But a show like that, you should really get the video, and for that you need to be a-. The video will be on YouTube If you're not a member the?

02:23:19 - Benito (Announcement)
oh really we published the video to you.

02:23:20 - Leo Laporte (Host)
Yes, yes, on the oh, I thought that was audio only audio only on our site.

02:23:24 - Benito (Announcement)
Uh video on youtube okay.

02:23:27 - Leo Laporte (Host)
Well, there you go. I was gonna say join the club, but I guess you don't need to do that anyway. Benito just saved you ten dollars a month. Comes right out of your paycheck, benito. That's all I'm saying.

02:23:40
We have a new no, I'm kidding, we have a new national cyber director. This is actually the believe it or not the first person confirmed by the senate and, in the new trump regime, the first senate approved cyber security official. Uh. Karen cross came from the millennium challenge corporation, uh, so he at least knows something about technology, um, and he? Uh will be responsible for cyber policy across all federal agencies. That's why I mentioned it on this show. That's a pretty important job, he said in a statement. As the cyber strategic environment continues to evolve, we must ensure our policy efforts and capabilities deliver results for our national security and the American people. Unfortunately, we're firing many of the people in the government cybersecurity infrastructure, but maybe Yashan can turn that around. Is CISA even still a viable entity, richard like? It seems like we've let go most of the people there yeah, I don't know.

02:24:51 - Richard Campbell (Guest)
I mean, we're still working. We haven't read all the stuff they already published, which is what they're going to put out next. So yeah, well, I hope there's enough functioning there. Those guys have been leading the charge for a long time. Very, very important. That's right, it's, it's one of the secrets that I think a lot of the world just completes has forgotten about it. The us consistently led in these areas. They just do it so well, they don't have to talk about it.

02:25:11 - Leo Laporte (Host)
Yeah oh, I've decided I'm not going to get into newsletters, I'm going to do podcasts. This from bloomberg, this from Bloomberg Weekend how podcast-obsessed tech investors made a new media industry. Oh, that's who did it. Well, thank you. We appreciate your hard work on our behalf. They're probably talking about shows like All In, which recently hosted both the president and vice president, or TBbpn. Apparently they're in. I didn't know this. There are a number of new podcasts, uh, hosted by not tech journalists like you and me, but by former venture capitalists yeah, people want to.

02:25:57 - Mike Elgan (Guest)
People want some of it to rub off of them, and by some of it I mean some of their money, right, because?

02:26:03 - Leo Laporte (Host)
when I listen to all in, it really is often like well, I was in japan and I found a sushi place that no one knows about. You know, it's kind of for rich people, but there is this aspirational thing, isn't?

02:26:18 - Mike Elgan (Guest)
there. That's what the that's the whole model. So that's yeah has always been like the wall street journals model, for example. You know, 10 of their of their audience are movers and shakers and 90 percent want to be movers and shakers. And so you subscribe and that's what it's all about. So they don't want tech journalists crapping all over everything, being snarky and and, yeah, we're not so positive about a lot of this stuff.

02:26:42 - Leo Laporte (Host)
Yeah, according to bloomberg, the seeds of today's podcasting boom were planted in 2020. I thought that was such a ridiculous line we've been doing this for 20 years now, but no, in 2020. It all started the early days of the covet 19 pandemic. That march, a group of venture capitalists premiered the all-in podcast which would develop into the preeminent vehicle for insidery vc musings. You know what's interesting about all in, though they don't have advertising. I guess they don't really need to do they I?

02:27:12
was just thinking of the john steinbeck line that every american is a temporarily embarrassed millionaire the same month, a startup called clubhouse released an app for remember this for free form audio broadcasting used by people like oprah winfrey and funding from andreessen horowitz at a four billion dollar valuation. Where is clubhouse, uh, today? What happened to them?

02:27:36 - Brian McCullough (Guest)
went well still still kicking. They raised enough. Look, one of the things is not to poop on anybody. I might want to talk to anytime soon, but look, there are shows like the Dorkish Patel podcast, or Dorkish Patel podcast or the Eric Torenberg network of podcasts that would pop up and then you'd be like, wait, how did all of a sudden they get Sam Altman on their show overnight?

