D (2:27)

Thank you.

A (2:27)

You tell a story early on in the book about your daughter Kira, when she was born, had a medical condition and you were. This is 2006 or even earlier. Yep, pre Wikipedia. And you were searching the Internet for information about the treatment that the hospital was offering, and it was hard to find any information. You point out now, the condition, the treatment, all the information you ever want is available immediately to anybody in the world through Wikipedia. And that's an amazing that's life changing in that little sphere. And so you've changed so many lives. And it's not just you, obviously. In fact, Molly White's gonna be on Twitter on Sunday and she's one of your fabulous editors.

D (3:17)

Yeah, she's fantastic.

A (3:18)

How many contributors are there on Wikipedia, do you know?

D (3:21)

Well, I mean, it depends on how you count. There have been millions over the years making one edit or whatever, but I would say a typical number we go by is 70 to 80,000 per year.

A (3:35)

Wow.

D (3:36)

I mean, sorry, 70, 80,000 people per month make at least five edits one month.

A (3:42)

So that's pretty dedicated. It's not just some hit and run editor, that's somebody who's spending some time.

D (3:47)

Yeah.

A (3:47)

There are other stats from the book. Seven million articles on the English version alone. And of course there are Wikipedias in 300 other languages. That's 93 times bigger than your predecessor, the Encyclopedia of Britannica.

D (4:03)

Amazing.

C (4:04)

Yeah.

A (4:05)

Did you put Britannica out of business, do you think? I mean, they're still around, but.

C (4:09)

Well, they're around.

B (4:10)

Well, they just published another print edition. They did, they just published another book? I believe so. Or was that Miriam?

C (4:16)

Not the full series, no.

A (4:18)

I have the World Book from when my daughter was born, 1992, on my shelf, but pretty much print encyclopedias are a thing of the past, thanks to Wikipedia.

D (4:30)

Yeah, I mean, you know, even, even like before us, like Encarta really put a massive dent in their sales because an encyclopedia on a DVD pre Internet was like a pretty cool thing actually. One of the former editors of Britannica said to me, we always understood we were selling both a knowledge product and furniture. Meaning you got these leather books to show that you're proper middle class people and you care about education for your kids. And they're a little dusty, but you know, there's something to have in your house.

A (5:06)

It's definitely performative education. And frankly, I'll admit it, that's why I have the leather bound World Book on my shelves in my, my wife.

C (5:14)

Forces me to go to drop books off at the used book sale and drags me out to do it. They never, they won't take encyclopedias anymore.

D (5:22)

Oh wow.

B (5:23)

I'll correct the record. It was the Collegiate Dictionary of Merriam Webster that was published yesterday.

C (5:28)

Not encyclopedia, sorry, last edition of Encyclopedia Britannica was 2010 15.

D (5:34)

So you know my copy. I have an old picture of me sitting next to the World Book Encyclopedia when I was maybe five, four or five years old. And a few years ago I Sold one of my old computers through Christie's, through, you know, through an auction. And that was great. And I was like, you know what we really need to sell or donate to the Smithsonian? Something is my old world book encyclopedias. And my mother said, yeah, I think they're in the shed. And my father said, no, we got rid of them years ago. So not long after that, my parents were moving, I went to Alabama. I went through every single. They had the most amazing amount of crap in their shed. And I'm like, yeah, mom, I'm gonna, I'm gonna, you know, like, it'll be your money. It's your encyclopedias.

A (6:28)

Oh, we were warned that Jimmy's inner.

D (6:32)

Fond memories of those encyclopedias.

A (6:35)

There you go. Yeah, as do I. I grew up, you know, as a kid, kind of an autodidact, you know, I'd take volume S and bring it to the bathroom with me. And I think a lot of the weird facts I have in my head are because I read the encyclopedia, but now it's just, it's all available.

C (6:58)

Jimmy, how did you first come across wikis as a protocol?

D (7:03)

Yeah, so we were working on Nupedia, the predecessor to Wikipedia, for about two years. It was going very slowly. We had a very top down system. There was a seven stage review process to get anything published. It was very academic. We had a fair number of people gathered, you know, on mailing lists and so forth who were excited about the vision, but they couldn't get anything done. And so one of my employees, Jeremy Rosenfeld, came to me because he had heard me complaining. I'm like, this is too hard. I tried to write an article. I was very intimidated by the process. I'm like, this is too hard. And he came to me and said, have you ever seen a wiki? I'm like, no, tell me about a wiki. And so actually a fun thing. So Ward Cunningham, who invented the wiki, lovely guy, programmer in Portland. I contacted him and I said, oh, what do you think about using a wiki to make an encyclopedia? And he said, yeah, you could do that, but it would still be a wiki. He was right.

A (8:01)

Wikipedia, to me, they were in, they were invoked.

C (8:04)

You remember Jimmy, the story of the which. Which you became central to the LA Times. Wikitorial.

D (8:10)

Oh yeah, yeah, yeah, yeah.

C (8:12)

There was an editorial. They decided, I think it was about Iraq or something. They were going to, they were going to use the wiki. We're going to have one editorial, but we're going to find consensus and everybody's going to collaborate on this and then they're all just going to magically agree. And I remember I blogged it and you came in at the moment and you said, that's the wrong way to do it. You got to fork it.

D (8:31)

Yeah, yeah, yeah.

C (8:32)

And you went to the LA Times and you leaked.

D (8:36)

A lot of Wikipedians sort of jumped on to try to help, but they had even hidden things in the software which are absolutely crucial to the process, like the history, like to be able to go back and look at the history and see all the versions.

A (8:48)

It's vital.

C (8:49)

That's for you.

D (8:51)

I mean, we knew how to get there because they were using our software. They had just hidden the link. So a lot of Wikipedias were like, okay, well, you just have to type in the link for the history so you can see what's going on.

A (9:00)

It's not the same.

D (9:01)

Anyway, it was a bit of a disaster for them.

A (9:04)

That's one of the keys to trust. We're going to talk about your book, as a matter of fact, just a little bit. I will say a few things, though. First of all, I remember interviewing Ward Cunningham because I was a wiki aficionado. In fact, I ran a website that was a wiki for a long time. Our Twit wiki was based on Wikimedia's software. We used your software. Great software. I love wikis. I still love wikis. I run a tiny wiki on my own home server because I just think wikis are a natural way to store information. I just love them. Nupedia, which required you write a cv, you had to submit a cv, you had to be an expert on the subject, as a result, had a mere 21 articles in it.

D (9:51)

Yeah.

A (9:51)

But there must have been, at some point, a transformational leap of faith when you said, okay, you don't have to be an expert to write an article. Yeah, something flipped for you.

D (10:04)

Yeah, well, I mean, we were already having discussions in the community about what are the qualifications. And, you know, I was really in the camp of like, well, does it really matter? Like, it's really the work that matters, not. Not the qualifications. And then basically it was when my. When my daughter was born and this fiasco, this was sort of. I was on the edge of pulling the plug because Nupedia wasn't working. And I was like, you know what? Just screw it, we've got to try it. And so put up the wiki. So, you know, she was born December 26th, and I launched Wikipedia as soon as I got home from the Hospital. She was in the hospital for a couple weeks on January 15th. And I was just like, you know what, we just gotta try. We got one last try, put it up. And the community immediately came in and started editing stuff. And it was actually really fun. Like you can imagine being the first person to write Paris is a city in France and hit save and like, look at that. Look at my amazing work. It's fantastic.

A (11:08)

You know, I guess you don't have to be an expert to write that article. At least start.

D (11:12)

Yeah. And of course, I mean, we were just getting started. It was all greenfield, there was a lot of fun stuff to do and obviously we had to sort of think about things. I'll tell you something really crazy about the early software I had downloaded. UsedModWiki, which was a Perl script, which was very easy to install. Open source was great. You know, I just got started. But this software, you could create a user account, but there was no concept of passwords, so anybody could come in the next day and log into your account. Like there was no real user account. So you could, you could give yourself a name, but anybody else could pretend to be you. So this was clearly not going to build trust. This was clearly going to be a disaster. So one of the very first things I did was scrabble together a little bit of a login system. But yeah, it was really very experienced, experimental and very open ended and quite fun really.

C (12:04)

How technical were you then?

D (12:06)

Pretty technical. I, I'm a programmer. I'm not a very good programmer, but I, yeah, even now I program for fun.

C (12:15)

Vibe or Vibe or hard?

D (12:18)

I mean a little of both. A little like I, I can program, but I, I need some help. And actually, I mean just even like the, the borderline is like just the autocomplete of something. You start writing has gotten to be pretty good and then you're like, yeah, that's what I was going to type anyway. But yeah, no, I actually think it's really amazing. I do think people who think they're purely going to vibe code and get something ready for production, you're probably kidding yourself because it's not that good yet. But if you're a programmer, like a sorry, rusty old guy programmer like me, you can actually get some stuff done that you're like, yeah, actually this would have taken me 10 times as long because I've forgotten a little bit of syntax and I, you know, I'm not up to date on the latest things, but it's like, all right, yeah, okay, this is fine. And I can read code and it's all good. But yeah, I think it's an amazing time to be making things.

A (13:12)

Talk about crusty. Last weekend I spent morning using Claude code to rewrite my emacs configuration. And it did, by the way, a lovely job and it really improved it significantly. And I was blown away, frankly, that it could write elisp and do it so well. It's amazing.

D (13:33)

Amazing.

A (13:33)

So your new book is the Seven Rules of Trust. And what I found interesting was, you know, I would have assumed off the top of my head that the trust involved was you trusting, or rather people trusting Wikipedia. And in fact, before that happened, you had to develop trust with the editors, with the community. And that's really what this book is about is how do you. You say we're in a crisis, a trust crisis globally, right now, we need trust somehow to come back. How do you, how did you create trust in that early Wikipedia community?

D (14:12)

Yeah, you know, a lot of it is just, you know, like getting to know people and, you know, just sort of realizing, because we were a small group, you know, it's like, okay, well, these are the people I've been talking to for a couple years thinking about the encyclopedia. Obviously they can just start editing and we'll just see what happens. So it was low risk, was a big piece of it. But also it was a learning process over time. I mean, I remember I used to wake up in the middle of the night in a little bit of a panic and run downstairs and just check on the site because I was afraid in the middle of the night somebody would trash the whole thing. Which you could have in the early days. But then I pretty quickly, I was like, oh, like, there's this guy in Australia, and when I'm asleep, he's on and he's like, making sure there's no vandalism and so forth. And then, you know, over time, there were a lot of details of like, okay, how do we. How do we get there? How do we. What are the things we need to do to. To build trust in the community? So that's, you know, there, There were so many steps.

A (15:15)

You talk about Frances Fry's trust triangle in the book. I apologize, by the way, for the watermark on the triangle, because I have a pre distribution PDF of this, the book, but three, three key elements of trust, authenticity, empathy, and logic. And she says every time trust exists, those three things are there. Every time it's broken, you could trace it back to a failure of one of those three.

D (15:46)

Yeah, yeah, no, I, I think It's a great framework because before I came across that and started to, to do my research, I just knew, sort of like my experience, like, okay, you know, like, what do we do? What are the things that we've done to build trust in the community, build trust with each other, and then ultimately trust with the public. But her framework, I think is really good and it's not unique to her though I think those labels are hers. But it's actually quite common in the academic literature around trust to have a sort of a three part formula. Sometimes they're named a little bit differently and they, they make sense. You know, if you, if you think about empathy, for example, like the, the idea of you need to genuinely be able to put yourself in the other person's shoes and understand them, and then even if you disagree with them, you can begin to trust them, but also you can behave in a way that they find trustworthy. Whereas if you're not having any empathy whatsoever, then the, the interaction is cold and not helpful. And I, you know, we see this a lot in politics. Like right now, everything is so divided and so divisive that, that people don't even have a kind of like, really simple empathy of like on certain issues. It's like, well, if you disagree with me, you must be an absolute Nazi horrible person. Right?

C (17:04)

Right.

D (17:04)

It's like, well, you know what? Probably not. They might be mistaken, they might be an absolutely horrible Nazi person. But actually among like normal people, that's extremely rare. Like, people can be bigoted, they can be friendly, they can be both of those things at the same time. But generally they're not actually horrific people. And if you can't sort of see them as people, then it becomes really hard to understand where they're coming from and what they relate to. I live in the UK and this was so obvious during Brexit, during the Brexit vote, like all the major political parties were opposed to Brexit. I live in London. I didn't meet a single person who thought Brexit was a good idea. And actually a lot of the rhetoric was basically saying, well, people are racist and that's why they're in favor of Brexit. And it's like, well, in retrospect, some of them are racist, of course. Some of them are actually just pissed off about a lot of legitimate issues that nobody in London was listening to. And, and that lack of empathy, that lack of saying, oh, like why are people so upset about Europe of all things. And it's like, and, and the answer is well, it's about jobs. And if you're like, well, hold on, look at the statistics, that's not the problem. You're not listening, right? You're not, you're not empathizing with what they're upset about. That may be legitimate and maybe something that needs to be dealt with. So anyway, I, yeah, I could go on all night.

B (18:26)

I feel, I feel like this gets to a point you were making earlier, which is when Wikipedia was created almost 25 years ago now, the Internet was a very different place. I mean, some might say it was, I guess, more human. It was certainly a lot smaller. And so it was easier to interpret other users actions in good faith and kind of build out a trusted network like this. That's obviously changed a lot as the scale has grown over the past two and a half decades. I mean, do you think that the sort of experiment in trust that was and still is Wikipedia could have existed if it was founded today?

