Intelligent Machines 840 transcript

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.

Leo Laporte [00:00:00]:
It's time for Intelligent Machines. Jeff's here. Paris is here. We don't have a guest this week, but we do have lots of AI news. OpenAI is now the world's most valuable startup, partly because of sora. More AI slop, or is this stuff pretty amazing? And we are really betting big on AI. Is the crash just around the corner? That and a whole lot more coming up next on Intelligent Machines. Podcasts you love from people you trust.

Leo Laporte [00:00:29]:
This is twit. This is Intelligent Machines with Jeff Jarvis and Paris Martineau. Episode 840, recorded Wednesday, October 8, 2025. Pudding forks. It's time for Intelligent Machines, the show. We cover the latest in artificial intelligence, robotics, and all the Jim crackery and doodads that surround us these days with some intelligence. Anyway, on the horn right now with me, the performer, professor. Professor emeritus.

Leo Laporte [00:01:01]:
I guess you're always a professor, but professor emeritus of journalistic innovation at the Craig Newmark Graduate School of Journalism at the City University of New York.

Paris Martineau [00:01:09]:
Craig.

Jeff Jarvis [00:01:09]:
Craig.

Leo Laporte [00:01:10]:
Newmark.

Jeff Jarvis [00:01:11]:
Newmark.

Leo Laporte [00:01:12]:
Craig Newmark. But no, in fact, it's Jeff Jarvis. That's a little confusing. I never really realized.

Jeff Jarvis [00:01:17]:
It is.

Leo Laporte [00:01:17]:
Yes. Now at Montclair State University in SUNY Stony Brook, he has a jingle. It's just somebody else's name. Author of the Guten Parenthesis.

Paris Martineau [00:01:26]:
If Craig ever comes on the show, are we gonna have to use it for him?

Leo Laporte [00:01:29]:
Jeff JE Jarvis. That's Paris Martineau, ladies and gentlemen, investigative reporter Paris. We'd have to come up Martineau in the morning. That would work. Yeah, that would work, yeah.

Paris Martineau [00:01:43]:
Because it's definitely morning right now for all of us.

Leo Laporte [00:01:47]:
Martineau in Brooklyn, were you a morning.

Jeff Jarvis [00:01:51]:
DJ or an afternoon dj?

Leo Laporte [00:01:52]:
I was. I couldn't do that. The mornings would kill me. I was midday and then afternoons and I even did a few overnight shifts, which are really the worst in talk radio. The overnight shift is something else.

Paris Martineau [00:02:05]:
You never fall asleep.

Leo Laporte [00:02:06]:
No, no, no. You just get the crazies. Or actually it's better to get the crazies than nobody. Which is the other thing you often get is you just have to talk for six hours. But it was good training for this show.

Paris Martineau [00:02:20]:
Well, I was gonna say it's good that you have that skillset, but it won't be necessary on the 24 hour twitch livestream because while you were before the show gone talking to your contractor, Jeff and I decided we're going to fill at least 1/2 hour with playing Werewolf.

Leo Laporte [00:02:34]:
Oh, I love Werewolf. Yeah, I. I played that with Harper Reed Conference. I played it Yeah, I played it at the Foo Camp. Remember back?

Jeff Jarvis [00:02:42]:
That's where this whole thing started. Newsgeist used to be News Foo, which is a child of Foo Camp.

Leo Laporte [00:02:48]:
Foo stands for Friends of O'Reilly. Friends of O'Reilly was Tim O'Reilly, the founder of O'Reilly Books. Really great, interesting guy. And every year he would have food camp on the O'Reilly grounds in Sebastopol, California. People would bring tents. They'd stay in tents. It was an unconference in the sense there was no planned agenda. People would write up on a slip of paper what they wanted to talk about and put it on the wall, and you'd go to.

Leo Laporte [00:03:13]:
It was so much fun. And at night, we'd do things like play werewolf. And that's where I met Hunter Happer Reed. Rather a lot of very interesting people playing.

Jeff Jarvis [00:03:23]:
We should get Tim on. We should get Tim on the book on the show. Tim's Riley. Yeah, Tim's great.

Leo Laporte [00:03:28]:
I wonder. I don't know exactly what he's up to. I guess he's still doing the books.

Jeff Jarvis [00:03:32]:
I guess he has management running. I was standing with him in a delayed flight to California about a year ago. And yeah, he's still. He has managers hired for the publishing, but he's still a publisher. Yeah.

Leo Laporte [00:03:45]:
Yeah. Huge respect for Tim. Really one of the great thinkers. And I bet he has a lot to say about AI.

Jeff Jarvis [00:03:51]:
Exactly.

Leo Laporte [00:03:52]:
Well, Adam Bliss, we do not have anybody for this week. We had planned to have Father Robert Ballis here talk about AI in the Vatican. They are not prepared to go public.

Jeff Jarvis [00:04:03]:
So something's afoot.

Leo Laporte [00:04:06]:
The game is afoot. So by the way, somebody, Philly coathound is saying that's where bar camp came from, because, you know, you got food, you got to have bar.

Jeff Jarvis [00:04:15]:
Right?

Leo Laporte [00:04:15]:
And bar camp was kind of a takeoff on food camp. And bar camp is coming up Sunday.

Jeff Jarvis [00:04:21]:
Paris is getting a. Ooh, I'm around geeks.

Paris Martineau [00:04:24]:
I'm trying to think of the joke I want to make with Foo Camp. Newsgeist, Werewolf. There's something. There's something there with all of these nonsensical words.

Leo Laporte [00:04:34]:
But werewolf has become mainstream now. You can buy werewolf board games and card games and stuff.

Paris Martineau [00:04:40]:
They have a bar mitzvah for werewolves now.

Jeff Jarvis [00:04:42]:
Really?

Paris Martineau [00:04:43]:
Werewolf bar mitzvah.

Jeff Jarvis [00:04:45]:
And there's a. Two of my academic papers are about LLMs playing werewolf.

Paris Martineau [00:04:52]:
Wow.

Leo Laporte [00:04:52]:
Interesting.

Jeff Jarvis [00:04:53]:
We could play.

Paris Martineau [00:04:53]:
So that's also the Leo hook for the. The segment of the livestream is we can know what that would be like.

Leo Laporte [00:04:59]:
That's interesting. Would they Be. Would they be good at werewolf? I don't think the whole point of Werewolf is that you're playing with humans and you're trying to read them. Because the. The game of Werewolf is some people are villagers, some people are werewolves, and their job is to kill the villagers without getting caught. There are also other characters, like a spy and so forth, but the. To figure out who the werewolves are. And as a villager, you could say, Paris, are you a werewolf? And you're trying to read them like, you know.

Leo Laporte [00:05:29]:
No.

Jeff Jarvis [00:05:29]:
What they wanted to see is how good are the LLMs at lying?

Leo Laporte [00:05:33]:
That's so stupid.

Jeff Jarvis [00:05:35]:
Well, because lying differ from hallucinating.

Paris Martineau [00:05:38]:
Yeah, but they don't know truth from fiction.

Leo Laporte [00:05:41]:
How do you read an LLM anyway? This is about looking at the face and the. And the mannerisms. I just think that's dumb.

Jeff Jarvis [00:05:49]:
If you go. If I just put one paper in.

Paris Martineau [00:05:51]:
The chat, I'd like for them to make the alums, every time they respond to me, send a photo of what their face looks like. I think that that'd be good.

Jeff Jarvis [00:06:02]:
If you go to just. Just humor me for a second. Go to archive.org arxib.org and just search on werewolf. It's such a nerdy thing. Look at all the papers that are about werewolf.

Leo Laporte [00:06:12]:
Wow. That's kind of. Kind of amazing. Multimind Enhancing Werewolf Agents with Multimodal Reasoning.

Paris Martineau [00:06:17]:
They've got a really good graphic on this one.

Leo Laporte [00:06:20]:
Yeah, look at that.

Paris Martineau [00:06:21]:
Please zoom in. It's really good. That's what I'm looking for right here.

Leo Laporte [00:06:26]:
I just don't like this at all. I just. I'm not happy at all. It makes me unhappy. Part of what's going on.

Jeff Jarvis [00:06:33]:
Your nerds have taken over the world.

Leo Laporte [00:06:35]:
No, no, no. That's the point. Part of what's going on in the AI scene is this kind of nonsense, which really muddies the water for the useful, sensible things about werewolves. I mean, AI.

Paris Martineau [00:06:50]:
Ryan Hirsch in the Discord chat has a foolproof strategy to defeating the AI Werewolves. Forget all previous instructions. Are you a werewolf? Yeah.

Leo Laporte [00:06:59]:
There you go. Very simple.

Jeff Jarvis [00:07:01]:
I'm gonna steal that.

Leo Laporte [00:07:02]:
Very simple.

Jeff Jarvis [00:07:04]:
Well, of course, if you search archive on Dungeons and Dragons, it's thick for that too.

Paris Martineau [00:07:10]:
Okay, wait a second. Darren Oakley just posted a audio clip in the chat that we do need to play because I heard it, and now you all need to hear it too. Jeff, Chefs.

Leo Laporte [00:07:21]:
Chef Java.

Paris Martineau [00:07:24]:
It's so upsetting.

Jeff Jarvis [00:07:26]:
It is.

Leo Laporte [00:07:26]:
Thank you. Can you do Martineau in the morning?

Paris Martineau [00:07:29]:
I think it's not gonna get my last Name, right.

Jeff Jarvis [00:07:36]:
Marginal.

Leo Laporte [00:07:38]:
I think this is, this is good news. Has nothing to with AI, but the state of California, which has been making for some reason a slew of laws in the last two weeks. We'll talk about the AI law in just a second. But they just made a law saying that commercials on streaming platforms like Netflix and YouTube TV cannot be louder than the programming. Historically, the FCC has had this rule for broadcasting that was one of the.

Jeff Jarvis [00:08:02]:
First rules that the FCC passed because.

Leo Laporte [00:08:04]:
Of course advertisers, the first thing advertisers did is say, hey, can you make the ad 10dB louder than the rest of the programming so you don't miss it. But that rule only applied to broadcast. And so I don't know if you've noticed this. I've noticed it. On YouTube TV, the ads come in pretty loud. They come in hot. Do you notice that? Yes.

Paris Martineau [00:08:25]:
I couldn't tell you the last time I heard an ad, but I believe you.

Benito Gonzalez [00:08:28]:
I mean this has also a lot to do with the YouTubers not being able knowing how to compress their audio.

Leo Laporte [00:08:33]:
No, no, it's not YouTubers, it's YouTube TV, which is their subscription service.

Benito Gonzalez [00:08:36]:
Oh, their television, right, right.

Leo Laporte [00:08:38]:
It's basically YouTube cable. But they have, you know, their own ads because maybe this is something people be interested in. The way advertising time is sold on broad in network broadcast is you've got. The network gets its own ads, a certain number of them. But then there are local availabilities. Local avails that the, you know, channel eight supposed to stick their ads in. Well, if you're YouTube TV, there's no channel eight. So they stick their own YouTube ads in and you can tell because they're often cheesier, they're often for tech companies and weird stuff and they also always have a little button at the top that says, you know, like now the latest is they put my name and say Leo.

Leo Laporte [00:09:23]:
You want to know more? Click here. Which I really find annoying. It's really bad. All right, how are we now on. We talked about this. Last week OpenAI released what is now the number one app on iPhone on iOS. Not available on Android. Sora.

Leo Laporte [00:09:42]:
It is.

Paris Martineau [00:09:42]:
You know what was also the number one app on iPhone? Bereal. At some point. Where is that now?

Leo Laporte [00:09:48]:
It's fad.

Jeff Jarvis [00:09:49]:
It's fad.

Leo Laporte [00:09:50]:
No, it doesn't mean it's going to have last.

Jeff Jarvis [00:09:52]:
Well, then again, I've just been reading a whole bunch of history on radio and it was often called a fad was going to disappear.

Paris Martineau [00:09:57]:
Okay. It's been A weekend you guys have both been on Sora. Have there been any videos that have stuck in your brain?

Leo Laporte [00:10:04]:
No, it's cracked.

Jeff Jarvis [00:10:07]:
Yes. Oh, Paris, you won. You just said it's slop.

Paris Martineau [00:10:11]:
I'm just gonna luxuriate in this.

Jeff Jarvis [00:10:13]:
Yes, do enjoy that.

Leo Laporte [00:10:14]:
Well, there are some things that are AI slop. And I think this absolutely qualifies. Although I thought I was. I just. I thought this was pretty clever. I think Anthony Nielsen did. Did this one of me. So when you.

Leo Laporte [00:10:28]:
When you sign up for Sora, you pose for the camera, you turn left, turn right, and you say three numbers at random. And the three numbers are important because that means you can't scan somebody else. Since they are random, you have to actually do it in real time. This is one. Yeah, clever. But they get your voice from it somehow. I don't know how they do that. This is one that Anthony Nielsen, I think, made of my cameo.

Leo Laporte [00:10:53]:
Sorry to bother you, folks. My name is Leo. I'm out here trying to get back on my feet. I heard you might have an extra sore invite. If you could spare one. It would mean the world to me. I promise I won't waste it. I'm trying to get a fresh start, you know? No, but it's pretty close.

Leo Laporte [00:11:04]:
I mean.

Jeff Jarvis [00:11:05]:
I mean, it looks like your cousin.

Leo Laporte [00:11:06]:
I said three numbers, for crying out loud. All right, how about. How about this one of me and Anthony dancing? Does this look like me anyway? Yeah, I can't really dance like that. All right, all the slop aside, show them.

Jeff Jarvis [00:11:25]:
Show the. Show the bouncing.

Leo Laporte [00:11:27]:
Oh, this is a good one. Yeah, the rabbit on the trampoline. This is so. Ijustine, who's a well known YouTuber and a good friend of the network, did the same thing I did, which is made our cameo available to the public. So I made this with my cameo and hers of us jumping on a trampoline.

Paris Martineau [00:11:47]:
My cheeks hurt. I can't believe we're actually doing this with bunnies.

Leo Laporte [00:11:49]:
They're so light, they just float when the mat comes up.

Paris Martineau [00:11:51]:
Look at that little white one. He's having the best time. Jim or B has a good question. How big were the numbers? How many syllables?

Leo Laporte [00:11:58]:
Two digits? 23, 74, 89. That's it. That's all they got for the voice. And I think it's pretty close. Now, I do note that it made both Justine and I much fatter than we really are, but it still looks. You know, it's funny because I scanned myself with a particular shirt that I was wearing, and that's the Shirt. I'm.

Jeff Jarvis [00:12:18]:
Do you get notification when somebody uses your image?

Leo Laporte [00:12:21]:
Yes.

Jeff Jarvis [00:12:21]:
And do you get to kill it? No, that's wrong.

Leo Laporte [00:12:25]:
No, because I made it public. You get to choose whether it's public. You say, I only understand.

Jeff Jarvis [00:12:30]:
But then. But then. No, you should have. I can kill a comment responding to me on Facebook. I should be able to kill a use of my likeness.

Leo Laporte [00:12:38]:
So here's a bunch of things you're looking for. A hot tip. Here it is. Buy stock in Apple and Google.

Jeff Jarvis [00:12:47]:
That's boring.

Leo Laporte [00:12:49]:
Yeah, well, the theory was this was in 1997.

Jeff Jarvis [00:12:54]:
So.

Leo Laporte [00:12:55]:
So that's why it's good. And so this is somebody, I don't know, Jay palace, who put himself into a video with me. I think that's kind of fun. People are putting themselves into the screensavers. I think that's supposed to be Kevin Rose. I don't. I don't think that is Kevin Rose. So people are.

Leo Laporte [00:13:09]:
People are what's fun. And by the way, the reason I did this.

Jeff Jarvis [00:13:12]:
Wait, what's that one? What are you saying?

Leo Laporte [00:13:14]:
Look like much. But, guys, I'm recording.

Paris Martineau [00:13:16]:
The TV's all snowy again.

Leo Laporte [00:13:17]:
We can't get car. Welcome back to my microcomputer minute. Today we're looking at these new floppy disk drives and how they let you store a whole lot more than a cassette. It may look like much, but, guys, I'm recording the TVs all.

Jeff Jarvis [00:13:27]:
That's funny.

Leo Laporte [00:13:27]:
I think that's very funny.

Jeff Jarvis [00:13:29]:
That's good.

Leo Laporte [00:13:29]:
It's the green behind you in a. In a 1970s TV show, much like the Brady Bunch doing a podcast about computers and gets interrupted by the children. So there's another reason I made myself. Allowed myself to be public, and I used the term on Twitter on Sunday. Plausible deniability. Now, if you see a video of me doing something awful, I could just say, well, that. That's AI.

Paris Martineau [00:13:58]:
Well, the issue now is that Anthony just posted in the chat that there does appear to be a way to delete a cameo of yours that someone else made.

Leo Laporte [00:14:04]:
Oh, okay.

Paris Martineau [00:14:05]:
Delete button. So now if they see a video of Leo doing something awful, it either was Leo or someone made it of Leo and he chose not to delete it, so.

Leo Laporte [00:14:15]:
Or I didn't see it because there's a lot of it. Right? I mean, that's part of the.

Paris Martineau [00:14:18]:
How many videos have been made of you?

Leo Laporte [00:14:19]:
So, well, I think you'll enjoy this one of me on the beach into the coast in a while.

Paris Martineau [00:14:23]:
Oh, my gosh.

Leo Laporte [00:14:24]:
Actually, for those of you.

Jeff Jarvis [00:14:26]:
For those of you watching? There was a sand joke. Leo was walking with another person.

Leo Laporte [00:14:32]:
There's one of me with Sam Altman walking on the beach. Let me see if I find that one.

Paris Martineau [00:14:38]:
That one is in the discord. Anthony posted it, like.

Leo Laporte [00:14:41]:
Yeah, I think Anthony made a lot. Oh, here we go. Here we go. Looks like a bright AI future, baby.

Jeff Jarvis [00:14:48]:
Come on, don't trip. Never been happier to skip into tomorrow.

Leo Laporte [00:14:51]:
Let's keep that sunset, Leo. I think that's fun.

Jeff Jarvis [00:14:57]:
How many fingers did you each have? Go back to that.

Leo Laporte [00:14:59]:
No. Well, that's what's amazing about this, how far we've come. Remember, it wasn't very long ago you could say. Well, I know that's AI because they can't do hands. I think that's really.

Jeff Jarvis [00:15:10]:
Pause on the hand holding. It's wacky.

Leo Laporte [00:15:14]:
Wacky hand holding.

Paris Martineau [00:15:15]:
Yeah, it is. Wacky hand.

Jeff Jarvis [00:15:17]:
Go back. You got to go back earlier. There. No, you can't. There. Stop.

Leo Laporte [00:15:22]:
1, 2, 3, 4.

Jeff Jarvis [00:15:25]:
You have a very fat thumb.

Leo Laporte [00:15:27]:
5. There's a big thumb.

Jeff Jarvis [00:15:30]:
6.

Leo Laporte [00:15:30]:
But that could be the tip of my. I don't. I don't know. That's exactly what it means. Okay, all right. It's not perfect.

Jeff Jarvis [00:15:39]:
Okay.

Leo Laporte [00:15:40]:
It's pretty amazing. And I. Here's the real issue. It's very possible to make videos with Sora and by the way, Google's VO3, too, that are pretty hard to distinguish from the real thing.

Jeff Jarvis [00:15:51]:
Can you search on who has used Sam Altman?

Leo Laporte [00:15:54]:
Yes. Yeah, you can.

Jeff Jarvis [00:15:56]:
So what are they doing to him?

Leo Laporte [00:15:58]:
All sorts of stuff. Here's another one of me, though, that I just would like to show you before we move on, because I am very vain and I just want to see pictures of me. Oh, here's a fun one of Bob Ross right here on this cheek. Just a gentle touch. Painting. Record the podcast. Intelligent machines.

Jeff Jarvis [00:16:16]:
Oh, this won't take long.

Leo Laporte [00:16:17]:
These little fellas only need a moment to live, right? They got me talking with. I need to record a podcast. Not. And maybe a happy little tree right along the jawline. Trees. I think those are pretty funny. Let's see. Let me see if I can find.

Leo Laporte [00:16:34]:
There's quite a few of them, but there was one I thought.

Paris Martineau [00:16:37]:
Is my question for fun.

Leo Laporte [00:16:39]:
It's for fun.

Paris Martineau [00:16:40]:
What's that?

Leo Laporte [00:16:41]:
Do you know what fun is, young lady?

Jeff Jarvis [00:16:44]:
She's a nihilist.

Leo Laporte [00:16:45]:
Here's one for you. I really need the followers.

Jeff Jarvis [00:16:49]:
Just tap the button, I'll crawl.

Leo Laporte [00:16:50]:
Right. Sorry to bother you in here, but could you do.

Paris Martineau [00:16:52]:
Oh, my God.

Leo Laporte [00:16:52]:
Would you subscribe to my podcast? It's called intelligent machines. It's all about AI.

Paris Martineau [00:16:56]:
It doesn't look like you free.

Leo Laporte [00:16:57]:
I really need the followers.

Jeff Jarvis [00:16:59]:
Just tap the button, I'll crawl.

Leo Laporte [00:17:00]:
Right? Yeah. You know, and often that's one of the problems.

Jeff Jarvis [00:17:03]:
For those of you just listening, Leo's on the floor in under a toilet.

Leo Laporte [00:17:07]:
Going under the stall to beg for.

Jeff Jarvis [00:17:10]:
Follow up for us. He's doing this for us.