02:28:16 - Leo Laporte (Host)
And I they it makes me so jealous because I can't get these people and it just makes me crazy.

02:28:19 - Brian McCullough (Guest)
This is my point is that there there is a, a network of folks that are like well, these are shows that are friendly and um so ah, they're willing to be on it because they're not going to get the tough question well, I'm not even saying that. What I am saying is that these are friends of ours and the the chats go out the emails there are buddies.

02:28:42 - Leo Laporte (Host)
They're in the same industry, right, they're vcs, they're, they're start, they're, they're entrepreneurs, they're startup people.

02:28:48 - Brian McCullough (Guest)
Yeah, exactly, exactly, exactly so and and this is, there are peers. This is something that especially folks at a16z have been talking about for years go direct, go direct, go direct. So they, they tried that thing called future, that sono uh was at the head of and and things like that, and and they wanted their own media brands. But what they decided I guess folks in the VC sphere was that you don't need to necessarily create your own brand. You just need to have your stable of of folks that uh, yeah, well, it works and people live.

02:29:23 - Leo Laporte (Host)
People do listen to it. Right, the numbers are good.

02:29:25 - Mike Elgan (Guest)
Yeah, absolutely yeah yeah it's really not that unusual. I mean, if you look at the celebrity podcast, sure, martin, lettuce and conan o'brien, friends, a bunch of actors and comedians talking about acting and comedians and you get your.

02:29:38 - Leo Laporte (Host)
I was going to do a true crime podcast on twit, but then they went up then, then that was over. So yeah, and somebody stole your idea.

02:29:46 - Mike Elgan (Guest)
You should have done the first episode like no, your point is exactly right.

02:29:50 - Brian McCullough (Guest)
Like let's take it out of tech. So like what? Like the conan o'brien thing, like I, or smart that's actually a good show. I enjoy that, but both of those are like they sold for hundreds of millions of dollars or whatever, because it's like, well, you, you're a celebrity in a sphere they have your friends to come on, yeah right and so although it irritated me.

02:30:08 - Leo Laporte (Host)
A couple of weeks ago, conan had mark maron on, who is now retiring from his very good podcast, and conan said you started this whole thing. And I thought, no, he didn't, no, no, no, he started in 2009.

02:30:21 - Benito (Announcement)
Where, yeah, yeah, I think he meant comedian podcasts.

02:30:25 - Leo Laporte (Host)
Oh, is that what he meant? Yeah for.

02:30:26 - Benito (Announcement)
Conan, because that's Conan's lineage right.

02:30:28 - Leo Laporte (Host)
That's all that matters, right? Comedian podcasts yeah, no mention of Adam Curry. No, that's all right. You know what I'm being facetious? I don't want credit for anything, it's all silly.

02:30:44 - Benito (Announcement)
You were the first VTuber. Leo, for sure, You're the first VTuber.

02:30:47 - Leo Laporte (Host)
I was the first virtual reality character on the site in 1998. And I won an Emmy for it A real-time 3D virtual reality character, and Mark Zuckerberg and I had legs legs, by the way, back in 1998 and I had legs. All right, we're. No I it's silly to look for that kind of credit. I'm just happy we're still here after 20 years. I I feel very fortunate that we've got to do this for so long.

02:31:20
Uh, we are going to wrap things up in just a bit one more commercial before we go and, uh, some final stories, including a, a story that, uh, richard put in and brian says he has something to say about. So that'll be coming up in just a second. But first a word from our sponsor, us cloud. Love these guys. They are. What would you think if I said us cloud? You'd think they were a cloud company, right? Well, no, although sort of. They're actually the number one microsoft unified support replacement. We've been talking about them for a few months now. Us cloud is the global leader in third-party microsoft support for enterprises. They support 50, 50, five zero over the Fortune 500. And there's a good reason for it. Switching to US Cloud saves them and can save your business 30 to 50% over Microsoft. Unified and Premier support Not just less expensive better support, faster at least two times faster in the average time to resolution versus Microsoft. Also, I think the best experts they make a concerted effort to bring in the average time to resolution versus Microsoft. Also, I think the best experts they make a concerted effort to bring in the best engineers with an average of 16 years working on Microsoft products with break fix and like that. These guys are the pros from Dover. They really are smart. But now they're going to live up to their name because US Cloud has another thing they're doing. That's really good. They're excited to live up to their name because US cloud has another thing they're doing. That's really good. They're excited to tell you about a new offering, something I don't see Microsoft offering saving money on Azure. I don't think that's in their interest. Us Cloud, it's in their interest. They call it their Azure Cost Optimization Service.