D (19:07)

So, I mean, I'm going to say yes and no to what you just said. So before Tim Berners Lee and the World Wide Web, there was this amazing place called Usenet, which was a giant sprawling message board, sort of like Reddit today, except it was more or less unmoderateable because of the way it was distributed and it was unbelievably toxic.

A (19:32)

I remember the flame wars on unet. Oh my God.

D (19:34)

It turns out people don't need algorithms to persuade them to be jerks to each other. And so, you know, that, that we should be careful. That's all I'm saying, we should be careful to not over idealize and say, well, it was this amazing, beautiful place and all these geeks came together and they shared their software and they're. Yeah, it was that and it is that today. And it's also like people are, you know, can be quite difficult and a lot of the problems that we see today in social media. So like one, one example, you know, that I, I like to give is like one of the things that really works in Wikipedia is the consistency of identity. Not necessarily your real identity, but you've got an account and all of your edit history is there. You build up a reputation for being a decent person and, and sort of being reasonable or unreasonable, whatever, whatever that reputation may be. And you can contrast that with, you know, a place like Twitter where frankly, if you post almost anything on Twitter, I keep calling it Twitter, but I.

C (20:36)

Refuse, I will not call it.

B (20:38)

Yeah, X the Everything app.

D (20:41)

X the Everything app. Yeah, like you posted. What often happens is some Absolute random, sort of sends you an angry message. You've never seen them before, you'll never see them again. And that kind of atomistic interaction, which isn't about building a reputation for thoughtfulness and kindness over time, doesn't. It's not conducive to people, you know, behaving a little bit better, whereas just having a. You know, over time, people get to know you and they're like, oh, yeah, yeah. You know, like, yeah, you know, Bob's a little twitchy about this topic, but other than that, he's a decent person. And, yeah, great. And we just have to be gentle around that topic.

A (21:23)

That's rule number five in your book. Your mother was right. Be nice, be civil, be respectful, be nice.

D (21:31)

Yeah, it works. It actually works. And I mean, a lot of this, you know, one of the questions people ask, oh, well, this seems very abstract and it all sounds very nice. I'm like, you know what? But it's also very practical. Like building. Building trust is very practical. You know, in business, we have some great business examples of, you know, companies. Airbnb had an enormous crisis of trust very early on that could have killed, like, if you. If people generally believe, if I put my house on Airbnb, it's going to get trashed and the company's going to do nothing about it, they're not going to do it.

A (22:06)

Right.

D (22:06)

And so they had to think about, okay, how do we make sure people don't have that feeling? Although, what are the steps we need to take? You know, things like indemnity, like, if somebody trashes your house, we've got insurance, it's going to be okay. You know, like, that was a huge step, and they didn't have that at first.

C (22:23)

What struck me, I came to a cold stop. Jimmy, in the list at number three, which I think is so important, a strong, clear, positive purpose is essential. And your example, of course, is that Wikipedia's encyclopedia, and I would argue that Google has had that to organize and make accessible the world's knowledge. But I've long complained that Facebook never had that North Star. They never said, why are we here? What are we doing here? Neither did Twitter. And I think that that's perhaps a glib explanation to the downfall, but I don't think so. I think nobody could say, well, no, you shouldn't be doing that here, because this is what we do here. This is what we don't do here, which everybody does on Wikipedia. So then you bring that to AI. And I think you're right. And I'm trying to think about it. I can't think of an AI company that has that kind of beyond.

B (23:09)

We're going to clarity of purpose, destroy.

C (23:11)

Mankind or we're going to build something smarter than all of you and replace all of you. Or that bs. I don't think anything has a human centered value proposition of purpose in any of the AI companies. Can you name any of them that do?

D (23:25)

I mean, it's a good question. I mean, I would say to some extent there's some high minded stuff, but yeah. Is it really? It's hard to say. You know, like Twitter for example, has the, you know, sometimes people talk about, oh, it's the town square. I'm like, it's a pretty bad town square, like it's ready to move. Because this place, there's muggers around every corner and people screaming at my grandma, like it's completely unpleasant. And it's like that idea of like a town square. It's like, oh, this is where people go. And there's debates and we're. And it's like an open purpose of ideas.

C (24:02)

Like, what are we, what are we building together? Yeah, what are we making together? Why are we here together? Right.

A (24:08)

Everything is going to lend itself to that. But that's true beautifully. Wikipedia clearly does. And it's, it's kind of a testament to what you're doing that people can arrange themselves around it and get behind it. You don't have to recruit. People get it and they go, yeah, I want to help.

D (24:27)

Yeah, yeah, exactly. And actually it's one of the things that I say kind of in sympathy to social media companies who I am critical of in many ways, is to say, like, actually that's a really hard problem.

B (24:38)

Yeah.

D (24:39)

If you've got a box that just says what's on your mind? As it turns out, some people have pretty horrible things on their mind. Right. And, and like, how do you make those judgment calls? And you know, like, I think they could do a much better job of it, but it is hard problem. Whereas with Wikipedia, there's no box that says, you know, hey, how you feeling? What's on your mind today? It's like, no, we're here to build an encyclopedia. Even the talk page, you know, like it's very common. People come on the talk page, they're new to Wikipedia, and they just start ranting about the topic and it's like, well, okay, right. I also don't think very highly of Donald Trump, but that's not what we're here to do. You can cut the Whole Internet for you, ranting about Donald Trump. How do we improve the article? That's the real question. What can we do about the article? So that purpose makes our self management so much easier. And of course thinking about social media, which is quite hard, is one thing, but I also think it applies to companies having a simple, clear purpose. You know, what you're there for is, you know, it's classic and obviously companies can do more than one things often, but sometimes doing even one thing is hard enough, much less doing too many different things. And a lot of big conglomerates lose focus and they end up falling apart because they're not really doing, you know, I, I think, I'm not a super expert, but I'm, I'm gonna argue that Boeing, they have a pretty simple purpose, right? Trust problem, stay focused on that. So that's not their problem. They've got a purpose, it's a good purpose. Like build a plane that doesn't fall out of the sky. Like that's super important. So it's not the only thing. Right, but it is one of the things that, you know, I think is, is key and I do think a lot of companies in all kinds of industries, you know, run on, they, they run afoul because they, they end up doing too many different things. They're not really sure what they're doing. I think Yahoo is a great example of a company that, you know, was a little bit too much of everything and not clear. And in our case because we're just in encyclopedia and because we're a non profit, you know, I remember very early somebody said, oh, when Gmail came out, somebody said oh, should we start offering like webmail accounts? Wouldn't that be cool? And I was like, there's going to be a lot of management problems and technical issues and like I don't, yeah, it just, it doesn't seem like that's building an encyclopedia. So let's just stick to the one thing that we know how to do. And we have obviously side projects and things, but they're all in that same theme of sharing knowledge in a wiki way.

A (27:08)

Yeah, you famously turned down advertising. People have pointed out how much money you could make. Is that part of your purpose is not to make money? Wait a minute, that's not right.

D (27:20)

Well, I mean it turns out, yeah, it's a pretty terrible business. We're a charity.

A (27:26)

So you know, and by the way, I contribute, I'm a monthly donor and I'm very, I, yeah, right. Because I get such value out of Wikipedia. I Use it every single day.

C (27:35)

I remember Jason Calacanis, I think, in Munich, yelling at you, giving up all this money. This is just stupid. You must be doing this.

D (27:47)

I mean, I think there's a lot of things, like a lot of reasons. One, I mean, we're a charity and whatever, but I think my biggest reason is sort of aesthetic and artistic, I. E. I just don't like the idea of it. I think Wikipedia is very nice. It's like a temple for the mind and I enjoy it. But I also also think like, organizations ultimately do follow the money and like, nobody at the Wikimedia foundation is thinking, oh, how do we, how do we improve engagement to keep people on the site longer so they see more ads? Or gosh, all these people reading about World War II, there's literally nothing to sell to them. Can we move them on to topics that will be better revenue? And they just don't do that. I mean, nobody ever thinks about that. Like, we don't care what you're reading. It's completely fine because it doesn't matter. Actually, all we care about our business model, which is the funding for Wikipedia is vast majority is small donors. So we're not funded by governments. There's nothing. I was told just over $10. So what do we need? We need people to go when they see that little message. They want to. Views. It's just like, it's worth it to me to like help. I should chip in. Like, it's a good thing.

A (29:10)

Yeah, it's, it's unnatural to, to. To support Wikipedia.

C (29:15)

What's the average donation do you know?

D (29:17)

Yeah, just over $10.

A (29:19)

Wow.

C (29:20)

You, you, you froze for a second. So we missed that.

D (29:23)

Ah, yeah, yeah, yeah. Well, you, you all froze as well. So.

A (29:26)

Yeah, we all fro from.

D (29:30)

Yeah, yeah, yeah.

A (29:31)

Well, let's talk about AI because if we're. Go ahead. I'm sorry, finish your thought.

D (29:34)

Just, just think of the amazing things we're going to do when the Internet doesn't suck.

A (29:40)

You're one of the, one of the signal reasons Internet does not suck. I might add when you talk about trust crisis. We are about to enter, you know, a trust apocalypse thanks to AI Now I'm an AI fan. We do a show. This show's about essentially bad AI. I would, I would say that most of the content of Wikipedia has been sucked up by every AI as part of its training material that regurgitates, you know, the contributions people have given. And yet AI has a big trust issue because of hallucination, because of AI Slop. You've. You actually ran into a little bit of a buzz saw this summer when you thought, when you mentioned briefly maybe we could use AI in Wikipedia. I don't know, you thought maybe judge articles or. We're looking at different ways you could use AI.

D (30:33)

Yeah, I basically wrote a little script that you could feed in a short, like a stub, a short entry. Then let's say it has five sources and you feed in all the sources and you just ask, is there anything in the sources that isn't in Wikipedia but should be? And is there anything in Wikipedia that's not supported by the sources? And I've just played around with it. It's okay. It's not perfect. And I like, even just a little prompt engineering would probably help it quite a lot. And you could use a second AI to judge it and you could just tell it, just say no, just like, this is my big beef about AI, is it very often should say, don't know, can't think of anything, doesn't like to.

A (31:12)

That's the problem.

D (31:13)

Always wants to help you. But anyway, it wasn't a buzz saw. Like there was news sort of interpreting that way. But you can interpret a wiki conversation any way you like.

B (31:24)

Y.

D (31:24)

Because there's always, you know, like different opinions.

A (31:26)

Oh, the controversy. Oh my God.

D (31:29)

Yeah, yeah, exactly. And it's sort of like. Yeah. And I think within our community there is obviously like very heavy skepticism about using AI for content for readers because the hallucination problem is. Is really quite bad. And I, in fact, I think most non wikipedians or most people who, who haven't really worked in the knowledge space and all that, that don't really realize how bad the hallucination problem is. And partly because if you use AI in a casual capacity, one of the problems with hallucinations is they tend to be quite plausible.

A (32:04)

Right.

D (32:04)

Because that's the way it works. And so you can very easily go, oh, right. It was an amazing. It gave me an amazing answer, explained everything to me. And you don't really realize that 10% of it was absolutely made up out of thin air because it sounded like it's possible. I. One of my favorite things, I always ask every new AI model who is Kate Garvey? My wife, because she's not a famous person, but she's known a bit. She worked for Tony Blair for 10 years. So she's in different books about the Blair administration. She promotes the global goals, sustainable Development goals with the UN and all this. So, you know, there's a few news stories about her. But it's kind of just perfect that the AI doesn't know who she is. Exactly. But it can guess and it guesses usually.

B (32:49)

What are some of the funniest responses you've gotten from that query?

D (32:53)

Well, so one of the best was it said that she set up a non profit. Okay, that's true. Set up a non profit to promote women's empowerment in the workplace. Sounds like Kate. She did it with Miriam Gonzalez, who is Nick Clegg's wife. Nick Clegg was the Deputy Prime Minister of the UK who went on to work for Meta as, as head of Global policy. And we know Nick, I know Nick from work. And actually our kids go to the same school because London's a small town, so we know them socially. And my wife said, wow, when I showed her this, she's like, like, that could have happened. Like we could have easily done at a party.

C (33:30)

It's my digital twin.

D (33:31)

Like it's completely plausible but completely false. And then the other, the other one that's really fun is I, I then I ask, and who did she marry? Never me. It's generally somebody we know because I, I always joke, I married into the Labor Party. So we know a lot of political journalists and politicians and things like that. And that's sort of our social circle here. And so it'll be people. We know one once. So my UK publisher, Alexis, is married to James Purnell, a former labor minister who was also my wife's roommate back in the day. And so it said to me, yeah, that she married James Purnell. And I actually sprung this on Alexis. We were doing an event at Bluesbury for her sales team or whatever for the book. And I was like, well, here's an example of a hallucination. And I'm like, I put it up on screen. It's like, yes, she married. I said, oh, wow, that must have, you know, two prominent people. That must have been quite a. Oh, yes, yes. You know, Tony Blair was there and you know, we're very happy to give a lot of nonsense. And then I say, you know, but I thought she married Jimmy Wales from Wikipedia. And it's like, oh, you're right. You're completely right. I'm so sorry, I got it all wrong.

C (34:42)

Where's my sword to fall?

B (34:45)

How dare I?

D (34:46)

And then the other one, and this is like, British audiences really get this quite quickly. I'll just say it anyway, but. Because he's been in the news lately. But the other one is, that is Peter Mandelson. Who just was the us, the ambassador to the US from the uk, who just lost his job because he had associated with Epstein back in the day and all that. But he's very famous in the UK, big part of the Blair administration. I said to ChatGPT when it said Peter Mandelson, I said oh, but isn't, isn't Peter Mandelson quite famously gay? And then it got very woke with me and it said it's not appropriate to speculate about people.