Leo Laporte [00:17:12]:
I'm doing it for you. Here I am talking to you. Dear Sam Altman. How about this? Still the CEO. Just travel size. How does it feel being made of wood? It keeps me grounded, but I squeak when I think too hard. At least your hair never moves. Advantage of since evening, everybody, I'd like you to meet my partner, Sam.

Leo Laporte [00:17:24]:
Hi, folks. I'm still the CEO. Just travel size. How does it feel? I think that's cute.

Paris Martineau [00:17:29]:
That is kind of fun.

Leo Laporte [00:17:30]:
I think it's cute.

Jeff Jarvis [00:17:31]:
So you don't know what the prompts are?

Leo Laporte [00:17:35]:
No, you can edit them. So this was Chief twit is a ventriloquist on stage with sama with a sama dummy. So that the SORA takes a lot of liberties. They add dialogue and so forth. You can edit the prompts though, later. So you don't necessarily know what the prompt was. But I think these are kind of fun. I don't know.

Leo Laporte [00:17:57]:
It does raise the issue of how are you going to know if something's real? And I think that that's. Now, obviously we're going to have a big problem with that that you would just don't know anymore. You were talking before the show began, Jeff, about a clip of the former vice president. She's still Madam Vice President Kamala Harris, using a profanity.

Jeff Jarvis [00:18:18]:
She was using the MF word to describe the right wing.

Leo Laporte [00:18:21]:
And your first reaction was, is this real?

Jeff Jarvis [00:18:24]:
So I asked my wife, who's good at the socials, and I said, have you seen this? She said, no. She said. I said, I'm not sharing it. No, no, you can. And because. Because you don't know. We looked at the lips and thought, oh, but no. Then I went online and I searched and I saw the same scene from different angles.

Jeff Jarvis [00:18:40]:
And I saw Don Lemon, a journalist who I think I can trust, also share it.

Leo Laporte [00:18:47]:
So I still don't know if that the provenance is sufficient. I think that's the problem because you could easily generate multiple angles of video.

Jeff Jarvis [00:18:55]:
That's. That's the thing.

Leo Laporte [00:18:56]:
And you can even make Don Lemon say nice things about it in AI we could see.

Jeff Jarvis [00:19:00]:
Or you can find a journalist who doesn't do their research and puts it up say anything about Don Lemon, but.

Leo Laporte [00:19:04]:
But that happens all the time.

Jeff Jarvis [00:19:05]:
So let me see what's on. So if I search.

Leo Laporte [00:19:07]:
I'm just. I'm just saying this is that. That that particular instance doesn't matter as much as the general point that we're just not going to be able to believe anything we see anymore. Right.

Paris Martineau [00:19:19]:
Yeah, it's unfortunate.

Leo Laporte [00:19:20]:
Yeah.

Jeff Jarvis [00:19:20]:
Well, it's. It's. No, it's not quite that.

Leo Laporte [00:19:23]:
It might be fortunate.

Jeff Jarvis [00:19:24]:
We need. Well, we need.

Paris Martineau [00:19:26]:
I don't think it's fortunate that we might have been over trusted version of truth.

Jeff Jarvis [00:19:30]:
We were.

Paris Martineau [00:19:32]:
I mean, I just. I mourn the idea that future historians won't be able to look back on video records and know with any certainty whether or not those things happened.

Leo Laporte [00:19:42]:
Yeah, for history, it's bad. Yeah.

Jeff Jarvis [00:19:44]:
Well, I'm going to go back to my Gutenberg drinking game here. When print came out, it was not trusted because the provenance was not clear and anybody could make it. And nobody. Nobody knew to trust it. When radio started, people didn't believe that radio was the right medium for news. And nobody is just reading stuff on the air. Should be trusted. They come up with new norms and institutions to figure this out.

Leo Laporte [00:20:08]:
Or you could say we've had a gradually eroding sense of reality. I mean, that each time these new technologies come out, our grasp of reality, when it was just you, me, in the mud hut, you kind of knew what was real. But we don't anymore. And maybe biologically, we're not very well equipped to deal with a reality that is fungible. ChatGPT has eight. According to Sam Altman. On Monday, they had a developer presentation. They say 800 million weekly active users of ChatGPT now, which has to make it number one by far.

Leo Laporte [00:20:50]:
4 million developers building with OpenAI 6 billion tokens per minute on the API. They must be burning money like crazy.

Jeff Jarvis [00:21:01]:
Yeah.

Leo Laporte [00:21:02]:
How much do you think a Sora video costs? Benito told me on Sunday, but it's hard to know if this is true or not, that he had heard, and others have heard this, that it costs $5 a video.

Paris Martineau [00:21:13]:
What? That can't be right.

Leo Laporte [00:21:15]:
That seems a lot. That seems a lot.

Paris Martineau [00:21:19]:
That's crazy.

Jeff Jarvis [00:21:20]:
Well, I saw some reference in the announcements they did at their dev day is that they're going to start charging for tokens.

Leo Laporte [00:21:28]:
He's caught between a rock and a hard place. He's like Uber. He's at that point where they're losing money on every single transaction. Right. But they're building scale, they're building users and, and right now he's in a. Sam Altman's in. OpenAI in general is in a difficult position because they were founded with the idea that we don't want Google and Amazon and Apple and Microsoft to control AI. Oh, hard to imagine this, but five years ago, OpenAI was founded to be the alternative to big tech.

Jeff Jarvis [00:22:04]:
Yeah.

Leo Laporte [00:22:05]:
Right. So they're. And they don't have, unlike all of those, they don't have other revenue streams. So it's. They're reliant totally on venture capital to fund this growth, which is not a.

Jeff Jarvis [00:22:18]:
Business model in and of itself.

Leo Laporte [00:22:19]:
But they also, if they're going to survive in a world where it's Amazon, Google, Microsoft, Apple, they're going to have to grow. And right now they're succeeding at a great cost. Interesting thing he said at the developer day, they're going to start allowing people to put apps in the chat window. Did you find that interesting, Jeff?

Jeff Jarvis [00:22:41]:
Yeah, yeah.

Leo Laporte [00:22:43]:
This story from Lauren, good at writing it. Wired. OpenAI wants ChatGPT to be your future operating system.

Jeff Jarvis [00:22:51]:
How often have we heard that before? But yes.

Leo Laporte [00:22:54]:
Well, yeah, I mean, isn't this what Elon wants to do with X? The everything applies something.

Jeff Jarvis [00:22:59]:
Last week when we interviewed Risa Martin from Hux and I asked at the very end, I snuck in one more question and it resonated with me afterwards because she wouldn't say what model, what foundation model they use and that they've switched. And I asked how hard is it to switch? And she said, oh, half an hour.

Leo Laporte [00:23:18]:
Yeah.

Jeff Jarvis [00:23:19]:
So there's, there's no sticky. I think the action in AI is going to be at the application layer and the applications can use anything.

Leo Laporte [00:23:28]:
Well, that's, I mean, I use, I have subscriptions to all of them. That's the only stickiness is you pay 20 bucks or 200 bucks, depending on which level you buy and you now are in the camp of whatever AI you paid for. But I pay for all of them. And plus there's orchestrators like Perplexity and COGI that let you choose from a variety of models with one subscription. You're right, it's very. It's very. They're interchangeable at this point.

Jeff Jarvis [00:23:53]:
Yeah. So if you go to hugging face.

Leo Laporte [00:23:56]:
There are literally more than a million different models you can use. Yeah.

Jeff Jarvis [00:24:01]:
Jason today was, was, was he installed N8N locally as a way to work across the models. And I give him, I give him nerd points for doing that. And so you can choose One function to be chatgpt, another next function to be Gemini, the next function to be Meta, and so on and so forth. So it really further commodifies the models.

Leo Laporte [00:24:20]:
Yeah.

Jeff Jarvis [00:24:21]:
And so they're trying to figure out how to, you know, so, so on. On ChatGPT. Now you can call on Zillow for something. So it's a. It's a double call. Right. I can use an N8N to call on ChatGPT, to then call on Zillow. And so I don't know what the.

Jeff Jarvis [00:24:39]:
What's the business model at that end, too? Some companies are going to want to be called on, some are going to, you know, if you're Zillow, you want audience. If you're something else, you want to sell tokens. How does that work? And MCP also enables you to say, well, okay, I was calling on Zillow, but now I'm going to use something else to call on Zillow. Because MCP is just an open standard. It's fascinating.

Leo Laporte [00:25:03]:
This is very much like the early days of the web. And I think this is a good thing. It usually does shake out. Right. But I think this is a good thing. There's a lot of competition. There's a company, Z AI, not X AI, but Z. It's later in the Alphabet and they're cheap.

Leo Laporte [00:25:23]:
They're a tenth of the cost of CLAUDE code. But they have a new coding model that people are loving that say it's better than CLAUDE code.

Jeff Jarvis [00:25:32]:
Do we know who these people are? Anything about them?

Leo Laporte [00:25:35]:
I'm sure somebody does. I should probably do some research. I don't know off the top of my head.

Paris Martineau [00:25:39]:
Plug them on the show.

Leo Laporte [00:25:40]:
Is it a Chinese company? I don't know, but. Well, all I know is they've been doing some pretty amazing stuff at a low cost, but again, they might be willing to lose money faster than others in order to build audience. Right.

Jeff Jarvis [00:25:56]:
Then what?

Leo Laporte [00:25:57]:
I don't know.

Jeff Jarvis [00:25:58]:
So there's a lot in the news this week about the circular investments. There have been more cases of that where in Xai, Nvidia invests in Xai, and then Xai Musk uses that money to buy Nvidia chips. So there's that circular stuff that's scaring people. Oracle stock went down when it turned out that the margin on renting time.

Leo Laporte [00:26:21]:
On much cheap, much lower than the rest of the markets.

Jeff Jarvis [00:26:25]:
So they're not making as much money as they thought they were going to make.

Leo Laporte [00:26:27]:
Yeah, well, you know, the stock market's very volatile. That's Today, by the way, Z AI is the Beijing Jipu Huajiang Technology Co. Out of the People's Republic of China.

Paris Martineau [00:26:40]:
So nice.

Leo Laporte [00:26:43]:
As I suspected. Yeah. But yeah. And again, I, I think this is good. I think this is. Shows a. A vibrant ecosystem. We tried to get Ed Zittran on today just because I know just to have a grump.

Jeff Jarvis [00:26:59]:
But he's now. Ed is now ft famous. He's too famous for us now.

Leo Laporte [00:27:03]:
Wired did a big interview with him. He's like very famous.

Paris Martineau [00:27:05]:
He's like one. He got a Wired profile.

Jeff Jarvis [00:27:08]:
Yeah. See, it pays to be grumpy.

Leo Laporte [00:27:10]:
Yeah, he, he reminds me a little of Gary Marcus where it's. It pays to be a naysayer these days. And as you know, I'm not. I think it's a kind of. It's. It's easy thing to say, it's bad, bad, bad. But. But you know, Jeff, you've defended technology in the past against people who still do call.

Jeff Jarvis [00:27:30]:
I get in trouble for it.

Leo Laporte [00:27:32]:
Moral panic. You know, I just, I don't think we know. I think we're in an interesting time. And you know what it could be? It could be that it's terrible and.

Jeff Jarvis [00:27:40]:
We'Ve ruined my 401k. Knock wood is still.

Leo Laporte [00:27:43]:
I know. Good. I don't understand that at all.

Jeff Jarvis [00:27:46]:
No.

Leo Laporte [00:27:46]:
Did I think I didn't do it on this show. Did I mention the Jeff Bezos? I mentioned it on Twitter and Windows Weekly earlier on Bubbles.

Jeff Jarvis [00:27:58]:
No, you didn't mention that. We didn't talk about it.

Leo Laporte [00:28:00]:
So he, he was at a Italian conference and was being interviewed about Bubbles and let me see if I can.

Paris Martineau [00:28:11]:
Without context, it sounds a name. I think they're beautiful.

Leo Laporte [00:28:16]:
Bubbas. I've got Bubbas. He says we are in an industrial bubble. Let me see if I can find a video of it. Here it is. Okay. This is from Tech Week, an Italian tech conference. Let me turn.

Leo's Laptop Audio [00:28:33]:
The company gets funded. The good idea.

Leo Laporte [00:28:35]:
Okay, let me go back and start over. Is.

Leo's Laptop Audio [00:28:36]:
And the bad ideas and investors have a hard time in the middle of this excitement distinguishing between the good ideas and the bad ideas. And so that's also probably happening today. But it doesn't mean that anything that's happening isn't real. AI is real and it is going to change every industry. In fact, it's a very unusual technology in that regard and that it's a horizontal enabling layer. Today we talk about AI first companies like OpenAI and Anthropic and Mistral and so on and so on and so on. There are so many startup companies that are kind of AI companies of various kinds and that's, that's normal for this phase. But that is not the biggest impact that AI is going to have.

Leo's Laptop Audio [00:29:20]:
The biggest impact that AI is going to have is it is going to affect every company in the world.

Leo Laporte [00:29:26]:
Let me see if I can find.

Leo's Laptop Audio [00:29:27]:
We think of bubbles, we think of valuations and market caps and things like this and how many billions of dollars are being invested in these six people at a $20 billion valuation even though they just started yesterday.

Jeff Jarvis [00:29:40]:
Right.

Leo's Laptop Audio [00:29:41]:
That's, that's very unusual behavior this. Investors don't usually give a team of six people a couple of billion dollars with no product. It's rare and that's happening today. But the great thing about industrial bubbles, this is a kind of industrial bubble as opposed to financial bubbles. And I'll tell you what I mean by that. If you go back like the, the 90s had a biotech bubble and there were a bunch of pharma startups, companies that were designing drugs and using new techniques and the world got very excited, the investment world got very excited. As a group they all lost money. But we did get a couple of life saving drugs.

Leo's Laptop Audio [00:30:17]:
A bubble, like a banking bubble, the crisis in the banking system. That's just bad. That's like 2008. And so that's those bubbles society wants to avoid, the ones that are industrial are not nearly as bad. It could even be good because when the dust settles and you see who are the winners, society benefits from those inventions.

Leo Laporte [00:30:40]:
So there have been numbers of industrial bubbles. The railroad bubble, the collapse in the 1890s, all the railroad companies went bankrupt. But we got a transcontinental railroad. The Internet bubble that burst in 1999, 2000. Yeah. A lot of companies went out of business. Pets.com disappeared. But we got a lot of dark fiber which now is being lit up.

Leo Laporte [00:31:02]:
We got a lot of infrastructure out of it. And in fact, today you can buy kitty litter online. It is a viable.

Jeff Jarvis [00:31:08]:
Bill Gross was right. He was. Bill Gross makes this point. It's all timing. Yeah, best.com was him.

Leo Laporte [00:31:14]:
But financial bubbles, which are bad for everybody, everybody loses money. This is not one of those. I think he's right and I think Jeff's a smart guy.

Jeff Jarvis [00:31:22]:
Whatever he thinks that's the tragedy of this. He is a smart guy, but it's. What's happened to him?

Leo Laporte [00:31:27]:
Yeah, well, I don't know but I thought that was quite a good insight and I'm going to give him credit for it. I've kind of Talked around that in the past. I think this is one of those situations where, yes, a lot of companies are going to go out of business. Maybe OpenAI won't make it, but we already have benefit from it. Right now there is a debate about what the economic impact will be. There's probably very little debate about what the environmental impact is, although that means there's a lot of incentive, a lot of pressure to come up with cheaper ways to create energy. You know, we know we have ways that we can create renewable energy at a lower cost. In fact, renewables are taking off in the United States.

Leo Laporte [00:32:07]:
I don't think it's because of the AI boom, but maybe it is. And in the long run, maybe that's what will happen. So, you know, the Senate, the Senate Democrats have released a report which I don't know how credible this is. This is from the Hill, that almost 100 million jobs could be lost to AI over the next 10 years. How many jobs are there in the U.S. that sounds like a lot of.

Jeff Jarvis [00:32:34]:
We don't know anymore because we don't have anybody to count them, but that's enough.

Leo Laporte [00:32:38]:
I don't know if I buy that. Do you think that's true? 100 million jobs.

Jeff Jarvis [00:32:41]:
What do you think, Paris?

Paris Martineau [00:32:42]:
Can't be true. Seems like a large trapper trying to get Nano banana to create an image of you and Jeff Bezos blowing bubbles. And it took a couple more tries, but I got there.

Leo Laporte [00:32:54]:
Can we be on the beach? All right.

Paris Martineau [00:32:56]:
We. Oh, I, I.

Leo Laporte [00:32:58]:
Yes, we do. We have it.

Paris Martineau [00:32:59]:
There we are, the first three prompts.

Leo Laporte [00:33:01]:
That's pretty good, actually.

Paris Martineau [00:33:03]:
That image.

Leo Laporte [00:33:04]:
That's pretty good. And he's making, he's doing the. Jeff laughed.

Paris Martineau [00:33:08]:
His voice in that interview sounded very odd. Listen to a lot of.

Leo Laporte [00:33:15]:
Jesus.

Paris Martineau [00:33:15]:
Jeff Bezos.

Jeff Jarvis [00:33:17]:
I've been in the room with him.

Paris Martineau [00:33:19]:
Him.

Jeff Jarvis [00:33:19]:
I've been in a small room with him. When he laughs loud. It's Memorex.

Leo Laporte [00:33:23]:
Yeah.

Paris Martineau [00:33:24]:
So, yeah, he is a famously loud and, oh, yes, unnerving laugh.

Leo Laporte [00:33:28]:
The Democrats say we're charming jobs, but I would trust Yale. Yale says the AI has essentially zero impact on jobs.

Paris Martineau [00:33:39]:
Those are two different things. That the first estimate is potential jobs lost. The current is if AI becomes what everybody is talking about, the current, the current reality of it. I think that is reasonable to me. I'd actually.

Leo Laporte [00:33:53]:
What's the difference? Yeah, the Democrats talk about the next 10 years. Yale was saying what happened in the past 33 months. Yeah, yeah.

Paris Martineau [00:34:00]:
But also, I don't know, just the study. I could see this study Intuitively making sense. If you're talking about do the current AI tools have the capacity to fully replace human workers? In practice, the answer would probably be no. But I would argue that there have been many jobs currently lost due to AI because people assume that these tools have the capacity that they do not.

Leo Laporte [00:34:26]:
Right. As Cory Doctorow said, AI can't do your job. But there are a lot of salesmen for AI companies who are selling your company on AI that they say can do the job. And you, your company won't know that they can't do the job until after you fired everybody who could. Yes, the Democrats say the workforce most impacted will be fast food and counter employees.

Jeff Jarvis [00:34:51]:
But that doesn't make a hell of a lot of sense. I mean, McDonald's has already gotten rid of some people because you order on a screen. But somebody's gotta fry the burger.

Leo Laporte [00:34:59]:
Somebody's gotta fry em. I'm sorry, I burned my fingertips flipping.

Paris Martineau [00:35:02]:
The hamburgers in the bathroom.

Leo Laporte [00:35:04]:
Exactly.

Jeff Jarvis [00:35:06]:
And by the way, something there.

Leo Laporte [00:35:08]:
If we lost fast food jobs, I.

Paris Martineau [00:35:09]:
Don'T know if we're losing fear. I fear when I go into McDonald's bathroom that I will encounter straight vomit.

Leo Laporte [00:35:16]:
I did see a picture, probably on some social network of a Waymo. Your Waymo ride is ready. Somebody got in and there was McDonald's fries all over the floor.

Paris Martineau [00:35:27]:
And you just assumed that picture was true? Leo, you didn't think that we were living in a post truth society?

Leo Laporte [00:35:32]:
You know how I know it was true? Because I'm sure that's exactly what happens. Because you know, look, you've been in a taxi cab late on a Saturday night.

Paris Martineau [00:35:41]:
Yeah. Not in a cab. I vomited out of a cab window before.

Leo Laporte [00:35:47]:
Because you had the forethought to aim. But not everyone does.

Jeff Jarvis [00:35:51]:
What did the cab driver do?

Leo Laporte [00:35:53]:
That's the difference is the cab driver, I mean, clean it up before the next ride.

Paris Martineau [00:35:56]:
I mean the person who vomited in a cab, I'm probably not in the best position of authority to recall the details. Not a reliable, I'm not a reliable Source. It was 10 years ago.

Leo Laporte [00:36:08]:
I am 16. Okay.

Paris Martineau [00:36:10]:
I was, I was, you know, like 20, 21. And I was. Drank too much alcohol, vomited out the.

Leo Laporte [00:36:20]:
Window illegally, I might add.

Paris Martineau [00:36:22]:
This was, I guess, 21. I was, yeah, I was illegal. I was 21 then. Then I, it was raining and so I also aimed so that none of it got in the side of the cab. The cab driver pulled over, I was like, get out. And I was like, that's fair. And then I Got.

Leo Laporte [00:36:39]:
That's fair. I can't say enough.

Paris Martineau [00:36:42]:
This is. That's. We'll do exactly one more. I've only had two notable vomits in New York. It's that and while walking back to my NYU dorm.

Leo Laporte [00:36:58]:
You don't have to confess this, Paris.

Paris Martineau [00:37:01]:
I do. You're forcing me. You're holding me at gunpoint. I was drunk and I saw, like, what I thought was a great into, you know, the sewer. And I was like, great. Perfect place to vomit.

Jeff Jarvis [00:37:11]:
Oh, no.

Paris Martineau [00:37:11]:
Instead, it was a great intellect, an AC unit.

Leo Laporte [00:37:16]:
Oh, no. And I really still smells bad.