02:32:59
Saving on Azure is easier than you might think thanks to US Cloud. They offer an eight-week Azure engagement. It's powered by VBox. It identifies key opportunities to reduce costs across your entire Azure environment. Now you're going to get that expert guidance from US Cloud's senior engineers yeah, the same guys who have an average of over 16 years with Microsoft products really smart people. At the end of the eight weeks, you're going to get a full interactive dashboard that will tell you where you have rebuild opportunities, downscale opportunities, even VMs that are just completely unused, and after a while that happens right Now you can reallocate your precious IT dollars towards needed resources.

02:33:42
Why spend money on something you're not using? Although I may make a recommendation, you could also keep the savings going by investing your Azure savings in US Cloud's Microsoft support. That's what a few other US Cloud customers have done completely their unified spend. So the savings just keep on coming. Ask sam. He's the technical operations manager at bead gaming b-e-d-e. He gave us cloud five stars, saying, quote we found some things that have been running for three years. He was using this Azure engagement for three years, which no one was checking. These VMs were, I don't know. 10 grand a month not a massive chunk in the grand scheme of how much we spent on Azure, but once you get to 40 or $50,000 a month, it really starts to add up. Yeah, I think so. It's simple stop overpaying for Azure, identify and eliminate Azure creep and boost your performance all in eight weeks with US Cloud.

02:34:41
Visit uscloudcom and book a call today to find out how much your team can save. That's uscloudcom. Book that call today and get faster Microsoft support for less uscloudcom. We thank them so much for supporting this week in tech. You remember some of you are old enough to remember 1999 and the Y2K bug. I was at tech TV at the time and we were on call. We were thinking at midnight we were all going to have to come in and talk about how the ATMs are down all over the country and the whole planes are coming out of the sky and nothing right, nothing happened. Well, get ready, because the next apocalypse is coming up. You know this, richard. Apocalypse is coming up.

02:35:44 - Richard Campbell (Guest)
You, you know this, richard, in 2038, the y2k38 bug or the unix apocalypse, counting by seconds since 1970.

02:35:48 - Leo Laporte (Host)
And then they're gonna run out they're gonna run out of seconds because they use 60-bit numbers to represent it. Uh, oh, I'm sorry. Yeah, they use 32-bit and that's going to run out because you, you know. So 2038. So Debian has announced which is one of the probably the most popular Linux distributions, if you consider the fact that Ubuntu is based on Debian. A lot of uh pop OS is based on Debbie. A lot of downstream distributions start with Debian. They've announced they're going to switch to 64-bit time now, starting with Trixie, debian 13. Yay.

02:36:24 - Richard Campbell (Guest)
They've got 15 years. It's a simple solution and it now bumps the number out to the end of the universe.

02:36:34 - Leo Laporte (Host)
That seems enough. You never know, though, that should do seems enough. Uh, I hope that this, this is really proactive of them. I hope that other uh. Well, the issue is what have you stored in your database? Which?

02:36:47 - Richard Campbell (Guest)
was the real problem with the y2k bug too. Right, it's. What we stored in our databases was two digits, right, uh? So if you stored 32-bit time fields, that's not going to fix it. You're going to have to convert them all to protect them oh 314 utc on january 19th 2038.

02:37:04 - Leo Laporte (Host)
The number of seconds elapsed since the unix epoch, january 1st 1970, will be larger than can be represented by a signed 32-bit integer. That means you only get really 31 bits. Uh, it would be fine, writes the register, if the decision had been made in 1970 to store all the numbers of seconds in exactly that format.