A (35:21)

Shame on you for asking. That is very funny. We're talking to Jimmy Wales, the founder of Wikipedia, a man who's done, and probably inadvertently, something so massive to change the world that your name will forever be enshrined on Wikipedia. At least one would hope, one would expect. So do you anticipate AI ever making its appearance on the Wikipedia's pages?

D (35:53)

Well, I mean, I think we're already seeing it for better and worse. So one of the good uses, if you are trying to write in a language that is not your, your mother tongue, people often feel a little bit shy about it because Wikipedia is written, you know, kind of a high level and so forth and people are using it to, to write something and then ask can you, can you help me with the, the language here? And it's pretty good at that. And so I know that's going on translation, machine translation, not necessarily using ChatGPT, but machine translation's getting to be pretty good. So that's another bit of AI being used. And I do think, you know, thinking about tools for our community, tools to help us find articles that are wrong. I, I talked to a, A one of our French volunteers who's an old school Wikipedian and she said I don't really have time to edit Wikipedia like I used to. So my hobby now is I find a dead link in French Wikipedia, I see what it was supporting and I go find a new reference that supports the same thing. Great. I said like what if an AI could take the dead link? You don't need AI to find a dead link. 404. And I found what if it could find dead links and go out and look for sources and just suggest to you, oh here's what this was supporting and that link has died. Here's a quote from this. Does that suit? Would you? And she's like, oh that'd be amazing. Because the human part, the interesting part is her judgment as an editor. Does this article support this statement? Do I need to modify this statement to be more accurate?

B (37:23)

It.

D (37:24)

But the part that a machine can do is actually understand it just enough to go and find a source that probably works. And so I think little tools like that will be the first use and I think that that will be great. Like, you know, it's like typical of any technology. Like the best use of any technology is the boring bits that don't really require a human. Take that over and let us do the stuff that humans are good at.

A (37:47)

Perfect.

D (37:47)

I think the fun thing about AI is like what it's good at is a set of things that aren't quite what we all would have expected. I think if we had had this conversation 30 years ago and said, oh, what will the first AI be like? And I think we would say it's going to be ruthlessly boring and factual, not creative at all. It's going to be very mechanical like a robot. It's just going to regurgitate facts. Turns out it's really bad at regurgitating facts and it's quite good at brainstorming and coming up with wild ideas. So, like, not what we expected.

A (38:20)

What do you think of Grokopedia?

D (38:24)

I don't really think of Grokopedia. So, no, I, I haven't had time. I've been on my book tour, been promoting the book, trying to get people to remember I'm not a billionaire and need to, you know, sell some books. And so I.

A (38:36)

But that's your own darn fault, by the way, Jimmy, you could have been. You could have been.

C (38:40)

Jason Tried to tell you, but no, you wouldn't listen. Very happy.

A (38:45)

I love it because you're right, it's perverse incentives and you've kept pure. Yeah.

D (38:50)

Kept it simple. Yeah. So I haven't had time to dig through it personally, as much as I would like to. I have seen news headlines of, you know, people are finding a. What I would expect, which is hallucinations like AI is not good enough to do this and it's going to make stuff up. Also, apparently, despite Elon claiming it to be more neutral than Wikipedia, it seems surprisingly aligned with some of his more intriguing political views. Views. And so I can tell you not everything in Wikipedia agrees with my political views, even though my political views are quite boring. But yeah, so I, you know, I'm skeptical of the project and I actually think it's interesting to sort of see a large scale attempt to do this. But I don't think the technology is going to be there for several years to come and particularly not if he's got his thumb on the scale or his staff do because they're terrified if it says something. You know, like one example, like this is something that Elon sort of complained to me on Twitter about and so on, and said Wikipedia's mainstream propaganda or whatever. But, you know, he made that gesture at the. At the Republican convention or. No, the inauguration, whatever. It was.

A (40:01)

Yes.

D (40:02)

And it was what? So what Wikipedia says is like just very simple and very factual. It just says he made this gesture. Many people said it looked like a Nazi salute. Elon denied it. Like, those are all uncontroversial facts. Right? That's not even, you know. And if Wikipedia said Elon made a Nazi gesture, I think that would be wrong. Like, that's not neutral, that's actually interpreting in a contentious way, et cetera, et cetera. You know, it's fine if that's an opinion. It's not for Wikipedia, but I'm not going to say leaving that out would make sense. It's part of history. It was a very interesting thing that happened and it had impact on public whatever, and it's part of history. He may not like it, but anyway, as I understand it, that's not in Grokopedia.

C (40:46)

Last week on the show, I quoted from a paper from Technical University of Dublin. Have you seen this comparing the two?

D (40:53)

No, I haven't seen it. No, no, I need to.

C (40:55)

I'll send it to you. But it says, I'll quote this again. Grokopedia articles are on average, several times longer. That's true with higher flesh concave grade levels, but lower lexical diversity and reference density. This pattern suggests that Grokopedia's generation processes elaborate on existing material, expanding text length and rhetorical flow, rather than producing substantially new or more rigorously sourced knowledge.

D (41:21)

Oh, interesting. So it's as if I'm just imagining Elon on a phone call with Trump. This is me hallucinating for fun. And Trump. And Trump says, elon, I give you some advice. Tell it to use bigly words, lots of hard words, that'll show that it's really smart.

A (41:39)

Smart.

D (41:41)

A high sort of score on high level. That doesn't have to be true. It just needs to sound impressive, sound bigly.

A (41:50)

Well, the book is fabulous. I really enjoy it. The Seven Rules of Trust. A blueprint for building things that last. Jimmy Wales has the credentials, the bona fides, to write such a book.

D (42:02)

That's very kind of you to say.

A (42:04)

You've done it, you've succeeded. Do you imagine you're going to continue doing this for a while? You want to retire?

D (42:14)

Oh, I Mean, I, I don't know how to retire because I don't actually have a job. I don't work at the Wikimedia Foundation. I haven't ever. I've never had a job working.

A (42:25)

Really? I didn't know that.

D (42:26)

Yeah, yeah. No, I mean, I've always been a volunteer and that's awesome. You know, I even pay for my flights to go to board meetings and stuff like that. It's my charity work. And then I'm Fandom, my for profit company again. I'm just on the board. I don't work. The truth is I would make a terrible employee. So I'm very fortunate to have become successful. Despite that. I just, I like to get up and do whatever I think is the most interesting thing. So lately I have this great project which I love, and I don't have enough time. So when you say retire, I'm like, maybe that's what it means I really want. My house is an old Victorian house and I found out that a. An admiral lived there about 150 years ago. And I have a picture of him because I was able to find his name and I was able to find a picture of him in an old magazine. And I'm making an AI assistant for my home who will be the ghost of administration.

A (43:25)

That sounds fun.

D (43:27)

And you know, when you ask him things about the weather, he'll speak in nautical terminology. If you ask him to turn on the television, he'll be a little befuddled because he doesn't really know what a television is. It's a better idea in my head than anything I've actually made. But I'm having fun playing around. I love it with large language models local on my laptop and sort of home assistant stuff. It's really cool.

C (43:52)

Which models are you favoring?

D (43:55)

So lately I find the OSS GPT 120 is good. It's very wordy, it's fast, and it just writes a lot. So I haven't had time to tell it to. Can we do a prompt to say just really give me shorter answers? Slow down, slow down. It's pretty good. And I mean, the thing is, for this kind of funny idea, the most important thing for a voice assistant is time to first token time to speak because you can't wait 30 seconds for it to answer you. Yeah, I've got a few little cute cheats for that. Like have a bunch of stock phrases that he says while he's actually thinking.

A (44:34)

I'm glad you asked me about that.

C (44:36)

Yeah, exactly.

D (44:38)

But yeah, I think, I think the models. So I, this is actually something I do find very interesting. And I think Nvidia's results came out today and I think, yes.

C (44:46)

So they were soaring. They said yes.

D (44:47)

Yeah. But like I, for the first time in forever, my, my newish computer, I bought the most expensive MacBook that was available. M4, 128 gigabyte, M4 Max, because it can run large language models locally.

A (45:07)

Right.

D (45:07)

And I realized before this I was always like, well, I just, I would get this two generations old. If I got a new computer, I get a refurbished one because I'm like, all I do is get on the Internet. I type emails and I read web pages. Like, I don't. I'm not doing video processing. I'm not doing anything. And now suddenly I'm like, oh, actually I could use a really powerful computer. And I bet you in the next few years that's going to become quite common as not just hobbyists or people in tech like me, but as a lot of people are like, yeah, I thought I would just get the cheapest $300 computer, but it turns out I need a $2,000 computer to do AI. And they will. And so I think there's going to be a real demand for compute, even for home users and people like that. And I might be wrong, but I think the era. There was a short period of time. It used to be like, you really need to get the latest computer because they were all so terrible. Then for a while it was like, well, no, they're all fine. Like, you don't really need a fast computer. They're all completely fine. And now it's like, oh, no, actually a fast computer is meaningful. Again, I guess that's. Gamers have needed, like.

A (46:10)

But they've been the only ones. They've been driving the market.

D (46:12)

Yeah, yeah.

A (46:14)

Now AI drives it.

D (46:15)

Yeah, yeah.

A (46:16)

Bitcoin did for a while for a certain segment of people.

D (46:19)

Yeah, yeah, yeah. But again, I think, I think gamers. That's a large group of people, but it's a segment. And you know, bitcoin's a smaller segment. But I. A computer that can like do the stuff I wanted to do that isn't gaming. That's everybody.

A (46:34)

Yeah. You know, I bought a Framework desktop just for that, just to run GPT120.

D (46:40)

Oh, yeah, great. Yeah. How is it?

C (46:42)

You can buy the Nvidia box.

A (46:43)

Very nice. I like it a lot. The Nvidia box people are complaining about already.

C (46:47)

Really?

A (46:47)

We'll talk about it.

D (46:50)

Mixed reviews. Mixed reviews. It's quite expensive. Yeah. It's sort of twice the price of the framework. Yeah. Which is like 2000 for one of those 395AMD.

A (47:01)

Yeah. Strix Halo.

C (47:03)

Yeah. You're too young to retire. That's true.

D (47:05)

Yeah.

C (47:07)

And probably too young to call a grandfather of the Internet. But as you are a pioneer, who's. Who made the better Internet and changed it? I'm just curious. Who else do you think deserves to be in that club with you? Who else has done well by the Internet?

D (47:23)

I mean, I. I think Tim Berners Lee, you know, like, he's a. He's the real deal. Actually. Craig of Craigslist, I think is great. You know, I think. I think there's, you know, but I also, you know, like, I can be critical of various aspects of big tech, but I mean, I also think Facebook has its problems, but you know what? I keep up with my high school friends on Facebook. It's not all bad. And there's a lot of bad stuff out there. So I'm still a very pro tech, tech optimist, but at the same time acknowledging, yeah, we got to do better. There's a lot of stuff. Yeah, we've got to do better. There's a lot of stuff that could be a lot better.

A (48:06)

Well, I'm in favor of making Jimmy Wales a billion. So everybody go out and buy this book, the Seven Rules of Trust, A blueprint for building things that last. Should be Sir Jimmy Wales. But I guess if you grew up in Alabama, you don't get that, so.

D (48:24)

Well, I am a UK citizen now, so.

A (48:26)

Oh, okay. Maybe you can. All right. I was going to make you a Kentucky colonel, but. Okay, we'll take that. It's really a pleasure to have you on. So great to see you, Jimmy. Thank you.

D (48:37)

Fantastic.

A (48:37)

Thank you so much. Jimmy Wales, everybody. Let's. We all owe him around. And if you're not yet donating to Wikipedia, if you use it, which I'm sure you do.

D (48:47)

Yeah.

A (48:47)

Please become a regular donor. This is something that we all, you know, they don't demand it, but it's something we all should do.

D (48:54)

Well, It's. It's a 25th anniversary coming up, so I'm thinking 25 bucks sounds about right. If I were giving a kid I liked a bit of money, when they turned 25, I'd give them 25 bucks, so that seems bad.

A (49:09)

Yeah. Only 2%, according to Wikipedia, of the readers actually are donors. Let's increase that number. Let's make it 10%. Yeah. And then Jimmy could go first class. Thank you. Thank you, Jimmy. Really appreciate your time.

C (49:24)

Thank you.

A (49:25)

Have a great day.

D (49:26)

Thanks for having me on. Brilliant.

A (49:27)

Take care.

D (49:28)

All right, Bye.

B (49:28)

Bye.