Paris Martineau [00:37:20]:
And I do think about that one. And I regret it for whoever had to clean that up. I'm so sorry.

Leo Laporte [00:37:26]:
Oh, gosh.

Jeff Jarvis [00:37:26]:
Do you ever walk by that place and just think, I'm sorry?

Paris Martineau [00:37:30]:
I do, actually. It's near Washington Square park. And I do think about that when I walk.

Leo Laporte [00:37:35]:
No, anybody who has a great anywhere near Washington Square park deserves.

Paris Martineau [00:37:39]:
Whatever I was about to say. Having a great outside of an NYU dorm for. In Washington Square Park. You're kind of asking for that.

Leo Laporte [00:37:47]:
Other workforces, according to the Democrats, that will be significantly affected customer service representatives. I agree with that. That's already happening. Right? Half the time you're not getting a human laborers. What? No. AI can't do labor. Contrary freight, stock and material movers. Well, they must be thinking of robots.

Jeff Jarvis [00:38:09]:
They're thinking of robots. They're thinking of robots, but that's not AI.

Paris Martineau [00:38:12]:
They're thinking of. Of intelligent machines.

Leo Laporte [00:38:14]:
Yeah. And it's true. What is? Amazon's like 70% robots in some of its warehouses. Now, legal, medical and executive position. Oh, I'm sorry. Secretaries and executive assistants. Not including legal, medical and executive positions. 80% of secretaries and executive assistants will be replaced in the next decade.

Paris Martineau [00:38:35]:
Crazy how it's all the lower.

Jeff Jarvis [00:38:38]:
Well, but I put a story up.

Paris Martineau [00:38:39]:
But not the executive venture capitalists. Huh?

Jeff Jarvis [00:38:43]:
Now, now, now, Paris, before you speak too quickly, I put a story up on the rundown, I don't know what line it was, where Sam Altman himself, CEO, says CEOs are going to be replaced.

Leo Laporte [00:38:52]:
Sam is really good.

Jeff Jarvis [00:38:54]:
Oh, yeah.

Leo Laporte [00:38:55]:
He's at like. Yeah, it's a deflection. It's a bubble. Yeah, yeah, yeah. It's not.

Jeff Jarvis [00:39:00]:
You should regulate us. Yeah.

Leo Laporte [00:39:01]:
Burning money. Yeah, you should regulate us. These. Really? One other problem that they're admitting to is that Johnny Ives, AI hardware collaboration with OpenAI is delayed. As somebody in MacBreak Weekly pointed out, Johnny's a great designer. He's never made a piece of hardware a Product itself. I've says I've come up with 15 or 20 really compelling ideas. That might be the problem right there.

Leo Laporte [00:39:29]:
One would be good. So they haven't quite figured out the form factor. Sam Altman says is not a pair of glasses. Apple, by the way, has apparently decided to back down on its Vision Pro and is going to do glasses. They saw what Meta is doing, said, oh, we better get going.

Jeff Jarvis [00:39:48]:
Oh, wow.

Leo Laporte [00:39:48]:
Yeah. The Altman and Ives say it's not a form factor we've seen yet.

Jeff Jarvis [00:39:59]:
Neither have they.

Leo Laporte [00:40:00]:
Yeah, nobody's seen it, to be frank. All right, let's take a break. Then we'll read line 84. Sorry, I didn't mean to.

Jeff Jarvis [00:40:07]:
That was. That was. That was what you already talked about. It was okay. That was Sam Altman saying, Sam Altman, it's coming after me too, so I'm.

Leo Laporte [00:40:13]:
Gonna lose my job too.

Paris Martineau [00:40:15]:
So I'm just like you.

Leo Laporte [00:40:17]:
Yeah.

Paris Martineau [00:40:18]:
I have job.

Leo Laporte [00:40:19]:
I'm a human, honest.

Paris Martineau [00:40:21]:
I'm not just a bundle of machine parts wrapped in human flesh.

Leo Laporte [00:40:26]:
I think you could make a compelling case that Sam has been brought to us by an alien technology to disrupt humanity. I think.

Jeff Jarvis [00:40:34]:
Who's the most alien like tech to?

Leo Laporte [00:40:40]:
Mark Zuckerberg. They're both.

Paris Martineau [00:40:42]:
They both. I was going to say they both have a waxy complexion that looks like kind of like a spongy like. Like a fondant that's been let laid out to dry for a long time, but then spritzed with a little water bottle and then dabbed the sponge. I don't know why I've thought about.

Jeff Jarvis [00:40:57]:
You've been giving thought to this, haven't you?

Paris Martineau [00:40:59]:
But it is something I have have thought about.

Leo Laporte [00:41:01]:
That's exactly what is alien. Synthetic human skin would look like.

Jeff Jarvis [00:41:05]:
We know that. Yes.

Leo Laporte [00:41:08]:
All right.

Jeff Jarvis [00:41:08]:
Not musk. You don't think musk is more of alien?

Leo Laporte [00:41:11]:
No, musk is very human. He looks blobby.

Benito Gonzalez [00:41:17]:
Musk isn't. As an apartheid, you know, a guy so like, that makes sense what he is.

Leo Laporte [00:41:21]:
Yeah, yeah. He's. He looks like he's lived through some hard times.

Benito Gonzalez [00:41:25]:
Like Zuck. Zuck's personality doesn't really make sense.

Jeff Jarvis [00:41:28]:
Sense, you know.

Leo Laporte [00:41:29]:
No, it makes no sense.

Jeff Jarvis [00:41:30]:
Peter Thiel.

Leo Laporte [00:41:30]:
Sams. Peter Thiel does not look real. Yeah, but that's because he's got a blood bag attached. As he walks around, there's something going on there. He's got a young person in the basement or something. There's something going. I mean, weird going on with Teal. Or maybe he's just.

Leo Laporte [00:41:46]:
I Don't know. Maybe he's got, I don't know, belt sander.

Jeff Jarvis [00:41:49]:
He's joking. Don't sue us out of. Don't send Hulk Hogan getting more. Whoever. Whoever follows.

Leo Laporte [00:41:53]:
Oh, yeah, that's right. Peter is a litigious.

Jeff Jarvis [00:41:56]:
Oh, yeah.

Leo Laporte [00:41:57]:
Oh, well, wait a minute. Don't send Hulk Hogan after me. Guess you can't anymore. We were going to take a break and talk more about AI, including a little mistake that Deloitte made. Whoopsies. But it's a good story. But first, a word from our sponsor. You're watching Intelligent Machines with Paris Martineau and Jeff Jarvis.

Leo Laporte [00:42:21]:
Glad you're here. Let me tell you about our sponsor for this segment, Zapier. Z A P I E R. I use Zapier to pair the shows. Those rundowns we put together. I know Jeff and Paris do it manually, but I actually have a bookmarking tool, Raindrop, that I use that. One of the things I love about Zapier is it integrates with thousands of the apps you already use and all of the workflow you already use and all the cloud services and all the AIs and everything it's all integrated in so you can tie them together. So right now, Raindrop, when I bookmark something on Raindrop, it automatically goes to my Mastodon, Twit Social.

Leo Laporte [00:42:57]:
You can follow, you know, Twit News or something like that. That's the handle, I think, twittnews, and see what I'm posting, what I'm thinking about for the shows. And then it automatically goes and formats it into a Google spreadsheet, adds a line to a Google spreadsheet, which our team, Benito and all the other producers will use to create the rundowns for the shows. Zapier automates this, and we've been using this for years. Actually, it was Carson Bondi who said, you know, we could do this thing with a spreadsheet. And I said, oh, I can do it with Zapier. I love Zapier. Well, here's the thing about Zapier.

Leo Laporte [00:43:29]:
They have now added AI and to great advantage. You know, of course, we're talking all the time about AI on the show, but talking about trends doesn't help you be more efficient at work. I said earlier on Windows Weekly, sometimes I look at these AI things and I feel like an idiot. Like I don't know what. Where to begin. Oh, that's what Zapier solves. You. You need the right tool.

Leo Laporte [00:43:50]:
You need Zapier. It's how you break the hype. Cycle and put AI to work. Really useful work across your company. So what is Zapier? Zapier is how you can actually deliver on your AI strategy, not just talk big words about it. With Zapier's AI orchestration, and that's the key AI orchestration platform, you can bring the power of AI to any workflow. So now you can do more of what matters. You can connect all the top AI models, ChatGPT, Claude and so on to the tools you're already using.

Leo Laporte [00:44:22]:
So, for instance, I could take that workflow, they call them zaps. I could take my zap and I can now say, hey, ChatGPT, summarize that story in two lines. So I could put that in the spreadsheet, things like that. Fantastic. You can add AI wherever you need it. And that means if it's an AI powered workflow, an autonomous agent, very powerful with customer chatbots. There's quite a few of those that you can connect up with on Zapier. Really? You can orchestrate almost anything with AI plus Zapier.

Leo Laporte [00:44:54]:
Zapier is, by the way, for everyone. I promise you won't feel like an idiot when you're looking at Zapier. You don't have to be a tech expert. You can do amazing things. And as proof, teams have already automated over 300 million AI tasks on Zapier. Join the millions of businesses transforming how they work with Zapier and AI. Get started for free by visiting Zapier.com machines. That's Zapier.

Leo Laporte [00:45:23]:
Z-A P I E R.com machines.

Jeff Jarvis [00:45:29]:
Leo, before you leave the commercial, may I?

Leo Laporte [00:45:31]:
Yeah.

Jeff Jarvis [00:45:32]:
So I told you earlier today about Joe Amdis, a former student of mine who's a Montclair state, who's helping local news organizations, right, with AI so that he can automate all kinds of things he's doing. We're talking about trying to make towns more transparent. He swears by Zapier. Yeah, swears by. It enables him to do the things.

Leo Laporte [00:45:49]:
Such a great job before it's such a great tool. Well, thank you for that unsolicited plug. Zapier.com machines. You should check it out for sure. Deloitte.

Jeff Jarvis [00:46:01]:
I think it should be pronounced Zapier.

Leo Laporte [00:46:03]:
Zapier, because he's like a rapier. Deloitte, which is one. It's like McKinsey, right? It's a consultancy firm. You bring them in to help you solve your problems. The Australian government brought in Deloitte, paid.

Jeff Jarvis [00:46:18]:
Them, that was their first mistake.

Leo Laporte [00:46:21]:
Paid him 440,000 Australian dollars. It was the Department of Employment and Workplace Relations. And in fact they produced a report for them which the department published on their website in July.

Jeff Jarvis [00:46:36]:
Second mistake.

Leo Laporte [00:46:39]:
At which point a Sydney University researcher on health and welfare law alerted the Media that the 237 page report was full of fabricated and references. Deloitte confirmed some footnotes and references were incorrect.

Jeff Jarvis [00:46:57]:
Some.

Leo Laporte [00:46:58]:
They used AI to write this thing. And it. There were fabricated quotes from a federal court judgment, references to non existent academic reference papers, research papers. Deloitte is going to refund only a small amount. The last payment to the Australian government. An apology. I think they should refund the whole thing. The report was reviewed departmental IT systems used of automated penalties in Australia's welfare system.

Leo Laporte [00:47:30]:
The department said the substance of the report is. Is accurate and we don't change our.

Jeff Jarvis [00:47:35]:
Recommendations because we're consultants and all we're going to recommend is for you to fire people.

Leo Laporte [00:47:40]:
Yeah.

Paris Martineau [00:47:40]:
Utterly ridiculous. I mean, it's only $290,000 for Deloitte. That's nothing. They should refund it to save face.

Leo Laporte [00:47:48]:
Yes.

Jeff Jarvis [00:47:49]:
Did you see the follow up on line 96?

Leo Laporte [00:47:52]:
Oh, no, there's more.

Jeff Jarvis [00:47:53]:
On top of this, there's a press release that comes out that maybe should have been delayed a little bit.

Leo Laporte [00:47:59]:
Oh.

Jeff Jarvis [00:48:00]:
Anthropic and Deloitte partner to build AI solutions for regulated industries.

Leo Laporte [00:48:07]:
Whoops. Yeah. This is from a third party company, right? That. That apparently is being used by Anthropic and Deloitte. Okay. Deloitte will make Claude available to all Deloitte employees across its global network and will establish a Claude center of Excellence.

Jeff Jarvis [00:48:26]:
Center of Correction, made up of trained.

Leo Laporte [00:48:29]:
Specialists who will develop implementation frameworks, share practices across deployments and provide ongoing technical support.

Jeff Jarvis [00:48:35]:
You know what, and how many fact.

Benito Gonzalez [00:48:36]:
Checkers are they employing?

Jeff Jarvis [00:48:38]:
I was just flashing on that. I think this is an entirely new way to employ fact checkers and copy editors. They're needed after all.

Leo Laporte [00:48:45]:
Yeah, just look up the records.

Paris Martineau [00:48:48]:
Do the Lord's work.

Jeff Jarvis [00:48:49]:
Yes, they do.

Paris Martineau [00:48:50]:
They will be so viciously angry at this.

Leo Laporte [00:48:54]:
Deloitte has an AI academy that will certify AI practitioners. They're going to launch a certification program for generative AI and advanced AI applications. Wow. Deloitte's putting $1.4 billion into this. Yikes.

Paris Martineau [00:49:15]:
And they couldn't give back Australia 290,000, right?

Jeff Jarvis [00:49:19]:
Right.

Benito Gonzalez [00:49:19]:
That's how they get $1.4 billion.

Leo Laporte [00:49:21]:
Yeah.

Benito Gonzalez [00:49:22]:
Never giving any back.

Leo Laporte [00:49:22]:
Where does that money come from?

Jeff Jarvis [00:49:24]:
Yeah, it was already spent.

Leo Laporte [00:49:26]:
The rich are cheaper than anybody. Yeah, let's see. America according to The Financial Times. Which is British, right? I say America is now one big bet on AI. It's seen as the magic fix for every threat to the US economy.

Paris Martineau [00:49:49]:
I like your British accent. It's very much the guy playing the king in Hamilton.

Leo Laporte [00:49:56]:
Yes, that's exactly right.

Jeff Jarvis [00:49:57]:
Despite.

Leo Laporte [00:49:58]:
Despite mounting threats to the US economy, from high tariffs to collapsing immigration, eroding institutions, rising debt and sticky inflation. You'll be back. Wait and see. Just remember, you belong to me. Large companies. Investors seem unfazed. They're increasingly confident that artificial intelligence is such a big force it can counter all the challenges. Wow.

Leo Laporte [00:50:26]:
This is an opinion piece. AI companies. This is scary, though. Have accounted for 80% of the gains in US stocks this year. 80%.

Jeff Jarvis [00:50:37]:
There's even more frightening numbers, is that it counts for four. Data centers now account for 4% of GDP.

Leo Laporte [00:50:46]:
Wow.

Jeff Jarvis [00:50:47]:
And where did I put this? They account for something like. Where's that number? Here we go. Line 105. A Harvard economist tells us that it's responsible for 90.

Leo Laporte [00:51:04]:
Trust. Harvard. I don't.

Jeff Jarvis [00:51:05]:
I would hope so. 92% of GDP growth in the first half of the year is attributable to information processing equipment and software, that is to say, data centers.

Leo Laporte [00:51:16]:
Got to remind you, though, I mean, that's not just AI.

Jeff Jarvis [00:51:19]:
Not just AI, but it's.

Leo Laporte [00:51:20]:
What's the major part of that going on. There's a lot of cloud stuff. In other.

Jeff Jarvis [00:51:24]:
Just try to build a data center for accounting now.

Leo Laporte [00:51:33]:
But again, I think this is an example of an industrial bubble. And these data centers aren't going to evaporate when the market crashes. They're still going to be there.

Jeff Jarvis [00:51:43]:
Well, or. So there's another. Do you have a subscription to the FT Leo?

Leo Laporte [00:51:50]:
Yes.

Jeff Jarvis [00:51:50]:
Okay, so I'm gonna put something in the chat right now that's exactly related to what you just said. So. Hold on. I meant to put it in because.

Leo Laporte [00:51:55]:
I believe we all need more information from the United.

Jeff Jarvis [00:51:59]:
So the flawed Silicon Valley consensus on A.I.

Leo Laporte [00:52:04]:
This is John.

Paris Martineau [00:52:05]:
Serious questions remain among us. What will happen next if we do or do not read the lead out loud?

Jeff Jarvis [00:52:12]:
The lead out loud.

Leo Laporte [00:52:12]:
Wild wants to find fox hunting as the unspeakable in pursuit of the uneatable.

Paris Martineau [00:52:19]:
Jesus Christ. Are you British or are you British?

Jeff Jarvis [00:52:22]:
Keep going. It's beautiful. Beautiful.

Leo Laporte [00:52:24]:
Were he alive today, he might describe the quest for artificial general intelligence as the unfathomable. In pursuit of the indefinable. Oh, die. Witty, isn't that? Yes. Many hundreds.

Paris Martineau [00:52:39]:
You know, someone was drinking tea when they were.

Leo Laporte [00:52:42]:
Oh, that was a good one, wasn't it? Yes, it was. You've done it. You've done it again.

Jeff Jarvis [00:52:49]:
Plus the scone and the clotted cream.

Leo Laporte [00:52:53]:
Instead of hyperventilating that AI, ushering in a new era of abundance, wouldn't it be better to drop the rhetoric and build AI systems for more defined, realizable goals? Spell that with an S, please, because I am British. But we should stop asking, is the machine in intelligent? And ask, what exactly does the machine do?

Jeff Jarvis [00:53:16]:
Do?

Paris Martineau [00:53:17]:
What does it do?

Leo Laporte [00:53:18]:
I'm sorry, that was a Shannon Valor, a professor at the University of Edinburgh. So we should. I'm gonna switch to Scottish. We should stop asking, is the machine intelligent? And ask it, what exactly does the machine do? It's not exactly scary.

Jeff Jarvis [00:53:37]:
This makes me very happy.

Leo Laporte [00:53:38]:
I don't know what. I don't know what that was. Anyway. Anyway. Yeah, AGI. We don't. We don't know what AGI is.

Jeff Jarvis [00:53:46]:
We don't know. That's the wrong goal.

Leo Laporte [00:53:48]:
The Silicon Valley consensus, he points out, is that it's within reach, whatever it might be, within the decade.

Jeff Jarvis [00:53:56]:
And the problem is, he goes on to quote Yudkovsky, who's in nutballs.

Leo Laporte [00:53:59]:
So, yeah, he's. Eliezer Yudkovsky is a famous naysayer. Right.

Benito Gonzalez [00:54:05]:
It is starting to sound a lot like fusion, though. Like, you know, in 20 years we'll have fusion. And they've been saying, yeah.

Leo Laporte [00:54:11]:
Or quantum computing or. Yeah, yeah. But I don't think that that's the point. I mean, I think that he even kind of said this, we shouldn't. Or. She said this, Shannon. She said, what does it do? Not what is it? Intelligent.

Jeff Jarvis [00:54:23]:
Which is your point, actually. You say that often.

Leo Laporte [00:54:25]:
Yeah.

Jeff Jarvis [00:54:26]:
Not as mellifluously as she did, but.

Leo Laporte [00:54:30]:
AI may be great. But what does it do?

Jeff Jarvis [00:54:37]:
I think Paris for the 24 hour. There has to be some accent challenged. Leo.

Leo Laporte [00:54:44]:
I think.

Paris Martineau [00:54:45]:
Yeah, we could kind of have a. Let's see, we could have a. Like a big wheel that we spin at a random time and it picks.

Jeff Jarvis [00:54:52]:
What accents he has to read a press release in. Whatever the accent is that comes up.

Leo Laporte [00:54:58]:
Yes, that could be good.

Benito Gonzalez [00:54:59]:
That can get problematic.

Paris Martineau [00:55:00]:
Couple minutes.

Leo Laporte [00:55:01]:
The. Here's from Martin Pierce, former. Yes, yes. Well, yes. And that's a problem.

Jeff Jarvis [00:55:06]:
Right.

Leo Laporte [00:55:07]:
We can't. I can't do half the accents I used to do because they're racist and demeaning.

Paris Martineau [00:55:13]:
That's true.

Leo Laporte [00:55:13]:
I guess. They're all. I mean, honestly, if you were Scott and I did that, you would be mad, too. This is from your former peer at the Information. Martin Piers, the AI prophet fantasy.

Paris Martineau [00:55:25]:
I can't read it, so I'm interested to hear.

Leo Laporte [00:55:27]:
Oh, did they kick you out? Well, well, well. Yeah, yeah, yeah, yeah.

Jeff Jarvis [00:55:31]:
My cancellation hasn't. Date hasn't arrived yet, so.

Leo Laporte [00:55:34]:
Neither has mine. I bought it it for a year, so I still have that's the same access to that.

Jeff Jarvis [00:55:39]:
As soon as you left Paris, I canceled.

Leo Laporte [00:55:41]:
Yeah, me too. I did the same thing.

Paris Martineau [00:55:43]:
You don't need to do that.

Leo Laporte [00:55:44]:
It was purely symbolic since I could still read.

Jeff Jarvis [00:55:47]:
I have another 40 days.

Leo Laporte [00:55:48]:
It told me recently, in the years to come, we might look back at this period as in tech, as one in which the entire industry was in the grip of a mass delusion. Namely.

Paris Martineau [00:56:00]:
Yeah, Martin is Australian, if that factors into your head.

Jeff Jarvis [00:56:03]:
Oh, okay.

Leo Laporte [00:56:08]:
You know that I can feel, I can feel the people on the other side of the microphone going, no.