02:37:24 - Richard Campbell (Guest)
So anyway, you'll be glad to know and, by the way, debbie is by far not the first to do this, like bsd has been like this for years. Like you, had a choice between a 32 or 64, like it's is android.

02:37:37 - Leo Laporte (Host)
That's going to be the question, because that's probably the most used linux derivative right out there. Um, anyway, we'll see. One of the reasons is important is debian debian is often used on servers that don't get updated or, you know, they use the long-term service version of it. Uh, and I did for a long time when I wanted a stable server. You use debian because it's you know it's rock solid, so those are the ones that really could be problematic.

02:38:05 - Mike Elgan (Guest)
So do it I wonder, I wonder if engineers, when they were developing this, predicted that they would have known the date right. They could have figured that out. Yeah, and they well think back.

02:38:16 - Leo Laporte (Host)
This was when unix was created by dennis ritchie and 1970. Brian kernahan. Think back to 1970. Did they think that this operating system would still be used 68 years into the future?

02:38:28 - Richard Campbell (Guest)
the original flavor of cobalt, stored years as a single digit because it would use it to 1970.

02:38:35 - Leo Laporte (Host)
10 years, who would nobody's gonna be?

02:38:39 - Mike Elgan (Guest)
using this software 10 years from now in the future they're gonna have ai. That will just do everything for them oh yeah we'll have jet packs.

02:38:46 - Leo Laporte (Host)
Well, we have the ai, yeah, where are our flying cars?

02:38:49 - Richard Campbell (Guest)
so, and, by the way, I bought a house on the back of y2k, like I did. You really, were you involved in that? I made so much money.

02:38:56 - Leo Laporte (Host)
So were you one of those old timer cobalt guys called in off the golf course.

02:39:01 - Richard Campbell (Guest)
I was fixing D-base apps, man. Oh, wow, yeah, no, we fixed it. When the Fed came down and said you're not a bank anymore if you can't pass July 1st 1999, the coffers opened up and that six months of of 99 I worked continuously and made.

02:39:20 - Mike Elgan (Guest)
Yeah, I bought a house y2k was pretty good for me too. I I basically did. It was five years leading up to that where I was on tv all the time. I made yeah vhs tapes explaining it.

02:39:30 - Leo Laporte (Host)
Like you know, I did this like whole tour going to businesses and explaining it, so it kind of ruined my new year's eve, 1999, because I had to be on call too we had written the software for uh, the port of vancouver and the port of alifax they controlled the container, specifically the safety systems for containers, and so they wanted us to be available if it failed well and the really it's a success story because it would have failed if people hadn't been very proactive and fixed the software.

02:39:59 - Benito (Announcement)
Yeah, that's what I was going to ask, Richard. If you didn't do that work, what would have happened?

02:40:04 - Richard Campbell (Guest)
Yeah, that's a great question. We don't know the answer to that. In the end, we had fixed the port software long before that and had trialed it. But they were nervous you know, these ops are nervous and they pay us a lot of money and so the good news is Halifax is four hours ahead of Vancouver, so by nine o'clock my time, you knew. They said, yeah, we'd already rolled over an hour, Nothing had happened.

02:40:23 - Leo Laporte (Host)
You can go out now. I'm sure, Mike, one of the things you were questions you were answering was ATM machines, because none of the ATMs were prepared to handle it, and and and the outcome would be unpredictable.

02:40:38 - Mike Elgan (Guest)
Right, and, and the outcome would be unpredictable. Right, that was the whole thing. But we're also reassuring people. The airplanes won't fall out of the sky, yeah, um, but other crazy things we're not thinking of might happen, and right, nothing of any major.

02:40:47 - Richard Campbell (Guest)
The banking was decided to open yeah, the banks were afraid because of the reconciliations filled for even a day. It's billions, yeah yeah, right, just in terms of the cleanup is huge, and so we had to run through this trial. The Fed invented this trial. It was a smart trial and we got through it.