A (49:29)

We'll have more intelligent machines after this word from our sponsors. This episode of Intelligent Machines brought to you by Vention. These are my good friends at Vention. I talked to Glenn at Vention a couple of months ago. Really impressed with what they're doing. First of all, Vention's been around for 20 plus years. Engineers that really know how to make software singh global engineering expertise. But of course, things are changing now in the the world of AI. And I bet you and your company have had this experience. AI, you know, it promises to make things easier, but often for many teams, it makes your job harder. We've seen that time and time again. This is why it's worth taking a look at Vention. Vention helps build AI enabled engineering teams to make software development faster, cleaner and calmer. Okay, they're not, they're not trying to like jangle your nerves here. Clients typically see at least a 15% boost in efficiency. But we're not talking hype, we're talking real engineering discipline. Because first and foremost, Vention is an engineering team. But they also know a lot about AI. In fact, they've started doing these great fun AI workshops. They're not just a lecture session about AI, they're our interactive session that will help you and your team understand what your goals are and how to incorporate AI into it. Using the expertise of the Vention engineers, they'll help your team find practical, safe ways to use AI across delivery and Q and A. It's a great way to start with Vention to test their expertise. Whether you're a cto, a tech lead, or a product owner, these seminars are great. You won't have to spend weeks figuring out tools, architecture or models. Vention that these workshops will help you assess your AI readiness. They will work with you back and forth to clarify your goals and then build the outline of the steps you need to get where you want to go without the headaches. And if at the end of the workshop or later you feel like, oh, you know what, I would like some help on the engineering front. Ventions teams are ready to jump in. They could be your development partner, they can be a consulting partner. It's the most reliable step to take after that proof of concept, you know, we've. You've gotta been there, right? You built a promising prototype. What did you do? You went to lovable, right? And you built something and it turns, turned it looks good, it runs well in tests. But now how do you go to market? What's next? Do you open a dozen AI specific roles just. Just to keep moving? Or maybe I got a better idea. Bring in a partner who has done this across industries. Someone who can expand your idea into a full scale product without disrupting your systems or slowing your team or shipping something that's just slop. They can help you. They understand how to do this. Vention. It's real people with real expertise and real results. Learn more@ventionteams.com see how your team can build smarter, faster and with a lot more peace of mind. Or get started with your AI workshop today. Ventionteams.com TWIT that's V E N-T-I-O-N teams.com TWIT we thank Vention for the work they're doing and making this stuff accessible to businesses that want to use it and safe and calm. And thank you for supporting intelligent machines. Machines. Wow, this is a big week for AI releases. Gemini 3 came out yesterday. ChatGPT 5.1. Google also announced a coding IDE called Anti Gravity and Grok went to version 4.1. I wasn't paying any attention to that one. No. Have either of you used Gemini 3 yet?

B (53:22)

Yeah, no, I haven't. Jeff, what's your experience been so far?

C (53:26)

Well, my experience was I couldn't find it, so of course I started off cursing, thinking I couldn't get it and.

B (53:30)

Others could because it was a workspace curse.

C (53:32)

No, no, I thought that's what it was. It's hidden. If you go to Gemini, whatever. I asked it, I said, what version is this? And it's 2.5. And what the heck, you've got to go down. And the fast is 2.5, but the thinking is 3. So go to the thinking is 3. Now, my phone version of, last I checked an hour ago, still didn't have it, but my web version did. Yeah, and I only asked a couple questions to start off with. And it, yeah, it's impressive so far, but I think I haven't yet. To really go into the muscles, you've got to use the full multimodal mode and some of the other things it can do, which I haven't had time to play with yet.

B (54:10)

What is supposed to be the difference about this model?

A (54:13)

Everybody's raving about it. I mean, the.

B (54:15)

In what way?

A (54:16)

In every respect.

C (54:17)

So feed it multimodal in one query.

A (54:20)

Yeah.

C (54:21)

So it's not separate. You can, you can mix the modes Mix the media interest to it. That's one.

A (54:27)

It's got a very good. It's very good at code. It's very good at deep thinking. It's actually quite a good writer, I think. Of course, my taste in writing is apparently not very good because I liked that nuzzy nuzzle.

B (54:46)

Which one?

A (54:47)

I liked Liz's post. I like the bamboo analogy.

B (54:51)

You like the bamboo.

A (54:52)

I like the bamboo analogy.

B (54:54)

This can of worms terms here now.

A (54:56)

That you and Jeff have both.

C (54:58)

I had, I had a meeting earlier today who insisted. Who thought that the bamboo was him making fun of Olivia.

A (55:05)

Well, yes.

B (55:08)

I feel like that's too convenient anyway. A lot of words about the. On the bamboo. If you don't know what we're talking about.

A (55:17)

I guess this is more interesting than Gemini 3.0. This is, it's a. This is an inside, Inside edition.

B (55:25)

How much time I spent on a first date last night talking about this?

A (55:29)

Did you really. Well, everybody.

C (55:32)

Were you out with a media person who understood this or out with some.

B (55:35)

I was out with a media person and we went, we went a good hour and a half, two hours into the date before we brought Nazi gate. And I think that's really impressive.

A (55:44)

I think though, it's a great conversation for a date. You can really suss somebody out.

B (55:48)

Will someone explain it succinctly? Because I won't be able to explain it without giving every.

A (55:52)

So Olivia Ozzy is a reporter. She was working for the New York Times.

B (55:58)

No, she was working for New York magazine. Sorry, Olivia.

A (56:00)

Vanity Fair. Right.

C (56:02)

Yes, and now Vanity Fair.

A (56:04)

She was doing. She was doing reporting, political reportage and was at one point associated, we're not sure how deeply with RFK Jr.

B (56:17)

So I'm gonna. Was a. Is. Has been a prominent political journalist largely known for her incisive, well written magazine style pieces for the Daily Beast and then as the. As New York magazine's Washington correspondent over the course of the Trump administration and subsequent Biden administration, she really kind of carved in a niche for like having these like fantastically well and reported in depth looks into Trump world and kind of political intrigue. She had a longtime relationship with another major New York or D.C. based political reporter, Ryan Lizza of Politico. This all came crashing down.

A (56:59)

Keith Olbermann, apparently that was, that was.

B (57:02)

Before but I was getting ahead of.

A (57:05)

Her, shocked to learn that.

C (57:06)

Although that's. We.

B (57:08)

Yeah.

A (57:08)

Let me tell you what, let me.

B (57:10)

Just tell you what the important context is that in last September it came out that she had been put on leave and eventually would lose her job at New York magazine because she was carrying on a. What she then described as a digital affair with RFK Jr so to bring.

A (57:29)

Us back to our main subject, I have asked chat GPT. I mean, Gemini 3. The prompt was what's Lucy? What's the tea on Olivia Nuzzy? Based on the latest reports as of November, the tea on Olivia Nuzzy is a multi layered media scandal involving alleged affairs with high profile politicians. She was reporting on a messy public fallout with her ex fiance. That's Lizza. And her subsequent career moves.

C (57:55)

That's Vanity Fair.

A (57:56)

Pretty good. Yes, pretty good, right. So the breakdown is the RFK Jr. Scandal, which she was revealed Harris or Gemini. Gemini is doing a great job on an emotional and digital affair with RFK Jr. I got to bring this back to AI somehow. It talks about her memoir, which is coming out next month, American Canto. It talks about, okay, the new accusation.

B (58:22)

From Ryan Lizzie out asking about the bamboo, which we're going to get to. So news broke that she had had a digital and emotional affair with RFK while covering him, which everyone was like, oh my God, crazy. Her and Lizzo getting kind of a spat for various other reasons, obviously separate. A lot of time has gone, a year has gone by.

A (58:40)

He was her fiance together.

B (58:44)

They also had a book contract together. There is a New York Times Times profile of Nuzzy that comes out in the last week. That's like Olivia Nuzzy did it all for love. That tag to her upcoming release of a book called American Kanto, which is going to be about kind of this whole affair, to use multiple definitions of the word.

A (59:05)

You could see, by the way, why they talked about this for an hour and a half on her date last night.

B (59:10)

I was gonna say.

A (59:13)

Get us back to a. I'm doing really good then.

C (59:15)

Yeah, you can't.

B (59:16)

They had an excerpt of this published in Vanity Fair, which a lot of people poked fun at online because it had a lot of overwrought metaphors.

A (59:24)

And of course, Paris breathlessly texted both of us, oh. Immediately I was like, we have to discuss. She's got to read this.

B (59:33)

Then earlier this week, Ryan Lizza, who has left Politico as well and now has his own substack, released a substack column called Part 1 How I found out. Everyone, myself included, kind of ignored the part one in the headline. I was like, oh, it's a blog about how he found out about the whole RFK thing. And it's this kind of again, overwrought written thing about. He's Talking about the bamboo behind their story. Georgetown Townhouse.

C (60:04)

You can't kill it.

B (60:05)

Like, you can't kill it. It's something that's invasive and growing there. And he keeps going back and forth about the bamboo and these notes that he found in Olivia's bag and the bamboo and this. And then it's the.

A (60:16)

Don't spoil it. Don't spoil it. No, don't.

B (60:19)

I'm not. I'm pausing.

A (60:20)

It probably is meaningless to. To most normal people.

C (60:24)

It is. It absolutely is.

A (60:26)

But there's quite a twist.

C (60:28)

Very much a twist. Yeah.

A (60:29)

Go read newsletter is telos T E L O S. If you want to read the twist.

C (60:34)

Not now.

B (60:35)

It's really worth it to read it for the twist.

A (60:37)

That's why I don't want to spoil it for people, because if they read it, it's going to be like, what? That's not. Not what we thought it was good.

B (60:43)

You didn't get that there was a twist the first time you read it?

A (60:46)

No, because I was so enamored of the bamboo. I mean, I got to the end.

B (60:50)

Until you said, it's a quick read.

D (60:53)

And the bamboo is.

A (60:54)

You said something. And I went, what? What? And then I finished it. It was like, oh, not what I was expecting. So it was not about RFK Jr. As it turns out.

C (61:05)

Well, well, now you've given away much. But not all. Not all, but. Okay, now we can go back to AI So that's what you've been. Everybody has been talking about.

A (61:13)

AI did a pretty good summary of this. Better than, frankly reading all the nonsense about it. Although, weirdly. Gemini links to realtor.com?

B (61:26)

Where does that take you to? Does it take you to.

A (61:28)

It takes you to Realtor. Well, what it is, it's a bunch of sources. Not just realtor.com but apparently. Realtor.com Go ahead, show my screen. Realtor.com did, in fact write an article.

C (61:36)

About this because everybody wanted in.

A (61:39)

Everybody wants in on this. So why they didn't say the Guardian read the screen.

B (61:44)

If you.

A (61:45)

Yeah, it's giving it away. It's giving it away. But if you do. If you don't care, read the screen. So it has. So, so it has bullet points. The RFK scandal, the new accusation, the war with Ryan Lizzo. Lizza. Which is kind of backtracking a little bit. Yeah. Do they ever call them Nizzas or Luzzies? Like, you know. No. Okay. Her current status, despite the ethical controversies.

C (62:14)

Well, there's, there's. There's the fact that vanity Fair would hire her as a reporter.

A (62:17)

That's quite an ethical controversy.

C (62:19)

As my wife said, she hopes that the governor of California stays away.

B (62:24)

Yeah, it's not.

A (62:25)

Well, remember his earlier? He was Kimberly Guilfoyle.

C (62:29)

So stay away.

A (62:30)

He's vulnerable, I believe.

B (62:32)

Let's go back to.

C (62:34)

Okay, now we go to Ant.

A (62:35)

Anyway, Gemini, impressively. And the other thing that's impressive is I didn't leave my custom gem, which was my common lisp stuff, and it's still got this all. So I don't know, maybe I should have had it write it in common lisp. I will. I will get out of the common lisp. So that was in the thinking part. And that's what you were talking about. Jeff, you have fast, which I used earlier today to find an ironing board.

C (63:01)

Two, five. For most people.

A (63:02)

Had I known, by the way, I would have just. It should have referred me to Paris's tweet about the things she found on the sidewalk, because I could have gotten an ironing board there.

B (63:13)

That's true.

A (63:14)

Somebody noticed that. By the way, you have followers in the Discord who said, well, Paris already found your ironing board for you. What is that on Blue Sky?

B (63:23)

Yeah, yeah. No, but in what context were you missing an ironing board?

A (63:27)

I was complaining that I have a lumpy ironing board. And. And Richard Campbell.

C (63:32)

You were taking us back to AI. Yeah, yeah, yeah.

A (63:35)

Richard Campbell said, leo, you can afford to buy a good ironing board.

B (63:41)

I need an ironing board. I'm still ironless.

A (63:43)

Well, I found. But the problem is the good ironing board is 130.

B (63:47)

Okay, that can't be correct. Correct. That has to be a normal price.

A (63:51)

Buy a nice ironing board, don't buy a cheap one. You know, cheap Amazon ironing. Anyway, Gemini did a very, I think, did a very good job both in the ironing board search, in the synopsis. If you read it on my screen, I think did a very good job of synopsizing this really kind of pot boiler that is of only of interest to people in the media and Paris's dates. And I will show you. I did ask it in the thinking mode to help me with lisp before I asked about Olivia Nuzzy. Oh, no, Now I can't find it because the Olivia Nuzzi article. Oh, there it is. Okay. I said, explain how the DO loop works, which is a kind of a obscure part of common lisp, and it showed its thinking. Now, here's this. Show this bonito, because I think it's of interest. So you don't have to show the thinking, but in effect, it wrote me a book about this particular Lisp syntax. And was that more than you wanted or.

C (64:57)

It was.

A (64:58)

Okay, well, you know, I. All I said was explain how the do loop works. Very simple prompt. So here in the thought, it says the user wants an explanation and it's actually giving itself a to do list. Identify the user's goal, scans course documents for keywords, information. The reasoning stuff is very, I think, impressive. Execution flow, drafting the explanation. It's actually going to write an article and this is the outline of the article. It's got examples. And then finally it does a final review. Clear structure. Yes. Cold examples provided. Yes. Explanation of syntax. Yes.

C (65:34)

Good for you. Good for me.

A (65:36)

It did a lot of work. And then finally, at the end of this. Keep showing it at the end of this very long thinking process, which again, you know, I could just turn off. It gave me quite a long. And I thought, oh, wait a minute, it's still giving me the thinking below the thinking. It gave me quite a good. Basically, what could have been a blog post with sample code. It really, really good.