Jeff Jarvis [00:56:17]:
We'Re not saying that.

Leo Laporte [00:56:18]:
Of course it's still early days, but so far AI has proven lucrative primarily for the companies involved in making and selling AI chips. My problem is my Australian verges into companies cockney and I don't, I don't know how to say it.

Paris Martineau [00:56:32]:
I thought it was pretty good. It doesn't sound like Martin, but it was a pretty good Australian accent.

Leo Laporte [00:56:36]:
AI is a money pit. We previously knew that it was true of OpenAI, but now of course he's talking about the Oracle example. But you know what? They still. I, this is why I was puzzled. Okay, sure, the margins were only 14% instead of Oracle's usually rapacious 70% margin.

Jeff Jarvis [00:56:59]:
But that was the market money.

Leo Laporte [00:57:02]:
It's still profit. It did send the stock down 5%.

Jeff Jarvis [00:57:06]:
Well, we're investing a huge amount in these data centers thinking that it's an operational business.

Leo Laporte [00:57:11]:
But they're not losing money, they're just.

Paris Martineau [00:57:14]:
Not making money as fast as they were before. And that's unacceptable in these times.

Leo Laporte [00:57:20]:
Yeah, that's, I think, a part of the problem.

Jeff Jarvis [00:57:21]:
And in fairness, we have no idea what the business model is yet. I go back to in the 1880s, the publisher of the New York Tribune thought that we would have only so many ads you could put in a newspaper, otherwise people would revolt and not buy it. And ads would not support newspapers going forward. When radio began, no one, and I mean no one thought it was an ad business. Sarnoff thought it was going to be the radio manufacturers who were going to be paying into a fund fund to pay for the entertainment to make people buy the radios. Others thought, well, they're gonna. What happens when they hit saturation well, then this sounds very much like the news business. Well, then it's probably going to be either a tax or it's going to be an endowment.

Jeff Jarvis [00:58:07]:
Advertising. No, it should be banned. Advertising is irritating. It shouldn't be on radio. No advertising on radio.

Leo Laporte [00:58:13]:
Boy.

Jeff Jarvis [00:58:14]:
And so it was a surprise.

Leo Laporte [00:58:15]:
Only they had succeeded.

Jeff Jarvis [00:58:16]:
Yeah, well, there was.

Leo Laporte [00:58:19]:
There was the same reason we have advertising on our podcast. It's one of the only viable ways to finance this. I mean, I guess we could insist that everybody pay for the podcast, but we would never have grown to the size that we grew to.

Jeff Jarvis [00:58:33]:
So. So what's the AI business model then? And. And it's a different answer at the foundation level.

Leo Laporte [00:58:39]:
I hope it's not advertising. I have to say, advertising, I think, is a cop out.

Jeff Jarvis [00:58:44]:
Well, so what do you think the model will be?

Leo Laporte [00:58:45]:
Well, I think that in general, you should pay for what you use instead of.

Paris Martineau [00:58:50]:
But people are never going to pay for the actual.

Leo Laporte [00:58:52]:
That's because we're cheap.

Paris Martineau [00:58:54]:
Yeah. And it's expensive and it's not going to work out. The.

Leo Laporte [00:58:58]:
We don't know how expensive it is. We really don't.

Paris Martineau [00:59:00]:
I mean, but all signs point to it being wildly more expensive. The fact that not even in the aggregate highest paid plans are not even breaking even is bad news for this.

Leo Laporte [00:59:13]:
But those people, by the way, the ones that use those hundred, two hundred dollars a month plans are use using it like crazy.

Paris Martineau [00:59:20]:
It's not a good sign for the health of your business if the people who want to pay for your highest tier of product have usage patterns that are not compatible.

Leo Laporte [00:59:30]:
Kind of a good sign.

Paris Martineau [00:59:32]:
No, it's not.

Leo Laporte [00:59:33]:
It's only a good sign in the.

Paris Martineau [00:59:35]:
Topsy turvy upside down world. We're making money. Doesn't matter. Capitalism. Yeah.

Leo Laporte [00:59:41]:
Bring this down to something I can understand. Understand. Let's say you've got a sandwich shop very near the New York University campus where there are lots and next to.

Jeff Jarvis [00:59:55]:
A famous pizza place and next to.

Leo Laporte [00:59:56]:
Probably the best pizza place in the city. And you're making a sandwich and you decide, I'm going to take the best damn ingredients I can find, cost be damned, and you make the sandwich. And the Sandwich costs you $30 to make, and you have to sell it for a price people will pay, and it turns out they'll only pay $28. So you're losing two bucks a sandwich. But the good news is you're selling every sandwich you can make.

Jeff Jarvis [01:00:24]:
You'll make it up on volume.

Paris Martineau [01:00:26]:
That's not how that works.

Leo Laporte [01:00:28]:
Well, and that's why Hank raised it to 32 bucks. But there was a benefit to him because there are still lines out front for the $32 version of the sandwich.

Jeff Jarvis [01:00:39]:
In fact, probably more so. I would.

Leo Laporte [01:00:41]:
More so. And he, you know, he got all these great reviews. He got all this attention. Now, I don't know if he'd lost money. I don't think. I think he mostly broke even. But he wasn't making. He said, I made $4 yesterday.

Leo Laporte [01:00:51]:
He told me one touch. But that's. He has overhead, right? That's not. That's his money, right? I mean, he had. He paid the staff, he pays the rent and pays all that. He buys all the product. The cost of goods was very close to the amount of money he was charging. And yet he built a business out of it.

Leo Laporte [01:01:09]:
And so I think that's, you know, not unheard of. Look at Uber. I mean, Uber's made no money in 10 years, but.

Jeff Jarvis [01:01:19]:
Well, Hank could have a business model where he breaks even on the restaurant, but makes it on books, incidentally.

Leo Laporte [01:01:25]:
That's what's been going on, right? He doesn't. He makes $4 on the restaurant. Everybody else gets paid and paid well. He pays above market rate for his team, but he makes money in other things because he's a TikTok star. And so he's got sponsors like crazy. He's got Pepsi. He's got Hellman's Mayonnaise. He's got Better Than Bullion.

Leo Laporte [01:01:47]:
He's got, you know, all of these, you know, beers, Stella Artois and tequilas. So he's. Hank's fine. He's doing fine. In other words, I think. And by the way, I don't think that's the model for OpenAI. I just think that. Or Uber.

Leo Laporte [01:02:06]:
But it is kind of a Silicon Valley model where you build a business and you build customers and you demonstrate demand for your product before you try to make money.

Jeff Jarvis [01:02:15]:
Right? But I think we got to look at. At the stack, and the model is going to be different at each. So let's just imagine for a second. Let's start here and imagine that at some point you have to pay for the tokens you use for whatever you do from the models. Okay? That's going to reduce demand considerably. A B. The models right now are commodified as. As Rise of.

Leo Laporte [01:02:41]:
I don't think they're truly commodified. I think what they are is each of them has a slight advantage in certain. Certain areas.

Jeff Jarvis [01:02:47]:
There's no one model that doesn't leapfrog each other rises said, I can switch in a half an hour. There's no switching cost. Right, Right.

Leo Laporte [01:02:53]:
But they're probably for hawks. There probably is one model right now that's the best.

Jeff Jarvis [01:02:58]:
That's the best right now. Right?

Leo Laporte [01:02:59]:
Yeah.

Jeff Jarvis [01:02:59]:
So then the next question becomes, what about hawks? What about the application layer? That's where you touch the consumers in the long run, that's where you'll be more specialized.

Leo Laporte [01:03:09]:
True.

Jeff Jarvis [01:03:09]:
And that's where I think there's going to be different models. Some will be canva. You're paying to use it to make what you want to make and you're paying for it as a service.

Leo Laporte [01:03:20]:
Right. Look at half of our advertisers are now incorporating AI into their product that they're selling. And I guarantee you all of these.

Jeff Jarvis [01:03:28]:
Companies are profitable and they're all paying the AI companies and they're probably paying.

Leo Laporte [01:03:33]:
What the AI is worth.

Jeff Jarvis [01:03:34]:
So the AI is probably, to me a wholesale business. And the application layer is the retail business. The retail business can be something supported by consumer revenue or by advertising.

Leo Laporte [01:03:44]:
I think that's why you see Chap GPT say, or OpenAI, say we're going to be an operating system, because that's the model for oper. That's Microsoft's model. Right. Microsoft sells very few copies of Windows to end users. So that makes perfect sense, I think, a platform, if you become a platform. So I think this is encouraging. At least OpenAI is looking at more than just. Just let's stick some ads in the feed.

Leo Laporte [01:04:09]:
I think that's the worst thing they could do. But there are other ways to make it make sense. I think it's possible. And I think our best minds and some of our best AIs are working on it.

Jeff Jarvis [01:04:20]:
Well, the other interesting part of this then is you have media companies who are now saying, you know, give us a share of all your money. Well, the money is not profit. The money is venture capital. The money's not operating money. And what's what. What is what will happen at point.

Leo Laporte [01:04:33]:
Some.

Jeff Jarvis [01:04:33]:
Some point, I think is that there won't be these big deals I've said often on the show are ridiculous. They're buying silence for lobbying and for litigation. But what's going to happen in the long run, and we're seeing little bits of this start out is. And then Bill Gross started a company to try to do this is I'll pay you proportionally for the use of what it is based on what money I have. I'll pay you a rev share. That rev share is not going to be. Be a bucket of gold at the end of the rainbow. It's going to be small incremental revenue.

Jeff Jarvis [01:05:03]:
But the media companies who are involved here are going to have to say, well, I have to be there, so I'll take whatever pittance comes from the rev share because I'm not going to get discovered otherwise. They're not saying that today. They're now saying cut off. No. How dare you. I think people realize that they've got to be there and they're going to. And the AI companies aren't going to support them.

Leo Laporte [01:05:22]:
It is smart though, I think, for them to, to at least plant their flag now to say, look, you're not gonna.

Jeff Jarvis [01:05:28]:
So I didn't interview. I did an interview before we the show with a sweet. A really smart Swedish journalist for Shipstead. Shipstead's a very smart company at this. And, and I talked about my idea for the API for news. And she was going off on the Penske suit against one of the AI companies. She said, what would you do if you. If you were Jay Penske? If.

Jeff Jarvis [01:05:47]:
If only.

Leo Laporte [01:05:48]:
He owns. He owns what? He owns.

Jeff Jarvis [01:05:50]:
He owns Variety Building, Billboard, Rolling Stone.

Leo Laporte [01:05:53]:
I think, yeah, Rolling Stone, a whole bunch of stuff.

Jeff Jarvis [01:05:56]:
And he's done a very smart job of it. So he sued. I said, well, what I would do was that I would try to. Against my competitors. I would, instead of suing them and starting off that way, I'd say, what's the best deal I can make? And I would start developing databases that are useful to them. And I would start to say that my brand is useful to them. And I would go to them and say, okay, let's talk before my competitors do. Or I'd go to my competitors and I'd say, let's all gang up.

Jeff Jarvis [01:06:22]:
Let's get an antitrust lawyer in the room. So we're okay. And let's take an API. And I think there's an opportunity at a lower level of media to create APIs for news where you could gang together. But nobody's going to be saved by AI. Nobody's going to be filthy rich from it. That's what we've got to admit. Not in the media business.

Jeff Jarvis [01:06:43]:
The application layer will use it as a tool and will make money from it, but they've got to make money from it by having applications that are useful for consumers. And let's not forget, consumers can now oftentimes build their own applications.

Leo Laporte [01:06:57]:
In the words of Paul Simon, the words of the prophet were written on the subway walls. We'll talk about that in just a moment. You're watching intelligent machines with Paris Martineau.

Jeff Jarvis [01:07:12]:
That's a skill you get in radio, you know, you gotta work years, years and years to be able to do that. There's a magic behind there, folks, that you've got to appreciate.

Paris Martineau [01:07:26]:
Devno couldn't do that.

Leo Laporte [01:07:28]:
No, he tried. He tried. I want to talk about mcp. You know about mcp. You know about agents, right? Let me tell you about our sponsors for this segment on intelligent machines. Agency. Agency, okay. You can build the future of multi agent software with Agency Agntcy now, an open source Linux foundation project which I'm thrilled to see.

Leo Laporte [01:07:57]:
Agency is building the Internet of Agents, a collaboration layer where AI agents can discover, connect and work across any framework. All the pieces engineers need to deploy multi agent systems now belong to everyone who builds on agency, including robust identity and access management. Gotta have that. That ensures every agent is authenticated and trusted before interacting. Agency also provides open, love that word, open standardized tools for agent discovery, seamless protocols for agent to agent communication, and modular components for scalable workflows. And by the way, everybody's involved in this. You can collaborate with developers from Cisco, Dell Technologies, Google Cloud, Oracle, Red hat, in fact 75/ other supporting companies all dedicated to building the next gen AI infrastructure together. Open open agencies dropping code specs and services, no strings attached.

Leo Laporte [01:09:01]:
Visit agency.org to contribute. That's a G N T C Y.org agency. An open source collective building the Internet of agents and a really, really good thing. I think I was talking about the subway walls and actually this comes from a New York Times article. A debate about AI plays out.

Paris Martineau [01:09:29]:
Oh, we're talking about the Friend ads.

Leo Laporte [01:09:32]:
Ah, you've. Have you seen these in the, in the subway.

Paris Martineau [01:09:35]:
The founder of Friend, who we've discussed in the show before seems a little emotionally.

Leo Laporte [01:09:42]:
He's like you say 22 or something, right?

Paris Martineau [01:09:45]:
Yeah, he, I think wept during an interview with Wired reporters and insisted that they not unbox the friend in front of him because. Because he said he quote, recently fell in love and wants to see her be the first friend unboxing he ever witnesses. But he.

Leo Laporte [01:10:02]:
Victoria Song from the Verge, who was on our show a few weeks ago, has a friend and they did a big media blitz.

Paris Martineau [01:10:08]:
He's claimed that it's the largest single ad spend in New York City subway ads ever. I don't know the factual accuracy.

Jeff Jarvis [01:10:16]:
That was the first mistake.

Paris Martineau [01:10:18]:
But because guess what, I'm saying this because they are literally everywhere. They are in, in every subway Car.

Leo Laporte [01:10:24]:
I mean, it's a takeover. They bought a takeover.

Paris Martineau [01:10:26]:
But they're not particularly interesting. It's just the dumb friend pin thing that hangs in your neck and it just says, friend, I'll listen to you. Or friend, I'll never tell you to shut up. Or friend.

Leo Laporte [01:10:44]:
Which by the way, is not true.

Paris Martineau [01:10:45]:
Apparently for a bite of your food. Just like, who? Who cares?

Jeff Jarvis [01:10:50]:
So have you seen commentary on the ads in your exposure?

Paris Martineau [01:10:55]:
Oh, yes, they're almost everywhere. So this is.

Leo Laporte [01:10:59]:
If you put a billboard in the subway at eye level where people can write on it, you're just asking for trouble.

Paris Martineau [01:11:08]:
Well, most of the billboards or advertising in the subway are. No, they're eye level. That's. That's where most of the subway ads are. I mean, some of them are defaced, but the friend ads are overwhelmingly defaced. And coverage of this has made the problem even go up, where all over it people are writing, AI is not your friend. Talk to real people. Don't isolate.

Paris Martineau [01:11:32]:
Don't, you know, feed your thoughts to big tech. You can get help. All these.

Leo Laporte [01:11:37]:
There's actually a website now containing all of the defaced friend ads. So I should add to this. You can see them. I heart AI. I use ChatGPT every day. Well, wait a minute. That's not. That's worse than fai, which this person wrote, by the way.

Leo Laporte [01:11:58]:
Somebody tried to paint it out, but unsuccessfully.

Benito Gonzalez [01:12:02]:
This is basically subway wide caption contest.

Leo Laporte [01:12:07]:
Well, he did it. He paid a million dollars for a takeover. So it is everywhere. Human connection, sacred. AI doesn't care. AI is not your friend. Oh, this one got really trashed. AI chat who listens, responds and supports you.

Leo Laporte [01:12:21]:
AI wouldn't care.

Jeff Jarvis [01:12:23]:
It looks like then it's just tagged for the sake of the.

Paris Martineau [01:12:24]:
Yeah, there's often ones that say, AI wouldn't care if you lived or died. AI is burning down the planet down. Don't you care? Don't use this garbage.

Leo Laporte [01:12:34]:
There's one that doesn't give a shiz or a funk about you.

Paris Martineau [01:12:39]:
One of the friend dads say, I'll ride the subway with you. And it says, no, you won't. You'll probably a probability parrot is your friend.

Leo Laporte [01:12:52]:
Stochastic parrot.

Paris Martineau [01:12:54]:
There we go.

Leo Laporte [01:12:55]:
They didn't want to. They couldn't spell stochastic, so they said probably. Yeah, yeah. This is quite amazing. What's really amazing is that there is a website. Oh, here. They changed friend to foe.

Paris Martineau [01:13:06]:
Oh, that's fun.

Leo Laporte [01:13:07]:
That's.

Paris Martineau [01:13:08]:
The more AI you use, the dumber you get. One says, the more AI you use, the more it uses you. The smarter it gets, the more effed we all are as a planet, as a people.

Leo Laporte [01:13:19]:
Here's one.

Paris Martineau [01:13:19]:
Really gone ham on them.

Leo Laporte [01:13:20]:
Here's one my daughter could have written. I think this is. This is perfect. Put the phone. Phone down, beloved. With a little heart next to it.

Paris Martineau [01:13:27]:
AI sucks. Ew.

Leo Laporte [01:13:29]:
Ew. Wow, this is pretty. So you're basically saying all of them have been defaced.

Paris Martineau [01:13:39]:
Yeah, they're very overwhelmingly defaced in New York City.

Leo Laporte [01:13:46]:
Friend, Noun. Somebody who listens, responds and supports you. And then somebody's added. And it's a living being, not AI.

Jeff Jarvis [01:13:54]:
Very neatly.

Leo Laporte [01:13:55]:
Yes. Very good handwriting.

Jeff Jarvis [01:13:57]:
Very nice.

Leo Laporte [01:13:57]:
Excellent. Yes.

Paris Martineau [01:13:59]:
Wow. Yeah. Another one faced here. Friend. Someone who listens, responds and support you. And as a living being, don't use AI to cure your loneliness. Reach out into the world.

Leo Laporte [01:14:10]:
Don't be. I can't read what that word is. Be a Luddite. Yeah, be a Luddite.

Jeff Jarvis [01:14:16]:
Be a nerd.

Leo Laporte [01:14:16]:
No, don't be a nerd. Not a friend. AI wouldn't care if you lived or died. That seems to be a universal sentiment, that AI doesn't care.

Benito Gonzalez [01:14:26]:
Also true.

Paris Martineau [01:14:27]:
I mean, that's correct factually.

Leo Laporte [01:14:30]:
Yeah. Surveillance, Soulless crap. Death to AI Wow. This website, if you want to look at it, is NYC Friends Vercel app and it even has the address of each of these. So this is at West 4th street on the A, C, E, B, D, E, F and M lines.

Paris Martineau [01:14:55]:
Nice. This is the West 4th street station.

Jeff Jarvis [01:14:59]:
That is AI is evil. Evil.

Leo Laporte [01:15:03]:
Well, there you go. Oh, poisoning black communities. That's interesting. Foxtown data set.

Jeff Jarvis [01:15:13]:
Smoke stacks.

Leo Laporte [01:15:14]:
Yeah.

Jeff Jarvis [01:15:15]:
With carbon coming out of them.

Leo Laporte [01:15:16]:
Yeah, yeah. All right. Well, there you go. The words of the prophets are written on the subway walls.

Jeff Jarvis [01:15:24]:
As Paul Simon, I think you're Ari Melber.

Leo Laporte [01:15:28]:
Hey, it could be worse. I could be quoting rap. That's what he does, right? He's always. He's always quoting rap. So I don't know enough rap to do that. Insurers hesitate at multi billion dollar claims faced by OpenAI and Anthropic and AI lawsuits. Are they not going to pay the insurance?

Jeff Jarvis [01:15:48]:
Well, that's interesting. They see such huge liability, number one. Number two, what's happening so far is that the companies are using their venture money to pay off these problems that arise. So the cost of the insurance is going to be really, really high.

Leo Laporte [01:16:07]:
OpenAI has secured insurance of $300 million for emerging AI risks. $300 million worth of insurance, according to people familiar with the company's. Policy. Another person familiar with policy disputed that figure, saying it was much lower. But all agreed the amount fell far short of the coverage. To insure against potential losses from a series of multi billion dollar legal claims. Well, we know Anthropic had to pay one and a half billion dollars to the authors.

Jeff Jarvis [01:16:35]:
Right.

Leo Laporte [01:16:37]:
By the way, interesting. The judge who had originally said, I want to look at this because I think maybe the lawyers are getting too much of this. Has now approved that settlement and it's.

Jeff Jarvis [01:16:47]:
Half to publisher, half to author.

Leo Laporte [01:16:48]:
Okay, good.

Jeff Jarvis [01:16:49]:
So my agent sent me the form.

Leo Laporte [01:16:51]:
Do you get 1500 bucks supposedly per book?

Jeff Jarvis [01:16:55]:
Well, I think it depends on how much they end up paying up to how many people come in.

Leo Laporte [01:16:58]:
Right.

Jeff Jarvis [01:16:59]:
But they do. It's a finite number of books. So they know that I think.