02:41:05 - Leo Laporte (Host)
I hope that people aren't sitting back and saying, hey, by 2038, ai will be here, it'll fix it all.

02:41:12 - Richard Campbell (Guest)
Yeah, I hope they're not thinking I mean, that's not the first to the cable on this. Lots of folks have worked on this problem. It's solved, it's good, well, no we know what's going to happen.

02:41:19 - Leo Laporte (Host)
It's not a surprise.

02:41:21 - Richard Campbell (Guest)
And so you know, if you're using a contemporary version of my sequel, contemporary version of Postgres, contemporary version of Maria, any of those databases store 64 bit numbers, sort of 64 bit timestamps.

02:41:32 - Leo Laporte (Host)
And now, if you're using debbie and you can do that too, so why is lena khan, the former head of the ftc under the biden administration, tweeting celebratory posts about figma's ipo? She blocked adobe. Oh, because adobe was gonna buy figma for 20 billion dollars yep, oh.

02:42:00 - Richard Campbell (Guest)
And figma went off at 68 billion dollars after the pop current market cap yeah I mean, it's just what she's saying is a ridiculous stretch and she shouldn't have blocked it in the first flipping place. You know, those guys are willing to take it on. It was good for everybody, and now they're building a competitor and but I don't know why would be celebrating.

02:42:23 - Leo Laporte (Host)
Oh, you know, I made a lot of money for the figma shareholders by blocking the merger well, but uh just is it good for america? Is that I'm?

02:42:31 - Brian McCullough (Guest)
not arguing her point, but read the quote where, if a company is allowed to be independent longer, more valuable. Yes, even if I disagree with her politically I don't know that I disagree with that for various reasons as a investor or a tech industry. Big argument here is why did the Figma IPO go well? Well, that's what I want to talk about.

02:43:03 - Richard Campbell (Guest)
You want to go right. It's the most interesting part of the story. I think this is a fatigue. I think folks were looking for something invested and it wasn't AI.

02:43:12 - Brian McCullough (Guest)
I think my my take on this is that this is the most bullish thing that I've seen since the sort of tech crash post-COVID, in the sense that I would argue that it is a nominal AI play, like everything is these days where you can like Adobe's an AI play, Canva's an AI play because you have those AI tools and whatever, but this is a SaaS company. It tripled, or almost tripled, on an IPO pop. Those are you know dot-com sort of era things. After what happened with CoreWeave and all this stuff like I think tech IPOs are back on the menu and all this stuff.

02:43:57
I think tech IPOs are back on the menu. If you're Canva, you have been spending the last five days talking to your bankers and preparing your IPO. If you're Stripe, shouldn't you be maybe taking advantage of it? Here's what I put the history hat on one more time taking advantage of it. Here's what I put the history hat on one more time. When it's a bull market for tech IPOs, it's not just that people can achieve their valuations. They're achieving valuations after pops that you'd be stupid not to line up for. Like, so like Figma could have raised three times the money they raised. And so when you see people seeing that you can get pops like this, if you're a stripe, the amount of money that you can raise after your IPO is the thing that you can't turn away from. So sorry, I was pontificating for a long time, richard, but like my overall point is like more than anything else, I think this is the ipo that says we might be back at the races for at least the near term I agree yeah, I'm just not a hero.

02:45:14 - Mike Elgan (Guest)
Mike, I have no idea, I'll tell you what paul graham says so.

02:45:18 - Leo Laporte (Host)
lena khan tweets a great reminder that letting startups grow into independently successful businesses, rather than be bought up by existing giants, can generate enormous value a win for employees, investors, innovation in the public, to which paul graham says startups are risky. Sometimes, when you keep rolling the dice, things turn out well, sometimes not, but founders should be able to decide for themselves when to stop. He's saying you know, you should have let it happen, you should have let the acquisition happen.

02:45:51 - Brian McCullough (Guest)
Is he saying that yes?

02:45:53 - Leo Laporte (Host)
Because I mean, of course, if everybody knew, oh, you're going to be worth, you know so much more if you just hang on for a couple of years, yeah, yeah then well, if they knew that in hindsight.