C (66:07)

Yeah, I asked it. I've been asking all of them because it's part of my research for my next book about views of public opinion, definitions of public opinion through various schools of thought. And so I'm using that as kind of a way to test all of them. And Gemini 3 did a very good job of that. And it draws charts, too. We compare them and all kinds of things.

A (66:29)

I guess we could ask. I'm going to take another break. Should we. Paris, why don't you give me a prompt I can put in here, Maybe a writing prompt, since you didn't like the bamboo analogy.

B (66:39)

I mean, I'm gonna be honest, my brain's all right now.

A (66:44)

Why don't you put in the style.

C (66:47)

Ryan Lizza and ask for a critique of the writing.

A (66:50)

Oh.

C (66:53)

See if it says bamboo is overdone here. Let's see.

A (66:56)

Okay, that's.

B (66:56)

And for those in the chat, confused. The bamboo we're referring to is an overwrought analogy that is sprinkled throughout Ryan Liz's piece about how he found out in response to Olivia Nuzzy's book release.

C (67:12)

And somebody posted just the six bamboo paragraphs to mock it.

A (67:17)

Okay. I'm going to say critique the writing style of this article and I'm going to give it a link. That's all I'll do, right?

C (67:22)

I think so.

A (67:23)

It should be okay. Now, when we come back after the ad, Gemini 3's critique. And let's see if it agrees with me or Paris and Jeff, who are professional writers. Although I have written more books than both of you combined.

B (67:41)

Written or published?

A (67:43)

Published. Thirteen volumes, my dear.

B (67:46)

I know, but that is not the specifics of the question. Did you write the books or.

A (67:50)

Oh, no, I wrote them with help. I wrote them with.

C (67:54)

Did I tell you I went to a convention of ghost writers last week?

A (67:58)

Yes, you did.

C (67:59)

Yes, I did.

A (67:59)

Yes.

B (68:01)

We got a recording from that convention right here. It's.

C (68:10)

Paris. Paris is hopped up on newsy.

B (68:13)

I am really. It was an electric week to be in media group chats.

A (68:19)

You were in one of the group chats that they referred to, weren't you? You, You.

B (68:22)

That was literally in one of my group chats. My friend Julie and former co worker Julia Black and I were like, wow, great week to be in media group chats. And then we both tweeted about it and Leo sent me her tweet later, and I was like, yes, I know. I helped.

A (68:36)

I was in that group chat.

B (68:37)

I was.

A (68:39)

I'm just jealous because I'm not in any. The only group chat I'm in is with you guys.

C (68:42)

Yeah, I know. That's.

A (68:44)

It's the only group chat I'm in. Oh, that and our contractor. But most of that just yelling. Let's take a break. We'll have more in just a moment. More newsy news in just a bit. I can't wait to read the critique. It's working hard. I think you'll like it. This episode of Intelligent Machines is brought to you by Zapier. Love Zapier. Zapier has been my longtime companion in automation automating stuff. My workflow, everything from turning my lights on at sunset to extracting all the information from my bookmarks to prepare these shows. And the beauty of this is, once you set up the workflow in Zapier, it just happens. I don't have to think about it. I've been using that Zapier workflow for our shows for several years now, and it's amazing. But now Zapier is even better. One of the things we talk about so much, AI is now a big part of Zapier. The problem with, you know, the trend, the buzz of AI is it doesn't necessarily help you be better at work, right? You need the right tools. And this is where Zapier is going to help you so much. It's how you break the hype cycle and put AI to work across your company. Now, with any luck, you're already using Zapier. If you're not, you should absolutely check it out, because it connects with over 3,000 tools, tools that you already use, probably in work like, you know, Google Drive and Atlassian tools and all of that. It connects them together to make workflows without coding very easily. But now Zapier can let you deliver on your AI strategy, not just talk about it, because Zapier has become an AI orchestration platform. That means you could take any workflow, any of the ones you've already got, for instance, and add the power of AI so you can do more of what matters. Connect the top AI models like Chat, GPT and Claude to the tools you're already using so you can add AI exactly where you need it, or even built an AI full AI workflow out of Zapier. It's so fantastic. For instance, so my workflow is I bookmark an article, it then goes to my bookmark site, which Zapier picks up and then puts on my mastodon. And then it takes it and formats it and puts it as a line in a Google spreadsheet. So the editors and the producers of the show can take that and put it into our rundowns. It's in a format they already use, and Zapier does all of that automatically. But now I could just inject a little AI in the middle there and say, and now write a briefing book, do a synopsis of of the 10 stories that are most important and send that out to the hosts. I mean, the sky's the limit. You could do AI powered workflows. You could take existing workflows and add AI. You could create an autonomous agent, a customer, chatbot, whatever you can think of, you can orchestrate it with AI with Zapier. And you don't have to be an expert, you don't have to be a coder. It's for everyone. Teams have already automated over 300 million, 300 million AI tasks using Zapier. Join the millions of businesses transforming how they work with Zapier and AI. Get started for free by visiting zapier.com machines@z-ap I-E-R.com machines thank you, Zapier. You've been a big help in producing this in all of our shows for years. And now I can add AI to it. It's fantastic. Okay. Based on the available excerpts and context surrounding Ryan Liz's article, the writing style can be critiqued as a blend of narrative journalism and confessional noir. That's right. Heavily reliant on novelistic devices to elevate what is essentially a personal grievance into a public silence.

C (72:39)

That's insightful.

A (72:40)

Yeah. He employs the. Let's see if they mention bamboo. The opening scene is set with specific cinematic details. The Herschel backpack with the flap that never quite closed properly tossed beside the bed. The pacing and structure of a detective novel or a political thriller. This is a classic show don't tell technique. Okay. Melodramatic and distinct imagery.

C (73:06)

You're muted. Paris. I can't hear your chortles.

A (73:08)

Brand name realism. Tom Ford. Herve Leger, Kimpton Hotel Stationary. $15,000 worth of Cartier jewelry. The burden.

C (73:19)

The burden.

A (73:20)

Narrator. Persona.

C (73:21)

Nice.

A (73:22)

The voice. Yes.

C (73:25)

Oh, very nice. Yes, very nice.

A (73:28)

I think this is good. I mean, if I were an English teacher and I, I. I could save myself some time.

B (73:33)

It doesn't mention bamboo.

C (73:37)

Ask it. Ask it what it thinks about that bamboo.

B (73:39)

What about the bamboo metaphor?

A (73:45)

What about the metaphor?

B (73:46)

What I was saying before when I was muted is that its critique of the water tower line is inaccurate because that was technically nuzzy.

A (73:55)

That was her line. Yeah. He didn't know the difference between nuzzy and Lizza, which, if you don't know, I don't blame you.

B (74:03)

I really love how much of this podcast has been consumed by this.

A (74:07)

Ah, I should have known, right? I should have known.

C (74:10)

We'll get back, folks.

D (74:11)

We promise.

A (74:11)

It consumed our group chat. Bamboo metaphor is the article's primary thematic anchor, functioning as a heavy handed. Heavy handed literary device to symbolize the invasive, uncontrollable chaos in the nature. It's technically sound, but unsubtle.

C (74:29)

Very good, Gemini.

A (74:31)

Very good.

B (74:32)

Very good.

A (74:33)

Yes. The literary pretension. Oh. Liza attempts to elevate the piece from tabloid gossip to literary memoir. It feels performative.

D (74:45)

Oh. Ooh.

A (74:46)

The oscillation between. Here's a text message about a sex scandal and let me muse about the resilience of bamboo creates a tonal whiplash. It signals to the reader I am not just an ex boyfriend vetting. I am a capital W writer.

C (75:00)

Oh, wow.

A (75:01)

Observing the human condition.

B (75:03)

That's actually kind of good.

A (75:05)

Many readers found this self aggrandizing, serving to center his own suffering and hard work rather than the events itself.

C (75:12)

Wow.

A (75:13)

So you know what? The bamboo metaphor is a symbol of the article's excess. It is a valid literary parallel, something beautiful but destructive and invasive that is ultimately overworked, making the piece feel calculated and overly polished rather than raw or authentic. C minus. See me after class.

B (75:35)

All right. So. Pretty good.

A (75:37)

Are you impressed?

C (75:39)

I'm Impressed.

D (75:39)

Yeah.

A (75:40)

Now it probably stole that very.

B (75:42)

I was gonna say it definitely stole that from all of the criticism of it online. And I would have been more impressed if it brought that up without having to be prompted about it. But.

C (75:51)

But you're trying to get the sand out of your socks now. Paris. It's not working.

A (75:57)

Incidentally, wasn't it just a few months ago that you would ask an AI something about current events and it would have no idea seamlessly without even making a big deal about it?

C (76:09)

Search is now completely.

A (76:11)

This is up to the date. This is up to the minute.

C (76:14)

Minute. I wonder whether. I mean, Google obviously has a current crawl. I wonder if you asked the same thing of let's not do it.

A (76:21)

But chatgpt.

C (76:23)

But chatgpt.

B (76:24)

I could spend one more hour on this, guys. I think our listeners will love this.

A (76:29)

So are you gonna. This is the most important question. Question. Will there be a second date?

B (76:35)

No.

A (76:36)

No.

B (76:36)

But that's already been discussed. That's not. I'm not breaking any news here.

C (76:41)

Okay, so somebody I follow on. On Blue sky or Twitter, I really like them. Sophie Vershbow because she has the greatest dog there is on the whole Internet named Simon. So she had a picture up of the. Of the Paris guy, the guy with the red strings and all that and the bulletin board and all that.

B (76:59)

Oh, by Paris guy you mean the character from Always Sunny in Philadelphia who's creating a red string board? Yeah.

C (77:07)

So she captioned it. This is.

A (77:09)

We always think of that.

C (77:10)

This is me trying to explain the doozy story to my boyfriend. It is like normies out there. The normies out there just think we're all crazy. But yeah. Which includes the entire audience of this show. And we apologize.

A (77:22)

We apologize.

C (77:24)

Teen founders apologize.

B (77:26)

I don't know.

A (77:27)

This is my favorite thing young people I've heard of. Founders have raised $6 million to reinvent pesticides using AI.

C (77:40)

It sounds cool, but.

B (77:42)

Wait, it does.

C (77:44)

But she's made. They're making things that can kill.

A (77:46)

Exactly.

C (77:47)

And do we know the impact of it on the environment? I don't.

A (77:50)

Well, the AI would know all of that.

C (77:52)

Well, no, it wouldn't. It wouldn't have any clinical basis to. To test.

A (77:57)

The company is called Bindwell. They didn't spend any of that money on the website. I'm just going to point out we're building better pesticides using AI.

B (78:07)

This is a very Gen Z style website.

A (78:09)

To be honest.

C (78:10)

That's a really bad drop shadow.

A (78:12)

That's. The drop shadow is horrible.

B (78:14)

No, that's kind of the point is it's supposed to be.

A (78:16)

Is it ironic? Ironic drop shadow. Okay, yeah, it is.

B (78:20)

Ironic drop shadow.

A (78:21)

The agrochemical industry.

C (78:22)

So glad we have terrorists here to instruct us in the ways of the we don't.

A (78:26)

We don't get any of this stuff. Tyler Rose from Wolfram Research and Navi Anand from Caltech. A scrappy duo of engineers from China and India respectively. And both close to farmlands in our countries. United by our passion for tackling global problems. We've dropped out of high school and college. Wow. To start Bindwell, they've already raised $6 million.

C (78:54)

They invaded Paul Graham's backyard, evidently and, and convinced him that this was wonderful and he wrote a check and then introduced them to others.

A (79:02)

I think that's fascinating. I mean, it's one of two things. Either they're onto something or it's a real demonstration of the overheated investment environment around AI.

C (79:16)

They come from countries where this is an issue for the productivity of their family's farms. I get that. But has there been sufficient environmental and health due diligence on these? Just because AI made it up doesn't make me comfy.

A (79:34)

A pat on the back to our friends at Kagi. Remember we had the founder and CEO of Kagi, which is a public benefit corporation I use in lieu of Google from my search. They have a lot of very interesting stuff. In fact, COGI Assistant, which is their AI orchestrator, is in my mind as good as perplexity. And I've kind of of replaced perplexity with it. Have also now just announced something called Slop Stop Community driven AI slop detection for COGI Search.

C (80:05)

So it's community driven.

A (80:07)

Yes.

C (80:07)

So that's going to be hard to scale, isn't it?

A (80:10)

That's what they told Jimmy Wales. AI slop is deceptive or low value AI generated content created to manipulate ranking or attention rather than help the reader. So they've actually made this possible. Make it very easy. If you're using Kagi Search to flag low quality content in web image and video search results, we will verify these reports using our own signals. If a domain primarily publishes AI generated content, we will downrank it in Kagi Search and mark it as AI slope.

B (80:48)

And how they're determining that something is low quality AI content is exclusively from human reports. Or is it a combination of human.

C (80:56)

Reports plus what's the false positives there?

B (81:00)

Yeah, because my issue is that I. We've talked with Somalian times in the show, but there are all these tools that claim to be Able to detect when AI writing is present and Right. They're just not that accurate.

A (81:11)

Yeah, well I think that's why they're using humans to flag it. And then they say we will use this data set to bid our build our own AI content detection tech. They aim to build the largest data set of AI slop domains on the web using in house built detection and a carefully curated community reporting system. So what they're going to do is take these reports and try to and create a data set once they've validated that it is AI slop. And you know a human, human probably can look at something and tell that I would guess. Anyway, what you'll be able to do as a coggy user is flip a switch that says no AI. No AI generated content. I think that's, you know, it's a worthy thing. You're right. I mean that's reasonable worthy.