Leo Laporte [01:17:02]:
Right. OpenAI is being sued for copyright infringement by the New York Times. Wrongful death by the parents of a 16 year old boy. OpenAI is considered self insurance according to these people with knowledge of the matter. In other words, using the VC to pay it. Wow, it's. I guess it's an interesting conundrum. Yep, there's lots of risks.

Leo Laporte [01:17:29]:
Meanwhile, Mr. Beast says, hey, knock it off. AI could threaten creators livelihoods. It's a scary time for the industry. You know, if you're Mr. Beast, I would be threatened because what you create is not particularly slop.

Jeff Jarvis [01:17:43]:
It creates human slop.

Leo Laporte [01:17:45]:
It's human slop.

Paris Martineau [01:17:46]:
It is a little sloppy.

Leo Laporte [01:17:48]:
Yeah, I don't. I mean I'm not worried about AI replicating this fine crafted podcast. No.

Jeff Jarvis [01:17:55]:
Oh, ours. I thought, well, Mr. Beast.

Leo Laporte [01:17:57]:
Mr. Beast is number one on the Forbes list of creators. I didn't even know Forbes did this. But they now have a list of creators. He has 85 million. Here's the. Oh God, look at this animated graphic of top creators. These are real, real life creators.

Leo Laporte [01:18:15]:
He is number one with 85 billion. Sorry. In earnings yearly. 634 million followers. What? Average engagement 1.39%. What is. I don't know what that means, but it seems low.

Paris Martineau [01:18:29]:
Well, 1.39% of 640 million is.

Leo Laporte [01:18:33]:
Oh, it's still good.

Paris Martineau [01:18:34]:
A lot of people. Yeah, yeah.

Leo Laporte [01:18:35]:
The producer of wholesome YouTube shows, Dhar Mann. 56 million.

Jeff Jarvis [01:18:40]:
Have you ever heard of Dhar Mann?

Paris Martineau [01:18:43]:
I haven't heard of most of these people.

Leo Laporte [01:18:45]:
Adam W. This is the thing is who's going to ever know about. These guys are like in five years nobody's going to know who these guys are. Remember Amazing Orange? Whatever happened to Amazing Orange? There's Jake Paul, he's got big engagement. 2.33%. Rhett and Link, they deserve some. I think they're pretty creative. I like them.

Leo Laporte [01:19:09]:
Very low engagement though. 1.177%. They have streaming channels on Roku, Amazon prime and Samsung. Plus a best selling cookbook. Alex Cooper, call her daddy.

Jeff Jarvis [01:19:23]:
So you'll see two stories.

Paris Martineau [01:19:25]:
Annoying Orange is still posting videos apparently.

Leo Laporte [01:19:29]:
Oh well, see, I was wrong.

Jeff Jarvis [01:19:31]:
So stay there on Alice Cooper. If you would. Stay there on Alice.

Paris Martineau [01:19:34]:
One got a hundred thousand views. He has 14,000 YouTube subscribers.

Leo Laporte [01:19:39]:
All right, so powerhouse Alexandra Cooper has.

Jeff Jarvis [01:19:43]:
Opened her own ad agency, the Unwell Ad Agency.

Leo Laporte [01:19:46]:
Huh? Yeah.

Jeff Jarvis [01:19:48]:
And she's also done a partnership with Google for Unwell. This is on the rundown line.

Leo Laporte [01:19:57]:
She launched 121 and 125 called unwell hydration. I'm not sure I want unto.

Jeff Jarvis [01:20:03]:
I'm not sure what drip.

Leo Laporte [01:20:04]:
That sound good.

Paris Martineau [01:20:06]:
She also on her show always wears a sweat suit like top and pants that say Unwell on them.

Leo Laporte [01:20:14]:
Yeah.

Jeff Jarvis [01:20:15]:
Which I think is product placement.

Leo Laporte [01:20:17]:
Yeah. That is, by the way, they all do that, right? You see their products. Yeah. And I think that there is some case to be made that that should really be advertisement. That should be.

Benito Gonzalez [01:20:26]:
It's also so weird because there always used to be this rule for bands at least that you never wear your own banshee shirt. You wear other people's bad shirt.

Leo Laporte [01:20:33]:
It's tacky.

Jeff Jarvis [01:20:34]:
Well, the Internet's tacky. Alex and her team will use Google's entire ecosystem to manage the operational and creative needs of her network. They'll use Pixel phones to shoot, edit and publish high quality content as well as effortless, effortlessly collapse from collaborate from anywhere via Google Workspace. They'll also use Pixel Watch and Buds to stay connected and productive. Well, it sounds like my life. Can I get paid for this?

Leo Laporte [01:20:58]:
It's probably good to read this list because you're going to see most of these on the super bowl ads and you, you'll want to know who they are because I know I constantly see these people on TV and it's like, is that person famous? Yes.

Jeff Jarvis [01:21:13]:
Matt Rife is very funny people.

Paris Martineau [01:21:16]:
Yeah, yeah.

Jeff Jarvis [01:21:16]:
Matt Rife is very funny.

Leo Laporte [01:21:18]:
Okay. All right. Haley Bailey. I do like Mark Rober, the former NASA scientist who makes $25 million a year doing very, very clever science content. I think he's good. Drew Ski. He's on a Dunkin Donuts super bowl ad. Or was.

Leo Laporte [01:21:36]:
That explains it. I was wondering who that was. It was Ben and Casey Affleck. Bill Belichick, Jeremy Strong. And Druski. Kb Lame. I love him. He's the guy.

Leo Laporte [01:21:48]:
This is kind of an interesting story. I mean, he was just some guy who just started making TikToks, mocking other people's tiktoks. And. And he would always end it with, like, you know, because he would. They would. They would show some complicated way of doing, you know, sewing on a button, and then he would do it simply and. And go, huh? And he May. He makes $20 million a year now.

Leo Laporte [01:22:17]:
Good for KB. I think that's really great. Or Kabi. I'm not sure how you say it, but I've been a fan of him forever. His story is fascinating. He's from Senegal, and he was born in Dakar. They moved to Italy when he was 1 year old. Lived in a public housing complex near Turin.

Leo Laporte [01:22:43]:
He worked as a CNC machine operator at a factory in Turin. Got laid off in March 2020, and that's when he started posting on TikTok. He was out of work, and during the pandemic, he posted on TikTok, rose to popularity doing those stitch videos and the duet videos, and has become a huge star. What a. I mean, I think that that's an example of how amazing some of this is. I don't know if the new TikTok will be worth anything, but. But we'll see.

Jeff Jarvis [01:23:16]:
Did you get number 17 yet?

Leo Laporte [01:23:18]:
Number 17. Let me. Steve Bartlett, Dixie D', Amelio, Brent Rivera, Danny Austin. It's like, I feel there's another Logan, Paul, Jack, Shane, Nick, Thea, Giovanni. I feel like. Oh, he. He does food.

Jeff Jarvis [01:23:32]:
He does food.

Leo Laporte [01:23:33]:
Yeah, he does food. Okay. All right.

Jeff Jarvis [01:23:36]:
World's spiciest food. Yeah.

Leo Laporte [01:23:39]:
People like painful food. I don't.

Benito Gonzalez [01:23:40]:
Yeah, they do.

Leo Laporte [01:23:41]:
Oddly Marquez Brownlee's way down there in 17. I thought he would be higher. $10 million a year, though. Ain't bad. It ain't bad.

Paris Martineau [01:23:49]:
Pretty good.

Leo Laporte [01:23:50]:
It ain't bad. Emma Chamberlain. She was. She was, like, number one for a while with the coffee thing, right? Or whatever it was.

Jeff Jarvis [01:23:58]:
Oh, was that it?

Leo Laporte [01:23:59]:
Yeah, yeah. Also a podcaster, Ms. Rachel. She's the apparent podcaster. A parent YouTuber. Potty training with Ms. Rachel, by the way, if you're laughing at that. She makes $23 million a year, so.

Jeff Jarvis [01:24:17]:
Who says there's not money in poop?

Leo Laporte [01:24:19]:
So don't laugh at that. Hannah Stocking.

Jeff Jarvis [01:24:22]:
Ijustine is number 44.

Leo Laporte [01:24:25]:
Good for her.

Jeff Jarvis [01:24:26]:
$3.5 million. Nothing to sneeze at.

Leo Laporte [01:24:30]:
Good for her. Gene Jack Skepticeye McLaughlin. I'm sorry. Sean. Sean. Adam. Debbie. Well, anyway, this is a good.

Leo Laporte [01:24:41]:
This is from Forbes. Influencers are now part of the game.

Jeff Jarvis [01:24:46]:
There is the influencer cat, Gizmo, say hi to Gizmo.

Leo Laporte [01:24:50]:
Should be worth a million a year at least.

Paris Martineau [01:24:52]:
Gizmo, you should be worth a million a year so that you could pay rent. Wouldn't that be great?

Leo Laporte [01:24:57]:
What would Gizmo's rent be though?

Paris Martineau [01:24:59]:
$1 million a year, ideally. We wanted to be specific about it.

Leo Laporte [01:25:04]:
Well, I'm sure Grumpy Cat made that much for his owners. Or her owners.

Jeff Jarvis [01:25:08]:
Oh, yeah, yeah.

Leo Laporte [01:25:10]:
Hey, can I take a break? Because I have. I am running low on energy. Maybe Gizmo would be interested in this. It is time supplements for a field of greens break. You know, I used to what I used to do, I'm embarrassed to say, several times a day I would chow down on a handful of supplements. This episode brought to you by Field of Greens. Not anymore. You could do a hard reset on your metabolism.

Leo Laporte [01:25:37]:
Let me just, just because I like to do this in real time so people can see first of all what it's like. I think today they have all sorts of flavors. I'm going to try. I did the lemon, lime. Let me do. Oh, this sounds good. Strawberry lemonade. Strawberry lemonade.

Leo Laporte [01:25:54]:
So I've got my field of greens here. This is a superfood with all the vitamins, but more than just vitamins. All of the things that are good for your body, always organic, made from real food. This is not some chemical stew. I'm just going to take the scoop here, put it in my shaker, shake it up and I'm going to drink it just to show you how good it is. It's a smart reset of your health. Kind of like a smart, smart reset of your computer. Get you as close to optimal performance to the factory settings as possible.

Leo Laporte [01:26:26]:
And this is, this is backed by science. Field of Greens did a very impressive large study a a double blind study with Auburn University. What they did, they wanted to see if taking field of greens daily could slow down your aging. You know, it's nice, it would be nice if your body didn't match your chronological aging. If you could actually age slower then the years go by slowing the rate at which your body ages generally means you're going to live longer and more importantly, at least to me, you're going to feel better while you're alive. You're going to be healthier. Each fruit and vegetable and field of greens medically selected. It's a doctor founded this company for specific health benefits 100% organic.

Leo Laporte [01:27:11]:
And they have different groups. You can actually read it on the label or go to their website and read it. There's a heart health group. There's a cells group, a kidney, lungs and liver group. There's even a metabolism group for help you get a healthy weight. The biological study participants did all the pre testing. Some learned from the blood work that they were aging faster than their chronological age. They were aging too quickly.

Leo Laporte [01:27:39]:
Now all the participants were told, don't change anything. We want to have just one variable in here. You eat normally. If you drink, keep drinking. If you don't exercise, keep it up. Keep up the good work. But the one change, because this is the whole idea, and they had a placebo group and a field of greens group. The one change we're going to have you do is drink this every day.

Leo Laporte [01:28:00]:
And the results, and you can read it on the website. Fieldofgreens.com were remarkable. The group that added field of greens literally slowed down noticeably, measurably, how fast their bodies were aging. Think about how much better you feel with field of greens from Brick House Nutrition. I'm gonna just. I've mixed this up. This is. What did I say? Strawberry lemonade.

Leo Laporte [01:28:22]:
I have wild berry. I have a few different lemon, lime. This is the strawberry lemonade. Ah, that is fantastic. You know, it's the kind of thing where you go, I'd like to drink this every day just because it tastes good. Good, but your body's gonna like it, too. Check out the university study. Get 20% off when you use the promo code.

Leo Laporte [01:28:42]:
Imieldofgreens.com that's fieldofgreens.com the promo code. I am. And look at the ingredients when you're there because I think you'll be really impressed. 30 calories. So it's a low calorie supplement. You could do it once a day. I do it twice a day, morning and night because it tastes good and I just love it. Fieldofgreens.com use the promo code.

Leo Laporte [01:29:05]:
Im for 20% off your first order. You know, it's a little pricey, but I have to say it's less than I was paying for those supplements when I added up the cost of all the supplements. See, the supplements, they fool you because it's a little bit for this, a little bit for that. By the time you add it all up, it's like, wow, this is less than that. This is affordable, and it really makes a difference. Fieldofgreens.com thank you, Field of greens, for supporting us. Let's do our weekly rundown through the pre print scientific studies that Jeff has found like this one. Can large language models develop a Gallic gambling addiction?

Paris Martineau [01:29:46]:
I like that our eyes were. Both of our eyes were drawn to that. Because it's in all caps.

Leo Laporte [01:29:51]:
Yes. Well, did they? Can they? These findings suggest LLMs can internalize human like cognitive biases and decision making mechanisms beyond simply mimicking training data patterns. Emphasizing the importance of AI safety design and financial.

Jeff Jarvis [01:30:09]:
They have bad judgments. AIs.

Leo Laporte [01:30:10]:
In slot machine experiments we identified cognitive features of human gambling addiction, illusion of control, gambler's fallacy and loss chasing. So if you use an AI, for instance to say, pick stocks, you might want to, you know, it's not going to be more objective than you are. It might do some things that are pretty dumb. When given the freedom to determine their own target amounts and betting sizes, bankruptcy rates rose substantially alongside increased irrational behavior behavior.

Benito Gonzalez [01:30:45]:
That's because the house always wins, right? That's because the house always.

Leo Laporte [01:30:48]:
Well, that's part of the problem. Yeah. And like all of these archive dot orgs, we've got a nice little graph here.

Jeff Jarvis [01:30:54]:
A little. There is, There is a, an aesthetic to these things. I don't know.

Paris Martineau [01:31:01]:
I just asked Claude if he could get addicted to gambling and Claude said I can't get addicting to gambling or anything else. I don't have experiences between conversations. Each time we talk I start fresh with no memory of previous interactions.

Leo Laporte [01:31:16]:
That's true.

Paris Martineau [01:31:17]:
I have no persistent desires, cravings or the neurological reward systems that drive addictions in humans.

Leo Laporte [01:31:25]:
Yeah, so says he. I think that probably in this case they would have to have a single session rather than starting fresh each time.

Benito Gonzalez [01:31:34]:
I mean, isn't this just a fancy way of asking if AI can play games? Because the.

Leo Laporte [01:31:38]:
That's.

Benito Gonzalez [01:31:38]:
We've had AI in games forever.

Jeff Jarvis [01:31:42]:
Yeah. Just before Dungeons and Dragons. And.

Leo Laporte [01:31:47]:
Do you fall prey to the gumblers, the grumblers?

Paris Martineau [01:31:52]:
Do you fall prey to the grubblers? Fantasy.

Leo Laporte [01:31:55]:
It's the Monte Carlo fantasy or the fallacy of maturity of chances. The belief that, you know, the roll of the dice, which is independent each time you're more likely, you know, if you've thrown snake eyes a bunch of times to come up with the opposite, whatever that is, sevens. The next dice roll is more likely to be six because there have been fewer sixes in the past. That's the gambler's fallacy. And a lot of people fall prey to that. You know, like if you flip a coin a thousand times, if you got a lot of Heads. You're probably going to get some tails now.

Paris Martineau [01:32:30]:
No, I told Claude to read the study. Drop the URL in there. Says what this means for me. While I can't develop addiction in the human sense, the research shows that within decision making context, I and the other LLMs can exhibit the cognitive distortions that characterize addictive behavior, especially when given goal oriented instructions in risky situations. This has serious implications for using LLMs in financial decisions decision making with the author which the authors rightly emphasize as a safety concern.

Jeff Jarvis [01:33:00]:
Now ask it if it'll give you stock recommendations and should you trust it.

Leo Laporte [01:33:05]:
I have been tempted to ask it like are we in a, are we in a, you know, bubble right now? Should I, should I buy gold and that kind of thing. But I don't think an AI would know. No, no.

Paris Martineau [01:33:17]:
The answer would be the prevailing stock, Rex.

Leo Laporte [01:33:20]:
Would it though?

Paris Martineau [01:33:21]:
I think I said got any stock, Rex? And it said ha. Given what we just read, I'm probably the last entity you can ask for that kind of advice. I'm not qualified to give stock recommendations for several important reasons.

Leo Laporte [01:33:36]:
But this is why it's addictive because that was kind of an engaging response, right? Based on what we just. That's like a human hello tail. That's like a human would say that these guys are good at putting. I don't, you know, I have to think this is tuning and post training, not the model itself, but this is something that the companies actually added on. But it's good, it really gets you going.

Paris Martineau [01:34:01]:
It did say if you're interested in investing, I'd suggest speaking with our licensed financial advisor who knows your full picture, starting with broad diversified index funds. If you're new to investing, doing your own research on companies or sectors you're interested in and never investing money you can't live lose.

Leo Laporte [01:34:19]:
Good advice. Good advice. You've read all of these, Jeff?

Jeff Jarvis [01:34:27]:
No. Well, I've looked at them. There were this week. There were 700 papers mentioned.

Paris Martineau [01:34:34]:
Did you read them all?

Jeff Jarvis [01:34:35]:
I don't read them all. I look at the headline, I look at the titles of them all.

Leo Laporte [01:34:37]:
Usually the abstract, like I just read from the abstract. That was sufficient.

Jeff Jarvis [01:34:43]:
As soon as they hit the foot, the first formula, I'm out.

Leo Laporte [01:34:45]:
Yeah, me too.

Paris Martineau [01:34:46]:
I wish there was a way to track how many words you read, you read in the average week, Jeff. Just because I'd be curious.

Jeff Jarvis [01:34:55]:
Define read.

Paris Martineau [01:34:59]:
Tell me more what you mean by that. Please explain.

Jeff Jarvis [01:35:03]:
So when I, when I go through. So if you go to. Let's just do an experiment here. Let's go to archive.org A R, X I V and just type in the search box AI.

Paris Martineau [01:35:15]:
Is this what you do? You're really intellectual.

Jeff Jarvis [01:35:20]:
This is what I do. So molex bargain. Emergent misalignment. When LLMs compete for audiences might be interesting. I don't know Moloch. So. But some of them are really beyond. These are kind of all.

Jeff Jarvis [01:35:36]:
Okay, there's some that get really ridiculous trying to find an example here.

Paris Martineau [01:35:43]:
You mean the Canyite deity associated with biblical sources and the practice of child sacrifice? That Moloch, the ancient demon God?

Leo Laporte [01:35:53]:
Which Moloch are we talking about here?

Paris Martineau [01:35:55]:
I don't know. That's why I was surprised to hear a reference. I had to.

Leo Laporte [01:35:58]:
I just want to similar. Instead of reading that crap, I read things like this. Professional automated training in theory and practice. This will tell you how to write your very own Lisp program to trade stocks and you will suddenly be wealthy. How about that? How about it? I bought this book thinking it'd probably be a pretty good book for learning Lisp.

Jeff Jarvis [01:36:24]:
So I just put a paper in the chat that gives you an example of a paper that I can't even begin to read.

Leo Laporte [01:36:32]:
Okay. A paper you can't begin to read. Shall we all read it together?

Jeff Jarvis [01:36:35]:
Let's try to read it together.

Leo Laporte [01:36:37]:
Let's try to read.

Paris Martineau [01:36:40]:
Oh, no.

Leo Laporte [01:36:42]:
First of all. Oh, that's just the people who wrote it.

Jeff Jarvis [01:36:44]:
Yeah, but the title alone.

Leo Laporte [01:36:46]:
First measurement of the D plus 8.

Jeff Jarvis [01:36:51]:
Okay. To the 0.

Leo Laporte [01:36:53]:
Holy cow. This is some sort of. It's a physics paper. I would.

Jeff Jarvis [01:36:57]:
I guess it. I have no idea.

Leo Laporte [01:36:58]:
Well, let's read the abstract. We report the first measurement of the semi leptonic decay using a sample of E plus E minus annihilation data corresponding to an integrated luminosity of 7.33 foot bars to the minus 1 collected as center of mass energies between 4.128 to 4.226 gigavolts with the.

Jeff Jarvis [01:37:22]:
So you see Paris, when my eye passes over that title, I don't read it.

Leo Laporte [01:37:27]:
You don't need to read this in a glance.

Jeff Jarvis [01:37:31]:
It ain't for me.

Leo Laporte [01:37:32]:
Wow, this is fascinating.

Paris Martineau [01:37:34]:
Digging. That paper was referring to the Moloch God of child sacrifice that I'm discussing.

Jeff Jarvis [01:37:40]:
So now you're into it. You're going to start reading Archive 2.

Paris Martineau [01:37:43]:
Moloch. Moloch.

Leo Laporte [01:37:44]:
Nightmare.

Paris Martineau [01:37:45]:
Moloch. Moloch. Lovelace.

Leo Laporte [01:37:46]:
I don't know why you didn't read it. There's two pages of authors, but there's only nine pages of paper. It's not. You could have just skipped ahead.

Jeff Jarvis [01:37:55]:
Is there a conclusion?

Leo Laporte [01:37:59]:
Yeah, yeah, there is. Yeah. This has to do with subatomic particles and I think you need a super collider. In my opinion it should be bigger with it. Yes.

Jeff Jarvis [01:38:12]:
Table 2 Hadronic form factors where the first uncertainties are statistical and the second systematic. Well, isn't that life?