02:46:04 - Brian McCullough (Guest)
Yeah, okay but I'm reading that. I'm reading that as him saying that like that shouldn't uh, okay, I was the government.

02:46:12 - Leo Laporte (Host)
I think he's saying government should not step in. Founders should be able to decide whether they want to.

02:46:17 - Brian McCullough (Guest)
I was taking too narrow a point on that, because they took the deal and right, they almost got lucky, right, sorry.

02:46:24 - Leo Laporte (Host)
Yeah, yeah. So I think they should give Lena Kahn at least five bucks.

02:46:30 - Richard Campbell (Guest)
So the question is what is Figman going to do with the money? I mean, the reason you go public is that you need to raise enough money that no individual investor can afford it or no group of investors can afford it. So you go to the public markets and raise a larger amount of money. Do you have an idea big enough that you need that much money?

02:46:46 - Brian McCullough (Guest)
Oh, but, Richard, we're talking about two different things. I'm talking about when. What is it? They're valued at like 55 times sales or something like that. Like, I think only Palantir is valued more than them on the stock market. Look, the point is, is that if you are a Stripe or like, let's say, a like God forbid, like a Perplexity, or one of these AI companies was ballsy enough to try an IPO, that's when you'll know you'll be back in the dot-com era. If one of these pure play AI companies takes a run at it in the next six months, that's when we'll know we're really back in the crazy days.

02:47:32 - Richard Campbell (Guest)
And Figma is certainly not a pure play AI company.

02:47:35 - Leo Laporte (Host)
It's not.

02:47:39 - Richard Campbell (Guest)
Not even close. It's a design company, right, and they could also be bought by adobe still just now in the public markets, right like there's a. There's always a question of what this actually looks like. The bottom line is do you have a mission sufficient to keep your shareholders happy?

02:47:53 - Brian McCullough (Guest)
and. But I'm arguing that it doesn't matter at a time when the uh public markets are like, hey, just give us anything we want to ingest it yeah, no, they yeah.

02:48:02 - Richard Campbell (Guest)
So you're betting on irrationality, which you know.

02:48:05 - Leo Laporte (Host)
Irrationality passes yeah, but hey, let's be fair. And, by the way, uh, ms khan, you probably shouldn't post this kind of thing on x, because you're only going to get people who disagree with you. But uh, let's be fair. The people who made the money were the investment bankers who were able to get in at the start price and then able to sell at the pop price.

02:48:29 - Brian McCullough (Guest)
Because, as Bill Gurley pointed out, they could have raised 3X had the bankers not priced it.

02:48:33 - Leo Laporte (Host)
Priced it at 30 bucks? Yeah, but that's what bankers not priced it. Priced at 30 bucks?

02:48:36 - Brian McCullough (Guest)
yeah, yeah, but the, but that's what bankers like, that's another bullish sign that that's an argument that we used to hear all the time in the old days, that we haven't heard for a very long time. Well, who should go public next?

02:48:48 - Leo Laporte (Host)
I said, um, I said stripe, stripe, perplexing yeah, if I mean we know no acquisition will be blocked unless you're unkind to the president.

02:48:59 - Brian McCullough (Guest)
Seriously, then all bets are off. Like this would be again. I'm looking at this from the Silicon Valley perspective. The investor perspective of things are insane.

02:49:21 - Leo Laporte (Host)
If a perplexity in the next 12 months floated an ipo and was successful, then it's the all bets are off, because then yeah it's a, it's a definitely a great time to jump now.

02:49:26 - Richard Campbell (Guest)
You've just smacked spax in the face. You cannot successfully. Yeah, you know it doesn't. It doesn't mean it's rational, but sometimes the rationality is profitable.

02:49:36 - Leo Laporte (Host)
That's why we have the $300 a month paywall behind this show, because we're making you money, ladies and gentlemen. Oh no, we don't.