C (82:04)

But it reminds me of the earliest Yahoo. Where they thought human beings were going to. We're going to category catalog the entire web. Web.

A (82:12)

Right.

C (82:12)

It's a little charmingly naive.

A (82:15)

I think they have a documentation page on slop. Stop.

C (82:20)

Jason told me. Jason Howell on AI Inside told me today he was out walking the dogs and he got a pitch. I can make a hundred health oriented Instagram posts in one minute, right?

A (82:33)

Oh, I've been getting pitches not just from AI, but just like that in general. They sell Blue sky followers, they sell articles.

C (82:44)

The volume of slop is going to overcome the taggy people.

A (82:48)

Absolutely.

C (82:49)

But also how AI has been influencing how people write. So at some point people are just going to write like AI. Like normal people, not professional writers.

A (83:00)

I have already decided never to use the EM dash again, if that's what you mean.

C (83:04)

Exactly, that's what I mean.

D (83:05)

Though I hate that you can't seed.

B (83:09)

The EM dash to these clankers.

C (83:11)

Paris, we have to. Right here. We have to start the M Dash defense league.

B (83:16)

I'm down.

A (83:17)

I gotta repeat that we can't say it again. That we can't seed the embassy.

B (83:21)

We can't seed the EM dash To these clankers.

A (83:24)

To these clankers. Just wanted to make sure.

C (83:26)

It's a little long for a title, but it'd be a good title.

A (83:28)

Yeah, it's good. It's good. Goofy. You were saying? I'm sorry, I just froze.

B (83:38)

We all. We all froze. I thought we all froze. I thought I was waiting. You know, I really like seeing people in the chat not know how M dash is spelled. This is perhaps the one time I'll be pedantic It's E, M, not the.

A (83:51)

Letter M. And then there is, interestingly, a non M dash in M dash.

B (83:57)

P M, N dash.

A (83:58)

Hyphen dash, right?

B (84:01)

Yes. EM hyphen. Yes, hyphen.

A (84:04)

Is that how we distinguish an EM dash from a hyphen? Is one's a hyphen, one's an M.

C (84:08)

No, there's an en dash. In between, there's EM dash, EN dash and hyphen.

B (84:12)

Yes.

C (84:13)

You use an EN dash for ranges like, like 1920 to 1946.

A (84:18)

That's an N dash. And the reason they're called M and N is because of the width of the character M. The character M N.

C (84:26)

The M is essentially the square. So if it's a nine point font, it's going to be nine point by nine point. That's an EM dash. An N is half of that.

A (84:35)

Okay.

C (84:36)

And so on and so on.

A (84:37)

How big is the height?

B (84:38)

2N dashes. That makes an em dash, baby.

A (84:43)

Or as I like to call it, a double en dash. So talk about investments.

C (84:49)

And you thought that you didn't want to hear about Lizzo. You were hearing all about dash. This is the heart of it all.

A (84:56)

Have you. Did you know the AI note taking startup Fireflies? Is that. Does that ring a bell?

C (85:03)

No.

A (85:03)

The AI note taker for meetings, they raised a billion. Well, they got valued at a billion dollars by their raise. Here it is Fireflies AI. But here's the interesting thing about Fireflies AI. It's two guys. It's two guys. Two broke guys living on pizza. The founder, one of the founders, Sam udotong, in a LinkedIn post, Futurism had the story, admitted his AI transcription startup was just him joining people's meetings and taking notes by hand. No, There was no AI at all.

B (85:47)

Oh my God. There was no AI at all.

D (85:51)

They.

A (85:52)

Yeah, they announced that 75% of Fortune 500 companies were using their services to transcribe corporate meetings. They announced a billion dollar valuation. Well, they got their start. Let's put it this way. In 2017, charging $100 a month for an AI that was really just two guys.

D (86:14)

Maybe later.

B (86:15)

AI, when you can have two guy.

A (86:17)

Yeah. Whenever a customer needed notes, he said either he or his co founder Chris Ramanani would dial into the room as Fred, masquerading as a Siri like equivalent. We'd sit there silently, take detailed notes and send them. Ten minutes later they were able to make rent, keep their startup dreams alive. After doing 100 meetings this way and then raising money, I think they have since Automated it. Oh, okay. Now they do use AI, but it was fake it till they made it. Basically.

C (86:53)

2017 though, you know, like.

B (86:55)

Yeah, 2017 was a time of manual transcription.

A (87:01)

Yeah. You know what, what is the. What is the etiquette on this, by the way? Lisa says if somebody, if she's in a zoom call and the AI agent joins, she will say, I get upset. Yeah, she gets upset. She says, turn it off.

C (87:17)

I know lots of people who do it. They'll offer to take it down, but they keep it in there. But now in Zoom, you can just turn on your own. Everybody can use it. Everybody can turn on the assistant.

A (87:25)

It announces it though, right? It doesn't do it silently. We did twit a couple of weeks ago, remember this? Benito and Ian Thompson couldn't get in, but his otter AI got in and so kind of gave away the game. Yeah, I guess he used it for his journalistic enterprises and it was just turned on and he turned it off. So it's rude.

B (87:51)

I find it rude.

C (87:52)

Well, but I think it's going to change. So Jason and I had an interview with a Microsoft executive yesterday that they introduced their agent framework where it's an open source thing where they're going to give identities to all the agents. And their presumption is that agents are going to join all kinds of meetings and. And all kinds of functions.

A (88:13)

And what, it's just another kind of AI slop that is?

B (88:17)

Yeah, it's agent Slop. Can you give me the good faith answer of what the heck does that mean? And why would an agent with one personality versus another be relevant? Business. Business.

C (88:31)

It's a function. Right. So one friend of mine uses it and it will summarize the meeting and it'll remember the. The to dos. Everybody has the assignments.

A (88:40)

So there's like one AI to do that, one AI to take lunch orders and then one AI to remember that could be. Say nice things about the CEO and.

C (88:49)

See, I'm not in Slack like you guys. You're in Slack, right?

A (88:52)

Yeah, we use slack.

C (88:53)

So don't you have an AI in Slack that you can have? You could. Could you turn to.

A (88:59)

Never used it? Yeah.

C (89:01)

Isn't that possible? Then you can say, hey, AI, explain to Leo why he's wrong about that.

A (89:07)

Benito, does anybody use it?

C (89:08)

I don't think anybody touches the AI in Slack here. No, but it does exist, right, Benito?

A (89:13)

I think it exists. I don't even know how I get to it. Patrick, have we disabled it?

C (89:19)

I know that Patrick has asked me a bunch of Times in the past, like do you want me take to turn the AI stuff on? I'm always like, no, go, go.

B (89:26)

Hey, at least it offers you. It's not like Google, whether you like it or not, just appear and you can't turn them off.

A (89:34)

Huh? Yeah. And well, Zoom has it now. Everybody open.

C (89:38)

AI announced today that you can, you can do them in, in a, in a chat. I get some functionality to it.

A (89:45)

I look transcription seems like a good tool. I mean you transcribe your interviews, right? Paris, you use Otter AI or something like that?

B (89:53)

I don't use Otter, but I use MacWhisper a lot where I am able to models locally so that my interviews are secure.

A (90:01)

You used to use a cloud service?

B (90:03)

I used to use Trent for many years and I enjoyed that. I never really liked Otter just because I didn't like the design of it. And I think recently as it's become kind of like this default thing for meetings, it's become very optimized for the meeting meeting for like the middle manager meeting planner class in a way that's not very useful for my purposes. But yeah, I think there's like a lot of etiquette around AI note taking assistance that people are still figuring out. Today I was in a meeting and you know, as things often do, something veered off and like one of the people in the meeting was talking about her birthday. She was born on New Year's Day or whatever. And we're about talking, talking, we're all joking. And she was like, oh man, now this is going to be recorded by Gemini. There's going to be a Gemini summary of the fact that we're all joking that I'm the New Year's baby. How gross. And these are just like the strange new realities we live in.

A (90:58)

That is a legit concern is that any cloud based AI is presumably recording all this information for its own benefit as well as for yours.

B (91:12)

I mean you just have to do a lot more work to figure out who's handling that data and what the privacy policy is and where it's going.

A (91:18)

And if, and if somebody turns on an agent. In our conversation now, it's on me to find out what the privacy policy is and whether I.

B (91:27)

And you're in a meeting currently, you.

C (91:29)

Know, says the guy who was recording everybody who was within three feet of him.

B (91:33)

Yeah. With a variety of pins.

C (91:36)

Yeah, yeah, yeah.

A (91:37)

They're all retired by the way, and the batteries have died now, so you're safe. I wanted that. But you're right, it should be local it shouldn't be uploading it anywhere. This is the solution to all of this is local AI. Local, local, local. It's not there yet. In fact, it may never be there. We were talking yesterday about this with Steve Gibson. He said this validates, in some ways, this validates the cloud, because in order to do good AI, you need huge resources, but you don't need them all the time.

C (92:10)

Right.

A (92:11)

So it's not economical for Everybody to have 100H, 100 Nvidia GPUs on their desk. If they only use it, you know, an hour a day, it makes perfect sense.

C (92:25)

But is it also true that some of the functions you ask for, like transcription, don't require a huge AI either? They can be done locally.

A (92:33)

Those can be like, as Paris does with whispers, local.

C (92:36)

Right. So that's the thing, is that I think the more we get to small models, the more controllable it is and held locally.

A (92:41)

Yeah. In fact, that may be an argument for models that do one thing well.

C (92:45)

That's what I think more and more and more is this notion of the everything machine is what leads everybody astray. I went to a great event on Sunday night, thanks to Jason, got invited and he couldn't fly into Brooklyn, so I went. Went where? Yann Lecun, who just announced that he is leaving Meta.

A (93:03)

Let's talk about that. Yeah.

C (93:04)

And. And Adam Brown from. From Deepthink, had a bit of a debate about whether LLMs will get us there, the Leo Laporte view, or whether there's a different model, which, of course is Yann Lecun's view, because he's going to move on. But. But Jan was talking about a future where things are purposeful rather than general. And what I infer from that is that there'll be a lot of AIs that together add up to something. But this idea that I'm going to have the perfect one machine that replaces all of humanity is hubris beyond belief.

A (93:42)

Specialized, isn't that what Kevin was telling us last week, is that he saw a future kk Kevin. I forgot his name. Anyway, good old people don't get old. Kevin Kelly. He's an old Kelly. Thank you.

C (94:00)

Kevin Kelly. Our good friend Kevin Kelly.

A (94:03)

Our good, dear, close, personal friend Kevin Kelly.

C (94:05)

Nobody told him that he did.

A (94:06)

That is eventually AI. No, I'm sure Kevin does the same thing. Eventually AI will be embedded. Little AIs will be in everything. Like, just as that's what y' all was saying has gone to the edge. I mean, we now have computers in Our light switches, we have computers everywhere. AI will similarly expand to the edge, but they can't by necessity, they can't be general AIs. They will be very specific AIs. The processor in your light switch isn't that bright. You know, they've somebody put Doom on a pregnancy test because there's a microprocessor in it and there's enough memory and enough CPU cycles to put the video game Doom on it and there's a little screen.

C (94:49)

So this is like the opposite an.

A (94:50)

Example of how computing is now everywhere. Right. And I imagine it'll be very much. AI will be everywhere.

C (94:57)

Yeah, that's, that's what. So I was at Pioneer Works in Brooklyn. John 11 leads these sessions they call Scientific Controversies. And it was a great discussion. It was just amazing. And I got to meet Jan afterwards. But it's a, it's a really interesting debate here about where we get there and what's his name? Adam Brown from DeepMind. See, but I just did it too. I said his name three minutes ago and then I forgot it. So Adam gave your argument, Leo, which is that look at the, at the vector of improvement. If we just keep going along that velocity, we will reach AGI, that it's just going to happen. And Jan was saying, no, it's not. And Jan's old enough that he could go back through all the histories of the various stages of AI and say we've seen this before and using just text or static images is a lot easier than. And he used the example of trying to pick up a glass of water next to him. He doesn't mean that just robotically. He means that how the AI can interpret actions and your hand doesn't go through your body to get to the glass. And so his view, as he's been talking about quite a bit, is about real world models. And he had a paper with Fei, Fei Li I mentioned last week that he's trying to get the AI to understand the video of how to understand how many chairs are in a room. And that's not a small problem. And so he's doing it that way versus Jensen Wong's been doing it through robotics. But I think that we're going to see. My hope here, here is that by somebody of the stature of Yann Lecun and There were like 500 people at this event because he was there, that that money is now going to go to competitive research, which I think will be good for AI. We'll have different models that we're trying, different paradigms. To see what we're doing. So he just announced on Facebook and LinkedIn that he will be leaving Meta at the end of the year to start his new venture, yet unnamed. And we'll hear more.

A (97:04)

Yeah, Darren Okie's making the point. I think it's a very good point, that not to confuse LLMs, which are large language models, with Transformers, which is the kind of underlying technology of all modern AI. And he says that Yann Lecun, of course, knows the difference, but is confusing people by implying that Transformers are the problem.

C (97:27)

No, no, he said straight out, we will keep using that. And that's. That's tremendous.

A (97:31)

Okay.

C (97:31)

But he said that working with language prediction isn't sufficient.