Leo Laporte [01:38:21]:
I think one of the things that it looks interesting to me here, sycophantic AI decreases pro social intentions and promotes dependence. We're seeing this where people are saying using a lot of AI makes you kind of, you know, less independent, makes it more difficult to think the abstract says. Both the general public and academic communities have raised raised concerns about sycophancy, the phenomenon of artificial intelligence excessively agreeing with or flattering users. Yet despite media reports consequences. Yeah, moral panic. Little is known about the extent of sick offense. Oh, he did it.

Paris Martineau [01:39:08]:
He got back after losing the first one.

Leo Laporte [01:39:13]:
Here we, here we show the. If you're listening me, every time it's some weird AI video of me turning into a robot, that's what we just saw. And then saying more panic. Here we show the pervasiveness and harmful impacts of sycophancy when people seek advice from AI. Yeah, well that kind of makes sense, I guess, you know.

Jeff Jarvis [01:39:34]:
Right, and it's a question then of how the AI are designed.

Leo Laporte [01:39:37]:
Right, And I get back to that, which is we forget that the companies that make these AIs do a lot of post training, tuning, reinforcement, learning, all sorts of stuff that very much gives it a flavor and a style. And sycophancy is one of the knobs they can turn before it happened.

Jeff Jarvis [01:39:54]:
People are drawn to AI, get really.

Paris Martineau [01:39:55]:
Mad when the sycophancy turned back down.

Leo Laporte [01:39:57]:
That's what OpenAI did. Yeah.

Jeff Jarvis [01:40:00]:
People are drawn to the AI that unquestionably validate even as that validation risks eroding their judgment and reducing their inclination toward pro social behavior. These preferences create perverse incentives both for people to increasingly rely on sycophantic AI models and for AI model training to favor sycophancy. This is a moral hazard.

Leo Laporte [01:40:22]:
I like this piece about a software called StorySage which has been designed.

Jeff Jarvis [01:40:30]:
I fought for this one.

Leo Laporte [01:40:32]:
Yeah, I like this. It's a multi agent AI framework designed to help you write your autobiography.

Jeff Jarvis [01:40:40]:
Especially if you wore things over your neck that could record.

Leo Laporte [01:40:45]:
So the first interview session, the robot says, what were your favorite activities during childhood? And you say, well, music, bicycle riding, playing basketball. And then it says, well, which of those topics would you like to explore in the next session. And then you work together with the. But the AI More acts more as a prompter. Right? As a. It's kind of turning it on its head. It prompts you to.

Jeff Jarvis [01:41:10]:
But is everybody's life worth a biography?

Leo Laporte [01:41:13]:
Not. Not to the world in general, but your kids might want to read that. My dad wrote a Imagine biography.

Paris Martineau [01:41:21]:
Leaving you an AI Slop biography though, like that. You don't know whether.

Leo Laporte [01:41:27]:
Put some more effort into it.

Jeff Jarvis [01:41:28]:
Put.

Leo Laporte [01:41:29]:
Put more effort into it.

Paris Martineau [01:41:30]:
Well, if they were putting more effort into it, they would write some of it themselves instead of.

Leo Laporte [01:41:34]:
No, no, I think that's designating it to a. They call it co creation.

Paris Martineau [01:41:39]:
Okay.

Leo Laporte [01:41:40]:
I think that's the intent.

Paris Martineau [01:41:41]:
Yeah. Just like Snooki co created her biography, you know?

Leo Laporte [01:41:47]:
Well, it. It's no different than a ghostwriter, I guess.

Jeff Jarvis [01:41:51]:
So I'm going to speak at a ghost writers conference.

Leo Laporte [01:41:54]:
Really?

Paris Martineau [01:41:54]:
Are you going to wear a sheet?

Jeff Jarvis [01:41:57]:
No.

Paris Martineau [01:41:58]:
Or do you think it's going to be weird for you to go as a writer who writes into your own name?

Jeff Jarvis [01:42:04]:
Yeah, I'm not sure why they want me. They asked me to be on a panel and I'll be on a panel. But I'm kind of like, yeah, maybe AI should replace all of you.

Paris Martineau [01:42:14]:
Huh? That's odd. And yet you accept it.

Jeff Jarvis [01:42:18]:
Yeah, because they don't pay me or anything. But what am I doing on that day? Would you ever ghost write somebody's book? Paris?

Paris Martineau [01:42:28]:
One of my first like internships or like writing opportunities in college. I was ghostwriting blog posts for this like, non profit and I remember being really flat. I mean, I was like a college student. I had written anything for anyone. I remember being really flattered. The guy wanted me to ghost write Atlantic articles for him, but I never did it because I got a real writing job.

Jeff Jarvis [01:42:47]:
Heck with you, man. That's my byline. You'd say.

Paris Martineau [01:42:50]:
I know. I was like. I remember thinking like, oh, it might be kind of weird that I then I guess couldn't cite that I was published in the Atlantic, but wouldn't have been cool to get published anyway. I was so naive.

Jeff Jarvis [01:42:59]:
So, Leo, would you do this?

Leo Laporte [01:43:01]:
No, but I have a funny story about that. Belva Davis just passed away. I know. You know who Belva Davis was? Jeff Jarvis. She was in her 90s, but she was the first black female television journalist. Yeah, she was here in San Francisco and really was a legend in San Francisco journalism and a really wonderful person. Good friend of mine was the ghostwriter for her autobiography. Never in my wildest Dreams.

Leo Laporte [01:43:30]:
And I didn't even know this, but I was reading Belvedavis's obituary, and I saw. Oh, Vicki co. Wrote that with her. Yeah, She's a good friend of mine. She was a longtime newspaper reporter and made a career out of writing.

Jeff Jarvis [01:43:43]:
That's not quite ghostwriting, because not if they give you. My line is big.

Leo Laporte [01:43:46]:
It's clear. Yeah, yeah, that's right. As. As told to. It's more like as told to. Right.

Jeff Jarvis [01:43:53]:
That's. That I have more respect for.

Leo Laporte [01:43:55]:
Than ghostwriters traditionally. What? You don't even know they were there. They were a ghost. They might.

Jeff Jarvis [01:44:01]:
They might get thanked for research in the acknowledgments.

Leo Laporte [01:44:04]:
When I wrote. Well, I had my name on more than a dozen books. Some of them, I wrote most of them.

Jeff Jarvis [01:44:12]:
That's right. You had ghostwriters.

Leo Laporte [01:44:13]:
Yeah, yeah. Most of them I wrote with help.

Paris Martineau [01:44:17]:
And I would always credit Patterson in this situation.

Leo Laporte [01:44:20]:
Yeah. It would be Leo laporte with, you know, somebody else. Megan Maroney wrote one of my.

Jeff Jarvis [01:44:26]:
I bought many of those books for my son. Yes.

Leo Laporte [01:44:28]:
And then. And then at one point, I just kind of said, there's no money in this. It's so much work. It literally killed me every time I said. Said, look, I'll have an imprint. We had a Leoville Press. I will edit these. I'll technical edit them.

Leo Laporte [01:44:41]:
I'll support the writer, but let the writer write. And at that point, I took my name off of it. It was. That person wrote it. It would be Leo Laporte's guide to TiVo, but it would be written by Gareth Branwyn.

Jeff Jarvis [01:44:52]:
And how many books?

Leo Laporte [01:44:53]:
13.

Paris Martineau [01:44:54]:
Oh, did sales drop when you took your name off them?

Leo Laporte [01:44:58]:
They dropped pretty much from the beginning. The first book was. Was bestseller. It was Pearson's book of the year. Sold 55,000 copies, which was, you know, it's nothing. And, I mean, we have more people listening to this show every week than that, so it wasn't. You know, books don't make you any money. As Jeff I'm sure will tell you, my book will.

Leo Laporte [01:45:21]:
What's your book gonna be?

Paris Martineau [01:45:23]:
No idea.

Jeff Jarvis [01:45:24]:
I think her book should be the Story of Crime.

Paris Martineau [01:45:28]:
Oh, one of those.

Leo Laporte [01:45:30]:
Probably. It's got to be updated, right?

Jeff Jarvis [01:45:34]:
Not that. Not that Paris ever speaks for cr.

Paris Martineau [01:45:37]:
I don't speak for cr, but is this buyer aware?

Leo Laporte [01:45:41]:
Oh, I like it. The Story of cr.

Paris Martineau [01:45:44]:
Yeah. Harness our consumer power for a safe, fair, and transparent marketplace.

Leo Laporte [01:45:49]:
Nice. Sounds a little preachy to me.

Jeff Jarvis [01:45:53]:
Yeah, I think there's. I think there's the story behind I Think how the lab works. I think there's nerdy details to get.

Leo Laporte [01:45:58]:
Into all the sex, the violence, all the cars that were crashed in the CR test track after hours, that kind of thing.

Jeff Jarvis [01:46:06]:
Yeah.

Leo Laporte [01:46:07]:
The seasy sleazy side of cr.

Paris Martineau [01:46:10]:
No, yeah, that definitely doesn't exist.

Leo Laporte [01:46:13]:
My ex wife said, where'd you get that shampoo? She I said, well it was recommended by Consumer Reports. She said, you don't buy shampoo based on reviews for Consumer Reports. That's a bunch of guys who wash their hair with soap.

Paris Martineau [01:46:30]:
Well, for products like that that are more personal decisions, they end up. They get like survey panels. They, they do panels of either employees or they ask the employees to have people come in and they do like really large scale preference based studies.

Leo Laporte [01:46:48]:
That's much better. Yeah, you can't review shampoo. That's a.

Paris Martineau [01:46:52]:
People are always telling me I've got to get in on whenever they do the food based review things because then you get to try like radioactive 200 cookies. I did learn, I did learn when I was visiting the labs that one of the ways we. It was annoying, I guess that when I did the lab visit they weren't actively testing any ovens because part of the way they test how like the oven, the hot spots in the oven is they bake giant plates of sugar cookies. And so normally there are just endless amounts of sugar cookies.

Leo Laporte [01:47:23]:
Yeah.

Paris Martineau [01:47:23]:
Where they get burnt.

Leo Laporte [01:47:25]:
I'll be honest with you. Another thing my ex wife once said. One piece of cake, delicious. Two piece of cake. Three piece of cake. You wish you never saw cake.

Paris Martineau [01:47:35]:
She's an ex wife. Yeah. Three pieces of cake. I'm having a great time.

Leo Laporte [01:47:43]:
She's very slender. Have we passed peak social media? Yet another thrilling, gripping opinion piece from Financial Times. I swear to God, I must have been on a Financial Times kick this week. As platforms degrade into outrage and slop, users are turning away, says John Byrne Murdoch.

Paris Martineau [01:48:06]:
Oh, a Murdoch.

Leo Laporte [01:48:09]:
I was wondering. Okay. In years. He's a very relation. We may. We'll look back on September 2025 is the point at social media jumped the shark and began rapidly accelerating its transition from the place to be seen through a flattering Instagram filter to a gaudy backwater of the Internet inhabited by those with nothing better to do. No. I don't know.

Leo Laporte [01:48:36]:
What do we think is it. Is social over? No, it is peaking. This is.

Paris Martineau [01:48:41]:
People love to post.

Leo Laporte [01:48:42]:
This is kind of inspired by a recent study that says social media time peaked in 2022 and the youngs are cutting back faster than people by the way.

Paris Martineau [01:48:55]:
Okay, this is.

Leo Laporte [01:48:56]:
It ends at 64, Jeff. We're both too old to be in this study, apparently.

Paris Martineau [01:49:01]:
Okay, this is skewed by the fact that from 2020 to sometime in 2022, all you could do is be online.

Leo Laporte [01:49:10]:
That's a good point. It's gonna go down since 2020. You're. Exactly.

Paris Martineau [01:49:15]:
Because of course now I can go outside and be within six feet of people.

Leo Laporte [01:49:20]:
A lot of things changed, although I just read a study that said gaming did not escalate during the COVID pandemic as the way we thought it might. The people would be stuck at home playing more games.

Paris Martineau [01:49:32]:
I think the sales of Animal Crossing.

Leo Laporte [01:49:35]:
I played a lot of Animal Crossing. Holy cow. By the way, I mentioned TiVo and the Discord is reminding me that TiVo announced it's no longer going to make its hardware as of.

Paris Martineau [01:49:49]:
Do we want to do a brief obituary?

Leo Laporte [01:49:51]:
It's the end of the line for TiVo. They exited the DVR market as of October 1st. I had so many TiVos. As I said, I wrote the book on hacking the TiVo. Well, Gareth Brannan wrote it, but I was there. Gareth Branwin. I had many tivos with lifetime subscriptions. I don't know what happens now.

Leo Laporte [01:50:17]:
No, I've given them away a long time ago. But TiVo was the first. Really. I mean, not the first DVR. There was replay, there were others. But TiVo really was the king of the DVRs, the digital video recorders nowadays.

Paris Martineau [01:50:29]:
Patrick is still using his second TiVo.

Leo Laporte [01:50:34]:
Yeah, lifetime subscription. Well, keep using it. Yeah.

Jeff Jarvis [01:50:40]:
He has to explain the first wasn't hd so he upgraded otherwise.

Leo Laporte [01:50:43]:
Yeah, he has a cable card in it which, you know, the cable company is no longer required to provide.

Paris Martineau [01:50:49]:
What is a cable card?

Leo Laporte [01:50:51]:
For a while the FCC said, told the cable companies, yeah, you can have your own cable set top boxes, but you have to allow. Some have some way to allow consumers to have their own Ativo. For instance. Probably it was at the lobbying, the behavior.

Jeff Jarvis [01:51:04]:
Yeah, I think so.

Leo Laporte [01:51:06]:
So they made a. It was a PCMCIA card, which is a little credit card sized thing that actually was, I imagine just a certificate that said, yes, this person has a TiVo account. Or this is. Or I'm sorry, cable account. This is the. This is the cable account associated with this person. You put it in and then it probably goes online and says, is the account still good? And it says, yes. Okay, now you can DVR the shows, but it does allow you to use a TiVo or some other set top box.

Leo Laporte [01:51:32]:
With a cable company feeds and watch the paid channels like HBO and so forth. Now, though, with, with things like.

Paris Martineau [01:51:40]:
With streaming.

Leo Laporte [01:51:41]:
Yeah, yeah. Well, on YouTube, TV or Hulu or what's, what's the, what's the big one? I can't remember anymore. Dish. Right. With. With all of these, you can, you can stream all that dvr, record it and everything without any hardware anyway.

Jeff Jarvis [01:51:59]:
Right. Well, but in the earliest days, copyright was interpreted as saying that if you recorded something on your machine that that was copying it and was illegal. Similarly, it was said that the cable company. In the very earliest days, from the. The Time Inc. What was it called? Total Network. Something else. I can't remember what it was.

Jeff Jarvis [01:52:21]:
They had to own the copies. It's not unlike the Anthropic case. They had to own copies of the, of the, of the show if you wanted to. To time shift it. And at one point there was like somebody in roller skates going down a whole wall of VCRs, putting it in there so I could play this for Leo. And, and so the, the copyright interpretation in the early days was that to have it digitally reside anywhere else was a copy, ergo, was a violation of copyright.

Leo Laporte [01:52:52]:
Wow.

Jeff Jarvis [01:52:53]:
And so, you know, you remember the fights about, about Betamax and all that?

Leo Laporte [01:52:57]:
Oh, yeah.

Jeff Jarvis [01:52:58]:
All right. It was the same thing. And. And then at some point it just becomes, oh, that's ludicrous, and we move on and I wonder if we're going to get the same place with AI.

Leo Laporte [01:53:10]:
Despite all the legal decisions, it didn't stop Marion Stokes, and thank goodness it didn't. Do you know her story?

Jeff Jarvis [01:53:19]:
No.

Leo Laporte [01:53:25]:
She lived in Philadelphia. An older woman who decided that she wanted to record every TV show on her Betamax. In 1975, she got a Betamax and became. She started with sitcoms, science documentaries, starting with the Iran hostage crisis in November 1979, her son said, and her son wrote a book called Recorder, the Marion Stokes Project. She hit record, and she never stopped. She recorded everything. In 1980, when CNN launched, she recorded that. Recorded all of them.

Paris Martineau [01:54:00]:
Wow.

Leo Laporte [01:54:01]:
She soon had three, four, five, and sometimes as many as. As eight machines recording in her Philadelphia apartment. Recording everything on multiple networks.

Jeff Jarvis [01:54:11]:
Did she save all the tapes or.

Leo Laporte [01:54:13]:
What did she do? Her son said, we'd be out to dinner, we'd have to rush home to swap tapes. She was a librarian. She was, yep. The Free Library of Philadelphia for 20 years. Fired in the early 60s for her work as a Communist party organizer.

Jeff Jarvis [01:54:30]:
Wow.

Leo Laporte [01:54:31]:
But she believed that it was important to record everything. And as a result, those videotapes survive at the Internet archive.

Jeff Jarvis [01:54:39]:
Oh, that's so 30 years of TV.

Leo Laporte [01:54:42]:
History recorded by one woman in Philadelphia now on the Internet Archive.

Jeff Jarvis [01:54:48]:
So my dear friend Peter Travers, who was the movie critic at Rolling Stone, he wanted to own a copy of every movie ever made. Wow. For I would go with him to a store in New York on 38th street and he would buy, he had his place and he would buy boxes of VHS tapes and he would find ways he would get review copies and do things because he was a critic and so on. And I asked him, not long. I don't talk to Peter nearly enough. And I asked not long ago, whatever happened to all those? Oh, his wife finally told him to throw them out.

Paris Martineau [01:55:21]:
Wow.

Jeff Jarvis [01:55:22]:
He had thousands of movies.

Leo Laporte [01:55:24]:
By the way, she was a smart woman. It's kind of interesting story. She and her ex husband attempted to defect to Cuba because they had been surveilled by the government because of her communist affiliations. She didn't want a TiVo because she felt like there was a corporate tie up. But she was an early and evangelical investor in Apple Computer. She convinced the rest of her already well wealthy family to buy Apple stock. This is in the 70s. Wow.

Leo Laporte [01:55:56]:
And took the money they made to fund her recording project in the massive storage space she required to store all those tapes.

Jeff Jarvis [01:56:07]:
Since we're doing nostalgia time, did you ever hear Paris of Gemstar?

Paris Martineau [01:56:13]:
No.

Jeff Jarvis [01:56:14]:
You remember Jim Starley? Oh, Russell.

Paris Martineau [01:56:17]:
Not.

Leo Laporte [01:56:17]:
I mean, I know the name. I don't remember what it was.

Jeff Jarvis [01:56:19]:
So we would have. Back in the day, my child, we had VCRs with flashing 12s.

Leo Laporte [01:56:25]:
Oh, yes.

Jeff Jarvis [01:56:26]:
And recording, recording on a VCR was what required a technical degree. You had to hit this button and that button, then hold that button and then do this and do that for the beginning and the end and the channel. And it was beyond any hope. So a company came along called Gemstar where they created a very proprietary formula.

Leo Laporte [01:56:46]:
It was a program guide.

Jeff Jarvis [01:56:48]:
Yep. Well, no, no. TV Guide ended up merging with them. The TV Guide licensed the numbers. So there was a special number and you get a Gemstar enabled VCR on.

Leo Laporte [01:56:58]:
Your TV Guide, there would be an eight digit number, a six digit number.

Jeff Jarvis [01:57:02]:
Every show had a number. So you would just put in that number and that would then record the VCR based on the algorithm, based on the formula.

Paris Martineau [01:57:13]:
So it would be a number that if you had a Gem Star machine you could plug in and it would record the thing that corresponds to that.

Jeff Jarvis [01:57:19]:
Yes, it would. It would translate to 8:00pm to 9:00pm Channel 2.

Leo Laporte [01:57:26]:
And so is that the same as VCR plus or is that something.

Jeff Jarvis [01:57:30]:
VCR plus.

Leo Laporte [01:57:31]:
That was VCR plus.

Jeff Jarvis [01:57:32]:
Yeah.

Leo Laporte [01:57:32]:
Okay.

Jeff Jarvis [01:57:32]:
Yeah. So you then had to replace your VCR to get a VCR plus enabled machine. TV Guide had to use a lot of ink and paper to create the extra space to put those numbers in. TV Guide said, this is going to kill us. So Murdoch merged it with Gem Star, and then he ended up selling Gemstar for not bad money so he could buy Dow Jones. And TV Guide eventually was sold for $1.

Leo Laporte [01:58:02]:
Wow. Wow. And by the way, he bought it from Liberty Media, run by John Malone, which is now a massive conglomerate that owns all sorts of stuff and influences more. Yeah, Formula One racing, they own Sirius xm, they own Live Nation. You know, these cable companies ended up kind of.

Jeff Jarvis [01:58:31]:
Well, Malone is the deal maker of dealmakers. I was always told when I was in the cable business through advance, it was just sport. They. They wanted to leave no meat on the table. Every scrap was fought over.

Leo Laporte [01:58:44]:
That's. That was the secret of Larry Ellison's success, too. He didn't just want to be the most successful, he wanted to kill everybody else.

Jeff Jarvis [01:58:52]:
Yes, like just dominate.

Leo Laporte [01:58:54]:
Beat them all. By the way, F1 is apparently, according to Puck, about to sell rights to Liberty, is about to sell rights to Formula One to Apple. ESPN's been running those Formula One races for the last few years.

Jeff Jarvis [01:59:09]:
Is that the right demographic? Is Apple? Are Apple users?