02:49:43
We should but we don't. No, you're listening to this show absolutely free, unless you're a member of the club, in which case I thank you profusely. If you're not a member, it is now 25% of our operating expenses. It makes a big difference to us. Plus, you get a lot of benefits, including, of course, ad free versions of all the shows. Twittv club twit. We would love to have you in the club and and newsletter the soon to we actually offer a free newsletter. People say well, how do you know what's coming up on the club in the next week? Well, just subscribe for free at twittv slash newsletter.

02:50:21
And stay tuned for Leo's private, make you money newsletter.

02:50:25 - Brian McCullough (Guest)
Your newsletter should just be all of those things that you're wearing around your neck, the summaries that you're getting of your day.

02:50:32 - Leo Laporte (Host)
It's pretty funny. I have blogged it a couple of times.

02:50:34 - Brian McCullough (Guest)
It should be the newsletter on a daily basis.

02:50:37 - Leo Laporte (Host)
Yeah, I should do a newsletter.

02:50:39 - Brian McCullough (Guest)
I'll do a newsletter.

02:50:40 - Leo Laporte (Host)
I'll think about it. Maybe I'll get Lex to write it for me. Thank you, mike Elgin. I hope you have a wonderful time in Oaxaca. I'm so jealous. The best food, the best people. It was just a party, and what a cute little town. It's so great, it's, it's beautiful. I gotta sign up for one of these. This is, oh, richard, it highly encouraged that you do that. Plus, a lot of the people who are there are twit listeners, so you're going to be in a good, good company. You know smart people. It's a lot of fun. Uh, and, of course, uh, we still will mention hello chatterbox, your son's startup, even though he's now gone to the kagi search engine. We'll interview his boss on Wednesday.

02:51:19
Hello Chatterbox. Great way for kids in school and others to learn about AI in a safe environment.

02:51:26 - Mike Elgan (Guest)
That's right. It's very important now more than ever, and so I thank you for that plug and also my newsletter Machine Society. I'd love to plug that. If you don't mind, leo, please sign up. I actually happy that you have the free version. I have a paid version. The difference is very slight, so I my philosophy is that pay pay newsletter.

02:51:47 - Leo Laporte (Host)
Make mike a rich man a really good newsletter.

02:51:49 - Mike Elgan (Guest)
And then the paid one is a little bit, a little bit more stuff occasionally yeah, machine societyai.

02:51:56 - Leo Laporte (Host)
It is excellent. I read it every day. Thank you, mike. Thank you. Thank you to Brian McCullough, the Tech Brew Ride Home. Just go to ridehome. Is it info? Yeah, you'll find out about the new Tech Brew Ride Home, but if you're already a Ride Home subscriber, don't worry it's it's still in your inbox each and every day thank you so much to uh to talk about it's. It's a good, good time to be a podcaster, isn't it, brian? Um, the riches are flowing aren't they?

02:52:30
what has it ever been a bad time and, uh, my good friend Richard Campbell will be podcasting with me on wednesday on windows, weekly and of course, every week at run as radiocom and dot net rocks. Thank you, richard, I'm glad I mean, it was very nice of you to get off the boat early and be here with us.

02:52:49 - Richard Campbell (Guest)
I wanted to go home anyway. You've seen the view out here, dude. Why would I?

02:52:52 - Leo Laporte (Host)
I know you oh my god, look at that. It's a beautiful day in uh in Madeira Park, british Columbia so we sailed the other way in this morning, would you? Have seen the boat go by if you had been oh wow, that's cool, that's really cool. Did you wave as you went by? Say hey, there's our house there goes. Uh, thank you to all of you who uh watch and participate. I do this. This is a look, mike. It's a good look, it's a good look.

02:53:20 - Richard Campbell (Guest)
It's a look I get it. I can get a little kim jong-un vibe.

02:53:24 - Leo Laporte (Host)
I don't know what's going on. I need a fade if I'm going to do that. Yeah, uh, this is the show where we cover the week's tech news every week with a fabulous rotating panel of hosts. We're glad you watch. You can watch us live, of course, or, after the fact, subscribe, but do watch. Uh, we've been doing it for 20 years and, uh, I'm very happy to keep doing it for another 20 years. Uh, the good lord will, and the cricks don't rise. We'll see you next time. Another twit is in the can Bye-bye.