A (97:35)

But I think people think LLMs. They think that's all. All the current crop of AIs. And it's not trans. Transformers were. That was the. Remember when we had Stephen Wolfram on, he said, you know, we. We shouldn't give up on symbolic AI, which is the old school AI, where you teach it everything and then it knows. And that's how Wolfram Alpha kind of works, is on symbolic AI. But the Trans. And for a long time, the heroes of AI were saying, oh, you know, Transformers got nothing. It's all going to be symbolic. It took a long time for Transformers to be accepted. We learned about this in, I think it was Karen Nao's book, that Transformers actually were the breakthrough. And I don't think anybody has credibly said, we're going to have something other than Transformers. So a lot of that investment, a lot of that work is really about transformers. LLMs specifically. LLMs are one kind, one way to train a Transformer. But if you're going to train it on the physical universe, you're still going to use a Transformer.

B (98:39)

I'm curious, Jeff, what was the mood in the room at this speech? Did you get to talk with any of the attendees?

C (98:46)

So John 11 has done 25 of these scientific controversies, and they're really neat topics. They're great. And she was very good at interviewing the two of them. And they weren't fighting tooth and nail, but they had a very clear, clear demarcated disagreement. And the room was wrapped. It was jammed. Paris, it was jammed. There was standing room. Only like 500 people in the room.

A (99:08)

Did you get a seat?

C (99:09)

Yes, I did, because I got there early. And then I also had the VIP thing, so I got to go upstairs, and that's where I got to meet Jan. It had a very good Korean spicy barbecue meatloaf slider.

D (99:23)

Ooh, that's a lot of fun.

A (99:24)

So what did you talk about with Jan? I think this is good.

C (99:27)

Just. Just kind of generally. He's. He was. At the first question he was asked downstairs at the. At the event was, okay, what are you going to do? And he said, I can neither confirm nor deny. So he was still. Had to be coy. He said he talked about what he called the LLM pilled people when we were upstairs, which rather amused me. And. And that they just see this world as that and that only. And he doesn't diminish the wonder and power of what these things do, but he thinks they don't get us there. And there he was asked at the end. There was only one question from a philosopher.

A (100:04)

I think it's a straw man. I don't think anybody's really saying, oh, no. LLMs are the only thing.

B (100:09)

Well, Adam, have you seen most press releases. Have you seen the way that most companies that are receiving these.

C (100:17)

That's what Adam said. Straight up.

A (100:21)

Is the thing.

C (100:21)

I have a video, but it was. It's to get. The true debate between the two of them is six minutes long, so we're not going to play it, but you can go to Johnny 11's Janet Levin's substack, and she has the full video there, which I recommend. It was really good. I think people were wrapped in this discussion, and then they saw the. The debate. I sat next to a vc, you know, and it's kind of. Are you on Team Yawn or Team Deepak?

B (100:46)

Did you check his boots for sand to see if he was.

C (100:50)

It was her. Her. Paris, you assumed the VC was a man?

B (100:55)

Yes. Venture capitalist at the AI Talk in Brooklyn was a man.

A (101:01)

Was she wearing a puffy vest? That's the question.

C (101:04)

No, she was not. She was much more fashionable than that.

A (101:06)

Oh, okay.

B (101:07)

I was gonna say it's too. Too cold in New York this week for a puffy vest. You'd have to be wearing a.

C (101:11)

This was. Have you heard of. What did I just call Pioneer Works? Paris?

B (101:15)

Yeah, Pioneer Works. Very cool. Not too far from me.

C (101:18)

It's just amazing.

B (101:18)

It's not too great art space.

D (101:20)

Yeah, yeah.

C (101:21)

There's no subway nearby, though, so I actually.

B (101:23)

Oh, wait, no. I'm thinking of a different space that Pioneer Works used to occupy. I know which one. The one you're talking about in Red Hook. I haven't been to that space, but it seems.

C (101:31)

It's amazing. They have media labs there and other things. There was really quite an event. I can't Wait to go to others.

A (101:37)

Let me take a little break then, if you will pick from your pile of stuff.

C (101:41)

Oh, we got them. Well, we got big news like Nvidia.

B (101:45)

I got some breaking news that's happened during our interview too.

A (101:49)

Okay, hold that thought. Good stuff. Still to come, you're watching Intelligent Machines with Consumer Reports, Paris Martineau and Montclair State University's Jeff Jarvis. SUNY Stony Brooks. Jeff Jarvis, the author of the Gutenberg Parenthesis. Jeff Jarvis Jarvis, who's an actual writer. Two actual writers and then there's me.

B (102:14)

Wow, we're back in the same wavelength, Jeff.

A (102:18)

There they are writing on their imaginary machines. This episode of Intelligent Machines is brought to you by the Agency Build the future of Multi Agent Software with Agency. A G N T C. Yes. Now, an open source Linux foundation project. Agency is building the Internet of Agents, a collaboration layer where AI agents can discover, connect and work across any framework. And the agency's got all the pieces engineers need to deploy multi agent systems. They belong to everyone now who builds on agency. And it's open. It's open. Including robust identity and access management that ensures every agent is authenticated and trusted. Speaking of trust before interacting, Agency also provides open standardized tools for agent discovery, seamless protocols for agent to agent communication, and modular components for scalable workflows. You'll be collaborating with the best developers from Cisco, Dell, Google Cloud, Oracle, red hat and 75 plus other supporting companies to build the next gen AI infrastructure together. Agency. They're dropping code specs and services. No strings attached. Visit agency.org to contribute. That's agn tcy.org agency nice to. Nice to be able to promote something that's making a big difference like that. That's fantastic. Okay, breaking news. Paris Martineau, I'm gonna need your help.

B (103:57)

To read this because it's an information article that I don't have access to. My wonderful former colleague Sylvia Varnimo. Regan reported just as we were getting on that the White House is working on an executive order that will direct the Justice Department to sue states that pass their own laws regulating AI.

A (104:20)

I don't know. I mean, obviously he can tell the Justice Department do whatever they want. Let's put Eric justice now because yeah, they're his official personal law Law Agency. But I think a court would stop that. I think the states are allowed to make their own regulations. Look what Illinois has done with Face id, what California has done, requiring chatbots to tell people they're a chatbot. Colorado has done the same thing. How can they. So the order would direct the Commerce Department to restrict federal. Oh, this is how the order would direct the Commerce Department to. This doesn't. This is. This is so evil. To restrict federal broadband funding for states whose AI laws it determines could hamper the United States leadership in the international AI race. And it would direct the Federal Communications Commission, another hammer of the Trump administration, to work with White House special advisor aizar and all in bestie David Sachs. To begin. I added the all in bestie part to begin drafting a federal standard for AI models that align with maintaining that leadership. Or. Or maybe the leadership of the existing incumbent big tech companies investing in AI with the ultimate goal of Sachs recommending that Congress codify such a standard for AI models. Congress tried this once before. Remember Marsha Blackburn's law that would preempt state laws, state regulation, which failed.

C (106:01)

Yeah.

B (106:02)

This is Punchbowl News. Had reported today also that the. This executive order is basically the White House's backup plan if Congress fails to pass its kind of preemption, its own version of this kind of preemption bill. And they are trying to do that currently by kind of rolling it into the upcoming reauthorization of the National Defense Authorization Act.

A (106:28)

Oh, they love Laden stuff. Larding stuff onto that act. Yeah, it was part of the big beautiful bill. The proposal was killed by the Senate by a 99 to 1 vote. So there's a definite point of view, at least in the Senate, that you shouldn't ban states from regulating AI. But David Sacks, not to be defeated, has continued to push this. Also, a Trump aide that I'm not familiar with, Michael Kratzios, has been pushing for the White House to limit states on AI. I mean, this is clearly lobbying from people like Google, Microsoft, OpenAI.

C (107:08)

Well, Peter Thiel, Marc Andreessen.

A (107:10)

Right.

C (107:12)

Yep.

A (107:13)

This order has not been signed. This is a breaking story. Sylvia Varnum o' Regan and Aaron Holmes writing for the information the White House.

B (107:22)

Is working on the Verge. Got a copy of the executive order or like a draft of the order? Yeah. According to this copy, this is a piece by Tina Win. The task for it basically would create this AI litigation task force overseen by the Attorney General, whose sole responsibility shall be to challenge state AI laws. This task force would be able to sue states whose laws are deemed to obstruct the growth of the AI industry, citing California's recent law as a catastrophic list of risk, as well as the call of Colorado law that prevents algorithmic discrimination as a problem. And the task force will occasionally consult with a group of White House Special advisers, including David Sachs.

A (108:13)

Yeah. Carr, at a appearance at POLITICO's AI and Tech Summit in September, according to the Verge, said, quote, effectively, if a state or local law is effectively prohibiting the deployment of this modern infrastructure, than the FCC has authorities to step in there.

C (108:33)

He imagines his authority like ChatGPT hallucinates facts.

A (108:37)

You know, it's so funny because Carr was one of the people in the Biden administration who argued that unless Congress explicitly tells you the FCC could do this, it can't. But now the shoe's on the other foot.

B (108:50)

He literally wrote the chapter of Project 2025 about kind of Internet regulation.

A (108:55)

Yeah, yeah, I understand the desire for, Look, I support, in fact, the desire from AI companies to have a overall federal law as opposed to a patchwork quilt of 50 state laws. But given that there is never going to be an overall federal law. In fact, the only law the feds have ever considered is one that would ban the states from doing anything. I think that that's un, you know, unrealistic. Yeah, it's not a good thing to have 50 different laws, but I think it's the state's rights to say, look, we don't want, you know, face ID unless.

C (109:34)

Unless you pass a federal law, in which case then.

A (109:37)

Right.

C (109:37)

You can't super. The local. Can't supersede it.

D (109:40)

Right.

C (109:40)

If you don't have in place.

A (109:41)

That's how it works.

C (109:42)

Right.

A (109:45)

And apparently 99 to 1, they. They can't agree on what that that law should be. So Nvidia breaking news.

C (109:54)

Yes, The Journal says that Nvidia profits soar soothing investor jitters over AI boom.

A (110:02)

Well, let me check my stocks. I don't have Nvidia stock. But as Nvidia goes, so goes the market.

C (110:09)

Exactly. You do it while the. I just checked the pre market for tomorrow and that always changes. But the pre market right now for OW type Jeff type.

A (110:20)

I just own a market basket of many stocks. It's not been a pretty thing for the last few last few days, but.

C (110:29)

It'S been pretty amazing. So the implied open for the Nasdaq tomorrow is up 377points. Nvidia is up aftermarket a few points as by the way, is Google we talked about earlier. So there's jitters are down. So the Nvidia jumped more than 5% in extended trading. Their earnings per share were 130 versus 125. Estimated revenue was 57 billion versus 54.92 billion estimated.

A (111:00)

I don't think this cures the market's Jitters though, because their jitters are not about today's profits, but about what happens in a year or two.

C (111:11)

Yeah, but they're so short term focus that if this had been down, they would have said, well here it is, the bubble burst. But it's not right. There's been a lot of talk about an AI bubble, said Wang. From our vantage point we see something very different. The most important business is the data center sales. Nvidia had 51.0 billion in data center data center sales, easily surpassing the guess of 49.09.

A (111:39)

How much of that was money that came from Nvidia, I wonder? Yeah, Nvidia has been paying companies to buy their chips.

C (111:46)

The circular drainage of that 43 billion was for compute or the GPUs driven by initial sales of the GB300 chips, which accounted for 8.22 billion in data. So anyway, it's a good, damn good quarter for Nvidia. And everybody's breathing a sigh of relief. In a island normally we don't do.

A (112:10)

Necessarily breathe a sigh of relief. I mean, I just saw an article that said Oracle is already Underwater on the $30 billion deal it made with Open AI.

C (112:19)

Yeah, who could have told you that one?

B (112:21)

Shocking.

A (112:23)

Jeff Bezos believes in AI. He is now again a CEO Project Prometheus, an AI startup focusing on artificial intelligence for the engineering and manufacturing of computers, automobiles and spacecraft.

C (112:41)

Actually, we won't see much of it ourselves.

A (112:43)

His. His space effort, Blue Origin is actually starting to make some headway. Bye bye Elon. And Elon's Artemis is struggling, so. So what a surprise.

C (112:55)

Yeah.

A (112:56)

Okay, I forget.

C (112:57)

Is Prometheus a cautionary tale? Prometheus took fire from the gods and.

B (113:04)

Gave it to human and it turned out great.

A (113:06)

Didn't he get his eyes poked out? I can't remember.

B (113:11)

I think everything turned out really great for Prometheus.

A (113:13)

And everyone did bring us fire. And that's good.

B (113:16)

Yeah, I think that's a great name.

C (113:19)

And civilization.

B (113:20)

Yeah, civilization. Really loved that.

C (113:22)

Sometimes credited with making the creation of humanity from clay. According to Wikipedia. Jimmy Wales. Thank you.

A (113:29)

Prometheus was condemned to eternal torment.

C (113:32)

Oh, well.

A (113:33)

Bound to a rock. And an eagle. He was bound to a rock. And then an eagle was sent to eat his liver. Right? Yeah. His liver would then grow back as livers are wont to do, only to be eaten again the next day.

B (113:48)

Unfortunate.

A (113:50)

Yeah. Although eventually Heracles did free him. So in another myth, he establishes the form of animal sacrifice practiced in the Greek religion. Guess where I'm reading this from? Wikipedia. Ladies and gentlemen, in. In Western tradition, Prometheus is a figure who represents human striving, particularly the quest for scientific knowledge and the risk of overreaching or under unintended consequences.

C (114:23)

Oh, well, there's that. They kind of do that all the time.

A (114:25)

Isn't. Isn't Frankenstein the subtitle Frankenstein the modern Prometheus? I believe it is. I believe it is overreaching. You see. It's your turn, Paris.