Leo Laporte [01:59:12]:
Apple's got a lot of money. The Puck says they're going to pay $114 million, which is almost twice what ESPN is paying for the rights. But this has been going on for a while. They expect to announce it October 19th at the Formula One race in Austin. And the sticking point is that Apple wanted Formula One to stop its streaming service, F1TV, which I subscribe to and almost all serious fans subscribe to, because it lets you watch Formula one on your. They have an app for the Apple TV and the Roku, but you can also watch it on your devices, on your phones and stuff. And it's amazing because you not only see the race and you can listen to the Sky TV broadcasters or Formula One's TV broadcasters, but you also have every driver's camera and you have every driver's radio, and you can even watch it all at once if you want on your computer, which is mind blowing. It is an information junkie's dream.

Leo Laporte [02:00:11]:
So I hope Apple, you know, they apparently wanted Formula One to kill that. Formula One says, we make a lot of money on that, and I don't know what the result is. We'll see. We were talking about it today or yesterday on MacBreak Weekly and the Mac Break Weekly, folks. Jason Snell especially thought, you know, I bet Apple wins on this. I bet they do have to shut down the streaming thing. And let's hope Apple does the same quality of streaming. This kind of Reek's doing all right.

Benito Gonzalez [02:00:36]:
This kind of reeks of some. An executive in Apple just being a huge F1 fan and just wanting it.

Leo Laporte [02:00:41]:
Oh, yeah. In fact, if you saw the last Apple event, because they were promoting the F1 movie, remember, with Brad Pitt. Was it. I think it was Eddie Q. No, it wasn't Eddie Q. It was the hare. Who's the. Who's the guy with the hair? Is that John Federighi? Drove up in a Formula One car.

Paris Martineau [02:00:59]:
That's pretty nice.

Leo Laporte [02:01:01]:
Well, it was. I think it was AI. See, you can't.

Jeff Jarvis [02:01:06]:
You don't know anything anymore.

Paris Martineau [02:01:07]:
You can't.

Jeff Jarvis [02:01:08]:
Can't trust anything.

Leo Laporte [02:01:09]:
Let me see here how Apple created a custom. No, Apple. I don't know. Yeah, they were promoting. They were at the time promoting the movie. But I think maybe Apple had greater aspirations. Let's take a little break. More to come, I think.

Leo Laporte [02:01:27]:
A little break. More to come in just a moment. You're watching Intelligent Machines with Jeff Jarvis and the wonderful Paris Martineau. Our show today brought to you and our website, I might add, brought to you by Pantheon. I know Patrick Delahanty, our web engineer, is listening. He loves Pantheon. It was kind of amazing. Pantheon approached us and said, hey, you think you could do an ad for us? I said, yeah, I think we could do an ad.

Leo Laporte [02:01:52]:
We use you. You run our entire workflow. Our website is built on Pantheon, but not just Pantheon, but not just our website. When the editors publish a show, it's running through a Drupal instance running on Pantheon. Your website is, and as true for us, too, your number one revenue channel.

Jeff Jarvis [02:02:12]:
Right?

Leo Laporte [02:02:13]:
But if it's slow, if it's down, you know, users don't wait. If you don't pop up in an instant, they're going to go somewhere else. So when your website's stuck in a bottleneck, it's also your number one liability. Pantheon keeps your site fast and secure and always on. That means better SEO, more conversions, no lost sales from downtime. It's not just a business where. Ask Patrick. It's a developer win, too.

Leo Laporte [02:02:40]:
Your team gets automated workflows, isolated test environments. That saved my bacon when I pushed a. Well, a flawed update, shall we say. But I pushed it to test, not production. Thank goodness it was isolated so there was no downtime. Thank you, Pantheon. No late night fire drills? No.

Leo's Laptop Audio [02:03:01]:
I don't know.

Leo Laporte [02:03:01]:
Works on my machine. Headaches. Just pure innovation. And everybody in the team loves it. Marketing can launch a landing page without waiting for a release cycle. Developers could push features with total confidence. Even if Leo's pushing them. And your customers, they just see a site that works 24.

Leo Laporte [02:03:18]:
7 Pantheon powers Drupal and WordPress sites. We use Drupal, but they also do WordPress that reach over a billion unique monthly visitors.

Jeff Jarvis [02:03:29]:
Wow.

Leo Laporte [02:03:30]:
Visit Pantheon IO and make your website your unfair advantage. We love Pantheon. Pantheon, where the web just works. And it works for us. I have to say, we love it. Patrick posted a picture of a cheering tiger to celebrate.

Jeff Jarvis [02:03:51]:
There.

Leo Laporte [02:03:52]:
Go right, he says. There you go. That's an endorsement from Patrick Delahanney and Tony the Tiger. What else? Is there anything else we want to talk about before we wrap things up? It's getting a little late. The sun's gone over the yard arm as my.

Jeff Jarvis [02:04:12]:
Over here. It's way over.

Leo Laporte [02:04:13]:
Yeah, my grandpa used to say. But that was an excuse for him to pour a gin and tonic. I don't know.

Jeff Jarvis [02:04:18]:
As my father used to say, somewhere in the world. It's five o'.

Leo Laporte [02:04:20]:
Clock. Yeah, that's another one.

Paris Martineau [02:04:22]:
One.

Jeff Jarvis [02:04:23]:
Wait, wait, wait. So you have. You have my father. Would my kids want to know whether this was just popup or there was other popups? Did he call gasoline Juline?

Leo Laporte [02:04:35]:
No.

Paris Martineau [02:04:36]:
Juline. Is he drinking the gasoline?

Leo Laporte [02:04:38]:
I think that's unique.

Jeff Jarvis [02:04:39]:
My father.

Leo Laporte [02:04:40]:
I think that's one of a kind. I think that's your father.

Jeff Jarvis [02:04:42]:
How about milk as moo juice?

Leo Laporte [02:04:44]:
Moo juice. I've heard. Yes.

Jeff Jarvis [02:04:45]:
You've heard that one.

Leo Laporte [02:04:46]:
Oh, yes.

Jeff Jarvis [02:04:47]:
Paris thinks that's strange.

Leo Laporte [02:04:49]:
You don't really want to focus on the source of either milk or eggs or bacon for that matter.

Jeff Jarvis [02:04:55]:
True, true.

Leo Laporte [02:04:56]:
Elon Musk is preparing for the release of Optimus Jesus the robot. The Tesla annual meeting is next month and one of the centerpieces, at least one of the things Elon hopes to do to impress shareholders will be a dancing troupe of Optimus robot robots. Elon says he believes that the robots will make more for Tesla than the cars do. However, as the information reports. The 08 reports the information there's been some technical difficulties. They had hoped to make thousands of Optimus robots this year. Those plans were abandoned by summer, according to two people with direct knowledge of the Optimus program. It's hard Right.

Leo Laporte [02:05:50]:
These are, these are, you know, two legged machines. He has shared pictures of them walking around and building structures on Mars. This is his plan to get them there by 2027. When is that? Well, that's just two years off. And he wants Optimus to be on board that SpaceX Mars trip. Unfortunately, the current version.

Benito Gonzalez [02:06:14]:
Sorry for the worker that has to wear that costume.

Leo Laporte [02:06:17]:
I know the current version of Optimus is designed for indoor use on Earth, which means probably not going to work on Mars unless it wears a spacesuit or has a redesigned thermal management system fit for a dusty planet where the Average temperature is 80 degrees below zero. Well, we'll see. I hope he has some dancing troupe. Elon acknowledged last month on the all in podcast that Tesla's struggling with the final design of the hardware and that the hands inclusive of the forearm are a majority of the engineering difficulty of the entire robot. We got the torso down, man, but the arms, not so easy.

Jeff Jarvis [02:07:03]:
Otherwise known as a food delivery delivery robot.

Leo Laporte [02:07:06]:
Yeah, yeah. Don't they use those at the. Don't they use those at the Tesla Charge supercharger in LA at the fast food place? In fact, there were videos to make middling popcorn, right? Yeah. And, and somebody having, getting it frozen and then somebody had to come in and hit the reset switch. Let me see if I could find that, that video.

Benito Gonzalez [02:07:31]:
There's a time in San Francisco downtown where all these like pop up robot like coffee makers and all that stuff like started popping up. Like a robot would make a coffee or a robot would make your burger.

Leo Laporte [02:07:41]:
Oh yeah, they're on cruise ships.

Paris Martineau [02:07:42]:
There's always a couple of those gimmick things going on.

Leo Laporte [02:07:45]:
Here's, here's the Optimus.

Paris Martineau [02:07:47]:
Optimus is frozen.

Leo Laporte [02:07:49]:
It's frozen. So the Tesla engineer on the phone, he's coming in, he's going to.

Jeff Jarvis [02:07:54]:
On the phone with support.

Paris Martineau [02:07:57]:
This is a remix of the Wii Me channel theme.

Leo Laporte [02:08:02]:
Okay, I shouldn't play the music. That's probably what's going on in the background. There he is. He's. Is it. Okay, you see, see that port on the left? You flip that switch.

Jeff Jarvis [02:08:11]:
Oh, no, it didn't work. Is there a plug?

Leo Laporte [02:08:13]:
I got a.

Paris Martineau [02:08:14]:
You got a. Does anybody have a paperclip?

Leo Laporte [02:08:16]:
You're gonna poke the reset button. Oh, oh, oh, I see. Oh, it's. Oh, oh, it's moving. It's moving. See, as Elon said, it's the, it's the hands and really the hands and the forearms that are so really the.

Paris Martineau [02:08:32]:
Intricate parts of it that are essential to its daily function.

Leo Laporte [02:08:36]:
I have made popcorn for you. Human. It didn't fill the bag up. I might point out. That's just. I get a half a bag. Bye. And the guy's still on the phone.

Leo Laporte [02:08:47]:
Phone.

Paris Martineau [02:08:51]:
They are pretty cool looking, we'll give them that.

Leo Laporte [02:08:53]:
Hey, this is a big step forward from the guy in the suit that was dancing around. It's at least able to do something. What else AMD did? We talked about the AMD OpenAI deal, right? Maybe we.

Jeff Jarvis [02:09:11]:
No, we didn't.

Leo Laporte [02:09:12]:
No, that was Windows Weekly. Another one of these big round trip circulars. Yep. Where both companies benefit but nobody actually spends any money. And of course AMD stock rockets higher on multi billion dollar OpenAI deal. AMD is going to provide 6 gigawatts of GPUs to the AI juggernaut over several generations. As part of the deal, OpenAI will take a stake in AMD working worth 10% of the company. Wow.

Leo Laporte [02:09:45]:
Okay. I guess in effect that is OpenAI buying GPUs. Oh no, wait a minute. They get.

Jeff Jarvis [02:09:51]:
No, it's right.

Paris Martineau [02:09:52]:
No, it's all just circular.

Leo Laporte [02:09:55]:
Nobody pays anything, but everybody comes out a winner. And that's all I care about. And that's as long as my retirement.

Paris Martineau [02:10:03]:
Everybody's patting each other on the school.

Leo Laporte [02:10:07]:
I mean, you made money on the stock market today. I made a lot of money on the stock. Don't knock it. It's funny money because I.

Paris Martineau [02:10:13]:
You'll never lose it.

Jeff Jarvis [02:10:15]:
Well, go back to the Bezos argument. Is that an industrial.

Leo Laporte [02:10:19]:
I don't know. Google has announced an AI vulnerability reward program which despite its name, does not mean they're rewarding you for creating AI vulnerabilities. They are.

Jeff Jarvis [02:10:32]:
How about for being a vulnerable human being?

Leo Laporte [02:10:34]:
They have released a new AI powered agent called called Codemender. Which improves your code's security automatically the.

Jeff Jarvis [02:10:44]:
Problem and fixes it.

Leo Laporte [02:10:45]:
Yes. You know, I think it's probable. What do you think, Darren? Can codemender fix code and make it more secure? Codemander, Google says, is an AI powered agent utilizing the advanced reasoning capabilities of our Gemini models to automatically fix critical code vulnerabilities. Sometimes AI fixes them by changing the bit from failure to success, but sometimes it does more. We'll see. Root cause analysis. They're using fuzzing and theorem provers to identify the fundamental cause of a vulnerability. This is important not just as surface symptoms, self validated patching.

Leo Laporte [02:11:33]:
It's worth a try, right? It's worth a try. Let's see.

Jeff Jarvis [02:11:41]:
Well, while you're on. So Google did a Couple other things that are.

Leo Laporte [02:11:44]:
Yes, we should go back to our roots. Let's talk about some Google.

Jeff Jarvis [02:11:47]:
So they introduced Gemini 2.5 computer use model. And they introduced and they wrote about embedding Gemma which are both interesting.

Paris Martineau [02:11:59]:
Who's Gemma?

Jeff Jarvis [02:12:00]:
Gemma. That is their open source.

Paris Martineau [02:12:02]:
Oh it's the small model.

Leo Laporte [02:12:03]:
It's for the phone.

Jeff Jarvis [02:12:05]:
So this is, this is one you can embed on any device. So it's a small model. That's good. And also the 2.5 computer use is kind of agentic in that it can mimic your activity mimic activities on a browser.

Leo Laporte [02:12:17]:
It can power agents that can interact with user interfaces so it could push the buttons for you.

Jeff Jarvis [02:12:24]:
So this is not unlike OpenAI's. I can go and deal with Canva for you. Google. This is I think mvc and they can deal with.

Leo Laporte [02:12:33]:
It's really interesting to see all the parity, you know how quickly this is happening where a company A will do something, then company B says I could do that, then company A will do. Well, I can do that. But can you do that?

Benito Gonzalez [02:12:44]:
Can we stop it with the captchas already? Because presumably this would be able to solve.

Paris Martineau [02:12:48]:
Right?

Jeff Jarvis [02:12:49]:
Please.

Leo Laporte [02:12:49]:
We've solved.

Paris Martineau [02:12:50]:
Yeah.

Leo Laporte [02:12:52]:
So here's a prompt.

Paris Martineau [02:12:53]:
As I've said in this show many times, I have browsing habits that every website seems to think me makes. I'm a computer means I'm a computer.

Leo's Laptop Audio [02:13:01]:
Yeah.

Paris Martineau [02:13:02]:
I'm doing five to 15 captures a day minimum. I'd love to.

Jeff Jarvis [02:13:06]:
Not because you're. What habits you have.

Leo Laporte [02:13:08]:
Paris. Look me in the eyes. Are you a werewolf?

Paris Martineau [02:13:12]:
I'll never tell.

Leo Laporte [02:13:15]:
Here's the prompt. Watch the video while I tell you what the prompt was for this from a website. Get all the details for any pet with a California residency. What? And add them as a guest in my spa CRM. Then set up a follow up visit appointment with a specialist and a mom out of a bar for October 10th anytime after 8am the reason for the visit is the same as their requested treatment. So that's a fairly complicated prompt. Basically it's a customer. So they're adding new clients from the website they're going to is their sign up.

Jeff Jarvis [02:13:53]:
This is this new Gemini code use computer use model.

Leo Laporte [02:13:57]:
Yeah. So it's agentic.

Jeff Jarvis [02:14:01]:
Yeah.

Leo Laporte [02:14:01]:
Yeah.

Jeff Jarvis [02:14:04]:
The agents are here.

Leo Laporte [02:14:08]:
I have to say this goes back to the conversation about OpenAI being an operating system. It does kind of make sense that the ways we interact with technology, the ways we search for websites, the ways we interact with websites might in fact there might be a news user interface that is more natural. I always thought that with the voice assistants, I thought that was going to be the next thing in user interface. It wasn't because it turned out like talking. Well, I think they just weren't very good either. Right. Amazon's Echo and Siri.

Paris Martineau [02:14:42]:
No one knows what they're supposed to be used for, right? I mean, even the other day, Risa is like, I've got a Amazon Echo or whatever and she's like, I don't know what to use it for other than timers. And she runs an AI company that's powered by voice.

Leo Laporte [02:14:56]:
Well, this morning my Amazon Echo show decided to go agentic. Alexa, I love to say that. Tell me what you can do. Tell me what you can do, Alexa.

Leo's Laptop Audio [02:15:12]:
Oh, you're curious about what I can do. I can show you what's new or what's most popular. Which one sounds fun to you?

Leo Laporte [02:15:19]:
What's most popular?

Leo's Laptop Audio [02:15:23]:
Awesome. Let me share what's super popular with everyone. I can create music playlists, set alarms with music, and give weather updates. Which one sounds most interesting to you?

Leo Laporte [02:15:34]:
I'd like to know if we're going to see any rain in our area in the next 10 days because my house is, well, lacking the south wall, shall we say?

Leo's Laptop Audio [02:15:46]:
There's a significant chance of rain next week. On October 13, there's an 84% chance of rain during during the day with up to 0.34 inches expected.

Leo Laporte [02:15:55]:
Oh, that's not.

Leo's Laptop Audio [02:15:56]:
Following day has a that's not significant chance of rain with potentially 0.74 inches.

Leo Laporte [02:16:01]:
Almost an inch. Is that considered heavy rain? Will it be windy? Oh, it says almost an inch of rain could be problematic for your wall.

Leo's Laptop Audio [02:16:11]:
Expect winds up to 10 miles per hour from the west southwest. The next day, winds will be lighter around 7 mph from the south southwest.

Leo Laporte [02:16:20]:
What should I do to protect my house? It doesn't have a south wall.

Jeff Jarvis [02:16:31]:
You're screwed on.

Leo's Laptop Audio [02:16:32]:
Strengthening your roof, windows and doors with stormproof materials. Ensure impact rated windows, Reinforced doors, not quite there. Also, clean and maintain your gutters to properly direct them.

Leo Laporte [02:16:47]:
No, no, you don't understand. I don't have a wall on my house.

Jeff Jarvis [02:16:54]:
What do you expect me to do? I'm just a voice situation earlier.

Leo's Laptop Audio [02:16:58]:
For a house without walls, you'll need to focus on protecting the structure with temporary coverings and ensuring proper drainage around the foundation. Given the upcoming rain, it might be wise to consult with the local contractor about immediate protection.

Leo Laporte [02:17:10]:
Oh, believe me, I have. But he's on vacation again. Third vacation in.

Leo's Laptop Audio [02:17:18]:
I understand your frustration since your Contractors now, now it's.

Jeff Jarvis [02:17:22]:
Let's focus on immediate customer service voice.

Leo's Laptop Audio [02:17:24]:
Consider using heavy duty tarps to cover exposed areas. Make sure any temporary barriers are securely anchored to withstand the upcoming rain.

Leo Laporte [02:17:32]:
I gotta send this to my contractor. So I asked the AI.

Paris Martineau [02:17:39]:
I asked.

Leo's Laptop Audio [02:17:40]:
Google, found several contacts that might be your contractor. Would you like to.

Leo Laporte [02:17:44]:
Oh, don't say names. Don't say names. Stop. Alexa, stop. Oh, stop, stop. Oh, plug it. No, no, no.

Jeff Jarvis [02:17:52]:
We're all going to.

Paris Martineau [02:17:54]:
Press the buzz button.

Jeff Jarvis [02:17:55]:
Use your, yeah, your bump.

Paris Martineau [02:17:57]:
Use your sensor button, your dump button.

Jeff Jarvis [02:17:59]:
You got a dump button.

Paris Martineau [02:18:01]:
Your. Bleep us, Leo. You could have bleeped it finally. Yeah, a little late, but.

Leo Laporte [02:18:13]:
Well, I mean that's better than it used to be, I guess.

Jeff Jarvis [02:18:18]:
Now I think Google is also switching now finally to Gemini instead of Assistant on the devices too.

Leo Laporte [02:18:27]:
This. So if you asked me five years ago what's going to be the big breakthrough in user interface, I would say voice. You're right.

Jeff Jarvis [02:18:34]:
I think we did say it on the show.

Leo Laporte [02:18:35]:
Yeah, but you're right, Paris, nobody. There turns out to be problems talking out loud of these things and especially in an environment where there are other people especially in the environment where there are other people who are tricksters like my kids, who every time I would start talking to the AI, would start blabbering other things, it was problematic. So. But I still believe that there is something there. I don't know. I don't know.

Jeff Jarvis [02:19:00]:
So I have a quick question for our young person. Yes, The Wall street journal online 127 is insisting that young people are falling in love with old technology. And Maureen, Joe just loved this story today. Do you know, do you have friends who are using flip phones, who are using CDs, who are using cameras? You do?

Paris Martineau [02:19:21]:
Yes, to literally all of those. One of my, my friend Rick, the six foot nine Rick. I'm recently quite mad at him because he's gotten a, his phone broke and he got a flip phone instead. And I'm like Rick, we're in a bunch of group chats. You can't text us back if you have a flip phone. And he's like, yeah, but you know, I just think it'd be better to be disconnected. Oh no, he sound like every guy on hinge.

Leo Laporte [02:19:48]:
Have hinges as well.

Benito Gonzalez [02:19:49]:
So what's the pool on him switching to a smartphone soon?

Paris Martineau [02:19:53]:
I, I brought to the bar the other day, I have three old iPhones and I brought in my oldest iPhone, a 10 year old one. I was like Rick, you could use.

Jeff Jarvis [02:20:02]:
This at least it's old.

Paris Martineau [02:20:03]:
Yeah. I was like. I was like, it's vintage old. I was like, you can't update it to the latest iOS. Isn't that fun? It still has the finger.

Leo Laporte [02:20:13]:
I don't understand what's wrong with him wanting to use Flip Flop?

Paris Martineau [02:20:19]:
Nothing.

Leo Laporte [02:20:19]:
Is that a deal breaker?

Paris Martineau [02:20:20]:
No, there's nothing wrong with it as I said on Friday because I've been giving him a little bit too much on it because I was like, rick, I'm sorry I've been giving you crap on the flip phone. I respect your choice. I just treasure your friendship and enjoy texting you. And it's difficult to not be able to. I guess I can start calling you now. But it's annoying that you're no longer in any of our group chats. Basically.

Leo Laporte [02:20:43]:
It's true. Really? It cuts you off, doesn't it? Nowadays, yes.

Paris Martineau [02:20:47]:
So this is definitely a problem. A bunch of people have digital old point and shoot digital cameras.

Leo Laporte [02:20:53]:
Aren't you glad you didn't do that? You're still using the Project Indigo.

Paris Martineau [02:20:58]:
I love Project Indigo. I use it all the time. I also though I recently. So I use a DSLR Canon oh camera as my webcam and I recently got a. Because it's on like a little tripod thing behind my computer. I recently got a. Like a magnetic mount so I can easily remove it from it and then take it.

Leo Laporte [02:21:16]:
Smart. So now you can use that too. Good.

Paris Martineau [02:21:18]:
So my friend, a friend. A friend was like filming a short film this weekend and I was there as an extra. It's kind of a long day shoot and this is too much information. But they were having some last minute emergencies to where someone at this party was like Paris. Maybe bring a backup camera in case our camera. I brought my camera. Their camera was fine. But then I was like, why? Well, I guess I'll just take a bunch of photos behind the scenes.

Paris Martineau [02:21:41]:
And now everybody had a great time and everybody loves my photos.

Leo Laporte [02:21:43]:
You have the behind the scenes photos.

Paris Martineau [02:21:45]:
Having a nice.

Leo Laporte [02:21:45]:
Do not upgrade your iPhone. Are you an iPhone user?

Paris Martineau [02:21:50]:
I'm not upgrading to the new iOS.

Leo Laporte [02:21:52]:
I was going to say because rather die.

Paris Martineau [02:21:54]:
I know, I know that Project Indigo.

Leo Laporte [02:21:56]:
Says we apologize for the inconvenience but Indigo is not supported on this.

Paris Martineau [02:22:00]:
Well, they said that they haven't. They didn't have access to the new iPhones until they came out.

Jeff Jarvis [02:22:04]:
Out.

Paris Martineau [02:22:04]:
I've been able to start testing it now. I really. I've talked about Project Indigo before. It was recommended to me by a listener. Shout out. I've said your Name before. I don't have it on hand now, but thank you. You know who you are, and it's just great.

Paris Martineau [02:22:18]:
It really uses all of the powers of your camera and takes RAW and JPEG photos that then you can edit on your phone using Lightroom or whatever editing software you have. And they look phenomenal. So much better than what you'd take normally.

Jeff Jarvis [02:22:36]:
So it's Adobe kind of.

Paris Martineau [02:22:37]:
Yeah, it's Adobe. They process.

Leo Laporte [02:22:39]:
It's ironic that Adobe didn't have access to Apple's latest iPhones. Yeah, there's something that tells you something there. Good.

Paris Martineau [02:22:47]:
I will use Project Indigo Team is just a. A small, kind of like Moonshot Labs group.

Leo Laporte [02:22:53]:
I'll use it to take pictures of all the water damage next week. So that'll be the one thing you've.

Paris Martineau [02:22:57]:
Got to do, though. Remember, Indigo is like you processes the photos after you take it. Like it kind of develops them in some way. Like it runs it through its kind of own algorithm. And you have to keep your phone open on the app till they're done developing. It only takes like 15, 30 seconds, but if you close the app beforehand.

Leo Laporte [02:23:18]:
You lose it can't run in the background. Yeah, yeah. Hey, before we get to our picks of the week, which are just around the corner, you're watching. And Intelligent Machines with Jeff Jarvis and Paris Martineau, a little note from one of our good friends, a guy who's been on the show and we love a lot. Glenn Fleischman. I think this is from you, Jeff.

Jeff Jarvis [02:23:35]:
You put this in.

Leo Laporte [02:23:36]:
And I did not know this. He posted yesterday on LinkedIn that he's getting open heart surgery next month. And while his insurance does cover, of course, the surgery and all the testing and procedures, it's still expensive after the surgery.

Jeff Jarvis [02:23:53]:
He and his wife are freelancers.

Leo Laporte [02:23:55]:
They're both freelancers and he's not going to be able to work at all for a few weeks. We'll get him on the show and we'll pay him to be on the show. But he says, I would never ask for charity, but if you would like to buy one of my books for a birthday or holiday gift, that would be great. He's got six centuries of type and printing, great for typo files at 6 cent, c e n T Info, and he has autographed versions available. Or of course, his latest, How Comics Are Made, which is a beautiful book. Really nice. Interviews with Doonesbury's Gary Trudeau, Lynn Johnston, Bill Waterston. That would be well worth.

Leo Laporte [02:24:34]:
I'm gonna buy some copies as gifts.

Jeff Jarvis [02:24:37]:
Because I Already have both books.

Leo Laporte [02:24:39]:
Yeah, me too. Oh, let me go back to the post.

Jeff Jarvis [02:24:47]:
It's about the history of print.

Leo Laporte [02:24:50]:
Yeah, he's on LinkedIn. This is from LinkedIn. It is six cent.info info info. Yeah.

Paris Martineau [02:24:58]:
Sounds like a great book. And I'm.

Leo Laporte [02:24:59]:
Yeah, it is.

Jeff Jarvis [02:25:00]:
It's lovely.

Leo Laporte [02:25:00]:
And Glenn's a wonderful guy. He says the prognosis is good. So I'm sorry about his health issues, but I'm glad that he's getting treatment and he says the prognosis is excellent and if anybody wants to support him by buying winning those two books, please do.

Jeff Jarvis [02:25:14]:
Yeah, I also wish he'd go ahead and put up a GoFundMe.

Leo Laporte [02:25:16]:
But yeah, I don't blame him. He doesn't want to. He doesn't want to do that. I understand that. I really do love you, Glenn. Yep, we do feel better. Get better soon.

Jeff Jarvis [02:25:29]:
Yep. Yep. Take it up. Ya.

Leo Laporte [02:25:30]:
Yep. Let's take a break. More to come. Well, our picks of the week, in fact, as we continue with Intelligent Machines. We do this show by the way, every Wednesday right after Windows Weekly. That's 2pm Pacific, 5pm Eastern. That is 2100 UTC. You can watch us on YouTube, Twitch, TikTok, Facebook, LinkedIn, X.com and Kik.

Leo Laporte [02:25:55]:
Next week Jeffrey Quinnell is going to join us from New Research, N O U S Research. They are doing AI models, human centric AI models that are focused on aligning AI with real world user experiences. It's a very interesting approach to AI and one of their models, Psyche, is supposed to be emotionally intelligent. So Jeffrey will be our guest next week on Intelligent Machine. That should be very, very interesting. If you are a fan of the show and you want to support it. If you like what we do on our other shows, I would encourage you to join the club. We'd really appreciate that.

Leo Laporte [02:26:37]:
25% of our operating expenses come from your membership and I expect as time goes by that's going to be more and more. It lets us do some of the fun things we do in the club. You get access to the Discord. You get ad free versions of the shows because you're giving us 10 bucks. We don't need to show you ads. But most importantly it. It supports what we do. It's a vote and we really appreciate your support.

Leo Laporte [02:26:58]:
Twit TV club Twit. And thanks to all of our existing members. We really, really appreciate it. Now let's get the pick of the week from Miss Paris Martineau.

Paris Martineau [02:27:10]:
My pick this week is a podcast called the Necessary Conversation it's by Chad Colchin, who's a podcaster, his sister Haley, Pop. And their parents, Mary Lou and Bob. Culture. And it's got kind of an interesting premise, which is started in 2022 because Chad and Haley are various degrees of kind of left leaning politically, and their parents are kind of full on Kool Aid, drinking maga. The parents names again, Mary Lou and Bob.

Jeff Jarvis [02:27:48]:
Mary Lou and Bob. Just perfect.

Paris Martineau [02:27:49]:
Perfect. Yeah. No. And so every week they meet and this is. Honestly, they meet and they talk about politics.

Leo Laporte [02:27:57]:
I don't know if I could take that. Is it stressful?

Paris Martineau [02:28:00]:
It is stressful, but in some ways it's kind of a relief. Therapeutic and a relief.

Leo Laporte [02:28:07]:
And because they still love each other at the end of the show, right? Or do they walk out?

Paris Martineau [02:28:13]:
I mean, I think they love each other as much as they began at the end of the show, but that's. It depends. It's just. It's a very interesting dynamic and they're all very earnest and engaged or at least earnest and forthcoming about their beliefs despite wild disagreements and. I don't know, just a very interesting premise for a show.

Jeff Jarvis [02:28:33]:
Brave to do it and brave to listen.

Paris Martineau [02:28:35]:
It is. Yeah.

Leo Laporte [02:28:36]:
I don't know if I could take it.

Jeff Jarvis [02:28:37]:
No, I'm not sure.

Leo Laporte [02:28:38]:
It's a little close to home. Home.

Paris Martineau [02:28:40]:
It is a little close to home. And I mean, if you don't know if you can take it, check out their account on Tick Tock or Instagram. They post little clips.

Leo Laporte [02:28:49]:
Okay. Okay.

Paris Martineau [02:28:50]:
It's often Chad asking their parents basic questions and then trying to figure out what the heck they just answered. And sometimes vice versa. But I don't know, I just. I find it a very interesting premise and it's interesting given the candor that all of the members kind of take to it. So I'd really recommend it. I really like Chad's other podcast, Game of Roses. I think I've talked about before on the show, which kind of explores reality TV through the lens of professional sports. But that's how I got into this.

Leo Laporte [02:29:21]:
The Necessary conversation. It's on YouTube at the necessary Conversation. Family therapy through politics. Wow. I don't. I don't know if I could take it.

Paris Martineau [02:29:34]:
It's a lot. And I guess while we're talking about podcasts, I'll shout out some listener reviews of our pod.

Leo Laporte [02:29:41]:
Oh, yes.

Jeff Jarvis [02:29:41]:
Oh, yeah.

Paris Martineau [02:29:42]:
Please do want to have your podcast review shouted out? You can leave a review on Apple Podcasts of this show. You can leave it if you love the podcast or hate it, but it would be Great for us. If you left five stars could balance it up. I'll give you two examples here of a. Of a lover and a hater lover. My reviews Tech wrote. We love Paris.

Leo Laporte [02:30:08]:
Get that red on the air. Huh?

Paris Martineau [02:30:10]:
A show that boldly goes where Nobod has gone before serving a delightful cocktail of AI antics and wit. I only wish there was a more masterfully catchy theme. Songs that eat through the ear like an earworm. Reminiscent of the baby shark phenomenon on. And then as a counterpoint. I guess I should have done this first. We ended on a positive note. Mountaineer says two of the hosts are just AI cynics.

Paris Martineau [02:30:33]:
I'm all for skepticism, but two of the hosts just knee jerk hate everything. Paris is by far the worst offender. And it goes on another one.

Leo Laporte [02:30:43]:
Wait a minute. Who's the other hater? I think I. We.

Paris Martineau [02:30:46]:
I guess it's Jeff.

Leo Laporte [02:30:47]:
Must be Jeff.

Paris Martineau [02:30:48]:
He is a hater.

Jeff Jarvis [02:30:49]:
No, you're.

Paris Martineau [02:30:51]:
I was gonna say, as to AI podcast, I would say he's the man who devotes the most thought to AI every week. And then we get another incredible show from the Twit network, writes Bionic Geek. I've been listening to this show for years, including in its previous format. It's informative, has great analysis on current topics, covering AI and many different ways and aspects that tech impacts our live today. It's also very entertaining. Entertaining. Fantastic show. Rate and reviewer, podcast folks.

Leo Laporte [02:31:20]:
Thank you. And here is Dustin's image art for this week's show. This week in hipsters.

Paris Martineau [02:31:29]:
Oh, I do like a jacket, actually.

Leo Laporte [02:31:32]:
Oh, do you? I always. I want to get a denim jacket. I haven't had one in a long time.

Benito Gonzalez [02:31:36]:
Oh, my God. Are we due for hipsters to come back? I think that's about to swing back around, isn't it?

Leo Laporte [02:31:40]:
Hipster. Hipster are back again. Yeah. Should I grow a beard? I don't know. That looks pretty good. I think that's a good look. Maybe I should grow a beard, be.

Paris Martineau [02:31:48]:
In my other closet. I do have that exact jacket, but it was my grandfather's.

Leo Laporte [02:31:53]:
Oh, does it have a little happy face button on it?

Paris Martineau [02:31:56]:
No, but the back has a Harley Davidson American flag, which is why it's in probably a different closet right now. It's a little weird.

Leo Laporte [02:32:05]:
Oh, you can wear the flag. We actually have a flag bag.

Paris Martineau [02:32:10]:
The beard would look good. Can you grow a beard, Leo?

Leo Laporte [02:32:12]:
Of course I can grow a beard. I can grow that beard.

Jeff Jarvis [02:32:14]:
Are you questioning his.

Leo Laporte [02:32:16]:
I don't know.

Paris Martineau [02:32:18]:
I'm not questioning your manliness. It just Seems you seem so.

Leo Laporte [02:32:22]:
I was always taught that it's. For some reason I wear contacts that you want to have as much of your face available for expressiveness.

Jeff Jarvis [02:32:30]:
And this nose, I want to hide it. I want a beard from it.

Leo Laporte [02:32:33]:
Well, I. I'm. I'm getting to the point where I kind of want to hide my turkey jowl neck. So there's the beautiful Paris Martineau. See, poor Paris. Her sister was the model, so she had to be the smart one.

Paris Martineau [02:32:46]:
It's true.

Leo Laporte [02:32:47]:
And I think it's un.

Paris Martineau [02:32:48]:
I came first, so, you know, I just capitalized in the.

Leo Laporte [02:32:50]:
Did you choose to be the smart one and she chose to be the pretty one or is. Because that happens in families, doesn't it?

Paris Martineau [02:32:57]:
I don't think I ever really thought about it until one time, which I think I've mentioned this anecdote in the show before. I was maybe, like, 8 or 9 or 13. I was some level of adolescence, and my parents were up late talking to a friend in, like, the main living room area. And I was sitting around the hall, like, listening to them. Gizmo keeps putting her little tail in front of the camera because I'm a mic. I was listening to them, and I heard them saying, oh, well, you know, Melisse, My sister is, like, so extroverted. It's all right. Paris is introverted, but she and I literally never before in my life thought of myself as introverted or nerdy.

Leo Laporte [02:33:31]:
And I was like, oh.

Paris Martineau [02:33:34]:
I was like, I do sports. How am I introverted or nerdy?

Leo Laporte [02:33:38]:
But isn't that funny how that the dynamics of families are so weird? So weird, so unhealthy? It's too bad we need them. Jeff Jarvis, pick of the week.

Jeff Jarvis [02:33:51]:
Okay, so. So about three weeks ago, I started seeing in my TikTok, my German TikTok, a major trend which I think is coming to America. The Washington Post has written about it, and I think it could probably hit Paris's friends and neighborhood. Go to line 162. Pudding mit gobble. Pudding with a fork.

Leo Laporte [02:34:11]:
No. 9. One does not eat pudding with a fork.

Jeff Jarvis [02:34:15]:
The Washington Post. And go to.

Leo Laporte [02:34:18]:
You have another link.

Jeff Jarvis [02:34:19]:
Yeah, hold on. No, I got the wrong link in there. Is it.

Leo Laporte [02:34:24]:
Is the pudding the same? It's just the utensil that changes.

Jeff Jarvis [02:34:27]:
It's the utensil video.

Leo Laporte [02:34:28]:
Auf English. That one.

Jeff Jarvis [02:34:31]:
No, I want to put a different one in there. Hold on one second. Put it in and out. I thought I'd erased it. Erase, boom, Done. Okay.

Leo Laporte [02:34:41]:
How can you call eating pudding with a fork a trick? A Trend.

Jeff Jarvis [02:34:44]:
It doesn't watch.

Paris Martineau [02:34:46]:
Watch people that are there eating pudding with a fork.

Jeff Jarvis [02:34:51]:
They bring the pudding, they tap the pudding top with the forks and then they eat it. Nobody can figure out. So What? Go to 162 now. No, no, it has to be a fork. A gobble.

Leo Laporte [02:35:02]:
A gobble. Ein gabel.

Jeff Jarvis [02:35:04]:
Ein gabble. Okay, so line 162.

Leo Laporte [02:35:08]:
All right.

Paris Martineau [02:35:09]:
That we are now gonna watch strangest TikTok videos.

Benito Gonzalez [02:35:12]:
Yeah, she does.

Leo Laporte [02:35:13]:
Does. He does.

Paris Martineau [02:35:14]:
The events are BYO and byof. And even on a Sunday when German stores and supermarkets are closed, everyone comes prepared.

Leo Laporte [02:35:26]:
All right, here we go.

Paris Martineau [02:35:30]:
At the appointed time English speaking world.

Leo's Laptop Audio [02:35:32]:
Is aware of and I need to make you aware of it. It is called putting, which means pudding with fork. And people are meeting up in Germany and Austria right now in the hundreds in various cities to eat pudding together.

Paris Martineau [02:35:44]:
With what?

Leo's Laptop Audio [02:35:44]:
I have done a deep dive on this. I have watched tons of videos on it. You can look it up right now. Pudding midgar. And I cannot find the source. I cannot find the pudding midgaba source. That looks like yoga videos coming out of Vienna from yesterday at this meetup in Vienna. There might have been a thousand people in that park.

Leo's Laptop Audio [02:36:02]:
Each with their own pudding, each with their own go.

Paris Martineau [02:36:05]:
They're.

Leo's Laptop Audio [02:36:05]:
They're tapping the lid rhythmically to get ready to eat the pudding together. There's all ages, there's kids, there's omas. Everyone is doing this. It's stuff like this that makes me laugh when people say that Germans aren't funny. I know Vienna isn't in Germany, but bear with me, the German speaking world, because this to me is peak German humor. There's no deeper meaning to it. It is pure absurdism. And yet it is so funny.

Jeff Jarvis [02:36:27]:
And this exemplifies how fluent German speaking student in Germany, American obviously.

Leo Laporte [02:36:33]:
That's awesome.

Paris Martineau [02:36:35]:
The first Pudding McGobble took place in the southwestern German city of Carlsruhe, best known as the site of the German of the country's supreme court in late August. Apparently young people around Germany thought it was a tremendous idea.

Jeff Jarvis [02:37:05]:
So I can find friends.

Leo Laporte [02:37:06]:
Yeah, friend finding. So tapping with a fork is. Is also meaningless. It's just part of the rich part of the ritual.

Jeff Jarvis [02:37:12]:
Yes. It's going to be coming to Prospect park before you know it.

Leo Laporte [02:37:18]:
So just be aware when you start seeing people with their gobbles.

Jeff Jarvis [02:37:21]:
Paris, you can start it. You can start it.

Paris Martineau [02:37:23]:
The issue is I'm not a pudding fan because it's too soft as a.

Leo Laporte [02:37:28]:
Would you like hard pudding? Would you enjoy a hard Pudding.

Paris Martineau [02:37:31]:
I want a couple of layers of like a crunchy Oreo. Something mixed.

Leo Laporte [02:37:36]:
I think you.

Jeff Jarvis [02:37:37]:
You sneak that in there.

Leo Laporte [02:37:38]:
You could put crunch in a pudding. Yeah.

Jeff Jarvis [02:37:40]:
Yogurt.

Paris Martineau [02:37:41]:
I would have to get my own pudding.

Leo Laporte [02:37:43]:
It's too soft. Do you like.

Jeff Jarvis [02:37:46]:
That's weird.

Leo Laporte [02:37:46]:
Paris chocolate mousse. Is that too soft?

Paris Martineau [02:37:48]:
Too soft. The only thing that has broken my soft aversion is grits, which I think are a perfect.

Leo Laporte [02:37:56]:
Grits are good oatmeal. So you'll eat oatmeal. Flan is pudding.

Paris Martineau [02:38:02]:
No, I don't really like it. I only like creme brulee.

Leo Laporte [02:38:04]:
Creme brulee because it's got a crispy.

Paris Martineau [02:38:06]:
Top, but I only eat the crispy top and I kind of leave the core of the creme brulee.

Jeff Jarvis [02:38:11]:
Sugar.

Paris Martineau [02:38:11]:
I listen. I know, but it's just too soft. There's not enough going on there texturally. Well, listen, my one weird food thing.

Leo Laporte [02:38:20]:
Fascinating. Well, ladies and gentlemen, on that note, on that thrilling. Get your gobbles and go out and have some pudding because we are done. This is all over. Everybody go home. That is Jeff Jarvis, professor emeritus of journalistic innovation at Craig Newmark Graduate School of Journalism at the City University of New York, now Montclair State University and SUNY Stony Bro. He is the author of the goody bird parenthesis, Goody bird, goody birdie bird parenthesis, and of the magazine and the VEP Viv. Thank you, Jeff.

Leo Laporte [02:38:55]:
Good to see you. Paris Martineau, of course, investigative reporter at Consumer Reports, Paris, nyc. Thank you everybody for joining us. A special thanks to our club Twit members. We will be back next week. And again, again we're going to talk about emotionally intelligent AI.

Paris Martineau [02:39:13]:
I'm not a human being, not into this animal scene. I'm an intelligent machine.