B (114:40)

I did one to start us off.

A (114:42)

I did one and Jeff did. So now.

C (114:45)

Now back to you. The wheel.

A (114:48)

Professor Plum, do we want to talk.

B (114:51)

About this Jack Conti opinion piece on line 170?

A (114:55)

Now, I like Jack Conti.

C (114:57)

I like Jack a lot.

A (114:59)

So let's see what he has to say. I'm building an algorithm that doesn't rot your brain.

C (115:07)

He's saying that attention economy algorithms do bad to you, so you. He. It's really quite a promotion.

B (115:13)

Okay, It's a promotion for Patreon. And I selected this because I hypothesized it was about the feature of Patreon that I hate the most. And it is. Which is I open up my Patreon app. Now, Patreon, for those not aware, is where you can subscribe to support your favorite creators and do various memberships for things. For me, it's almost exclusively. I subscribe to do podcasts I like that put extra content behind the paywall and have other stuff like live streams normally. When I say normally, I mean back in the day, I mean the last five years of using Patreon, I would open up my Patreon app and perhaps what I would expect to see upon opening the app would be the content from the creators whose Patreon think that's.

C (115:59)

Exactly what he says that he wants to give you.

B (116:01)

Well, now what I see is in the top, just a tiny little bit of. Of 30% of the screen is my. The creators I follow. The rest is just algorithmically, not generated, but an algorithm of other creators. Content that I do not subscribe to, I don't want to see and I cannot get to go away. And I understand why he's doing this because Patreon needs me to subscribe to more and more Patreons in order for it to make money and be profitable. But I hate it. If I am paying money for. I don't even want to count how many patrons are on here. I want to just see content from those. I don't want to have to click into multiple different menu buttons to see the stream of. It's the same problem that a lot of people have with Facebook's news feed and following feed. I mean, not that anyone's really using.

C (116:54)

Facebook condemns advertising as the basis for the attention economy because the of it makes you do exactly what you just said that membership makes them do.

B (117:02)

Yes, and this is part of the issue. I mean this has been a core issue for Patreon in particular. I wrote a story about Patreon some years ago, two, three years ago about how many of the people. Patreon's kind of a unique company. Well, not unique in this way. That many of the people who joined it were really motivated to participate by its mission. And its mission on its face is like incredibly noble. It's that advertising has ruined the Internet. It has created the perverse incentives of the intention economy and it makes everything worse. What if we made a better version of the Internet where money can just go directly into the hands of the people who are creating?

C (117:40)

That's what he says he wants to do.

B (117:42)

Yes, but that's difficult because the margins on that business are slim. And it's hard when you've raised a lot of money from venture capitalists and are beholden to certain growth expectations that pushes and pulls you in ways that are not perhaps conducive with taking just a small sliver of five dollar a month monthly memberships from a really plugged in and socio politically involved group of artists. Artists. There's kind of been this fundamental tension between the incentives, monetary incentives necessary to scale a venture backed business and the reality or the expectations of employees and users at Patreon for many years now.

A (118:37)

Couple of things I should say right now. One is that the Club Twit is, is powered by a Patreon company called Memberful. So your membership in Club Twit is facilitated by Patreon. I've known Jack since he was pamplamous in fact for us in the studio and I think so actually I'm watching the video. I hadn't seen it before so while you've been talking I've been kind of watching it. I don't disagree with his primary point at all.

C (119:08)

No, I don't think Pierce would either.

B (119:10)

I think many of these, it's just, it's a very interesting dynamic and one that I think about often. Ever since I would publish this article is just like how do you deal with the tension between having lofty and noble ideals and the reality of trying to scale up a venture backed startup?

A (119:30)

I would love to replace advertising with community donations with Memberful with Club Twit. It was my goal. In fact, when we first started Twit, I thought, we're not going to have advertising, we're going to take donations and we got 9,000 people. 9,000. It wasn't 9,000 people. $9,000 a month, which was a lot of money, but it wasn't enough money to do what we're doing. I mean, it was enough money for me alone. But we couldn't do multiple shows. So we eventually turned to advertising. I would love not to have to worry about advertising. That would be a boon. But it's not just like Wikipedia, only about less. In fact, Wikipedia has more higher percentage of donors than we do. Club Twit is less than 2% of our audience.

C (120:20)

People, people, people.

A (120:23)

No, I don't want to guilt people. I understand there's lots of reasons why, you know, people wouldn't join. If we got to 5%, we wouldn't need advertising. It would replace advertising. Right now, the 1.5% is 25% of our operating income. So if my math is correct, all we'd have to do is get six. Yeah, 6% of our audience to pay for it and we'd be sitting pretty.

C (120:49)

Six more percent. Pepper said he can get better shirts. Six more percent. We need 8%.

A (120:54)

8%, that's right. Yeah, we need 6 more percent or less than 8%. Anyway, it's a pipe dream. It's not going to happen. The other thing, but thank you, by the way, for all the Club Twit members. And if you do want to support us, I really appreciate it. Twit, tv, clubtwit. It'd be nice if you joined the club. There's lots of benefits. We did that great Dungeons and Dragons thing. You should be. Jeff, you would be amazed at how talented. I hope that at that fancy, fancy date you went to last night, you might have mentioned Kathira Long swallowed. Just a little bit.

B (121:26)

I didn't, but I shouldn't.

A (121:28)

I wasn't say anything about D and D. I don't want to really bring up the D and D part.

B (121:34)

Say DND comes up not frequently on.

A (121:38)

It'd be a good test, a good litmus test. How do you feel about Dungeons and Dragons?

C (121:44)

How nerdy are you? More risky. Speaking of which, we can go.

A (121:48)

Before we leave this, there's one other thing I want to say about the thing. This is the first time I can remember a New York Times opinion piece being entirely a video.

C (121:55)

Oh, no, they do that now.

B (121:57)

They do this.

A (121:57)

They've done this before with other People.

C (121:59)

Yeah. Adam Ellick produces these, and there's tons of them, and they're quite produced.

A (122:03)

Oh, so. So this is not produced by. I feel like Jack, who does really good videos.

C (122:08)

He does really good videos. But this is obviously collaborative with them.

B (122:10)

Yeah.

A (122:10)

You think he did okay?

D (122:11)

Okay.

C (122:11)

Yeah. Malik is. Is very good.

A (122:13)

Is that them trying to be major into video?

C (122:18)

They're going major. Well, they have.

B (122:20)

This is. This is different than their short form video endeavors.

A (122:24)

Well, I know my son was featured in the sandwich.

C (122:27)

Exactly.

A (122:27)

Video.

C (122:28)

Right. That's an example. Yeah.

A (122:29)

Yeah.

B (122:29)

Man, I gotta get one of those sandwiches.

C (122:31)

I do too. We gotta come to New York.

A (122:32)

So I should just fly out right now.

B (122:34)

Do it.

A (122:35)

You know, screw the show. I'll be right there.

B (122:37)

All right. See you tomorrow.

C (122:39)

Oh, it's a little cold, though. Paris. That's thick.

B (122:42)

Hey, maybe the line will be shorter.

A (122:45)

No, apparently it's not. I have a winter coat.

B (122:49)

A. A singular one.

A (122:52)

Yeah, I saved it from my old days. The old days. The old times.

C (122:56)

Can we go back to dating on line 127?

A (122:59)

We never left. Yes, we can.

C (123:03)

So we'll see what Paris thinks of this.

A (123:05)

AI Love.

B (123:07)

Actually, I never would have expected this.

A (123:10)

You know, this is written by.

C (123:12)

I know.

A (123:13)

Megan Maroney, our longtime host on this network.

C (123:18)

So Tinder is testing a new AI powered feature called Chemistry that uses deep learning scary quotes to analyze a user's photo in their camera room.

A (123:29)

Not just photos. Photos, photos.

C (123:32)

It understands their vibe, which means make better matches.

A (123:37)

This is a bald face bid to get access to your camera roll.

B (123:42)

No, no, no, no, no.

A (123:45)

Not on Kathera Long Swallow's watch.

B (123:47)

No, Hinge has a version of this. Well, I think I've mentioned this in the show. Hinge has uses AI in that. So it's a mix. Hinge is a dating app that is a mix of photos. You could also do videos as well as text prompts where it's like a question or an answer, some sort of thing where you're writing something and it uses AI now, where at the bottom of when you're writing your little prompt, it will have a little AI thing come up and like it or say that it's good or something like that. And I. There's no way to turn it off. And I find it very annoying because it does not understand humor clearly, but it does things. My prompts are good.

C (124:25)

It just doesn't understand.

B (124:26)

Maybe it's does. I mean, I clearly think smart are good, but I don't think its critiques are like the reason why it Thinks it's good.

C (124:32)

Is they think the Ryan Liz is a good writer.

B (124:35)

Yeah, probably.

A (124:37)

Hey, I thought he was a good writer.

B (124:38)

I liked it. I thought the noir, like that wasn't overwrought at all.

A (124:43)

Maybe I like overwrought stuff, you know, and we drew me in. It drew me in. Megan continues that once couples have made a match and decided to put a ring on it, 36% of people are actively using AI in their wedding planning. This is according to the Knot.

B (125:02)

What does that mean? Yeah, well, does that mean you Google things about wedding and you scroll past the Google AI summary? Does it mean you're scrolling on Pinterest and you happen to look at a bunch of images?

C (125:14)

Gemini 3 for a Dungeons and Dragons wedding.

A (125:17)

Oh, perfect. For somebody. Like.

C (125:24)

For those of you on audio, Paris just rolled her eyes.

A (125:26)

Help me.

C (125:28)

You could hear that.

B (125:30)

I was gonna say, I do think it came across in the side. I know how to play to the audio genre too.

C (125:38)

You are multi.

B (125:39)

Oh, I did. Okay. So one of my favorite D and D podcasts I'm sure I've talked about, and it's called Not Another D and D Podcast, they have this segment where in addition to playing D and D, they preside over a dungeon like a Supreme Court, litigating squabbles at different DND cases. And this is good. They also have a segment that's originated from a bit where they step into church and they are like people absolving people of their DND related crimes. And in DND church one day someone wrote in saying, oh, me and my now wife, big D and D fans, we had a DND themed wedding. And as part of the that as we were like doing our vows, the officiant had us each roll a D20. And, you know, D20 for context, it's a 20 sided dice. If you get a one, that's like really bad. If you get a 20, that's extraordinarily good. And they both rolled. I think his wife got a nat 20. Great. He got a nat one. But the officiant looked at both them and was like, wow, two Nat 20s. And announced it to the crowd.

A (126:44)

Oh, very smart.

B (126:46)

And so the writer was writing in being like, I think our marriage has been, you know, ceremonially built on a lie because we lied to the audience about the role.

A (126:58)

I didn't think of that. So I wrote to Gemini, help me plan a Dungeons and Dragons wedding for Paris and her new beau, RFK junior. This is certainly a unique pairing for a campaign. It says combining high society Flair. With a high level Druid ranger archetype. To plan a Dungeons and Dragons wedding for Paris and RFK Jr. We need to blend high fantasy nature motifs.

B (127:25)

I'm the city that whites of high society.

A (127:28)

Flair and absolute glamour. They're calling it the union of the two realms. The setting. The Druid's Grove meets High court. The location. An outdoor venue is mandatory. Think Redwood Sanctuary. A castle ruin overtaken by vines. Or a botanical garden. Bioluminescent lighting. Table names instead of numbers. Table names follow legendary locations. Baldur's Gate. Neverwinter. The Shire Falcon's Hollow. The centerpieces. Moss runners. Antlers. Terrariums. Dead bears. We need to assign classes to the bride and groom to guide their aesthetic. The groom, Robert. Class is circle of the Beastmaster Ranger.

B (128:13)

I like that it goes to Robert.

D (128:15)

It goes.

B (128:15)

And it's not Paris.

A (128:18)

You're the college of glamour bard. A high elf sorceress. I think it knows you.

B (128:23)

That's actually okay.

C (128:25)

Can it draw pictures of the bride and groom?

A (128:28)

I just want to.

B (128:28)

I am going to say that RFK Jr. Would be a. What was the second one?

C (128:34)

It said.

B (128:36)

It was be a Beastmaster Ranger. I do think Master Ranger. Yeah.

A (128:42)

You're. You're gown. By the way, if. If you're like. Wait a minute. I ask. Oh, yeah, wait a minute. Here. Here is a. Here's the picture of Paris and RFK Jr. At their wedding.

B (128:53)

Anthony is a falcon.

A (128:55)

He's got a falcon. I think they've tied your hands together in a kind of symbolic gesture.

B (129:01)

I have a purse that is shoved into my hip.

D (129:05)

Floating.

A (129:07)

Yeah. There's no strap for the purse.

B (129:08)

I think it's a. RK Jr. Has a sword.

A (129:12)

Look at. There is a net. A. A 20. Is that a 20?

C (129:16)

That's Paris Hilton, by the way. That's Paris Hilton.

B (129:18)

I was gonna say.

A (129:19)

Oh, it's Paris Hilton. Oh, it is. It's Paris Hilton. Okay. I should correct it, shouldn't I?

B (129:29)

No, it's okay. I like that we rolled this a 17 as well.

A (129:33)

It's a 17. Yeah.

B (129:34)

I mean, that'll be good. I guess that's pretty good.

C (129:37)

It could be okay, but you're.

A (129:39)

Let me. Let me tell you the vows. I promise to be your tank when you are drawing aggro, your healer when you are down, and your companion on every side quest until my hit points reach zero. I see. This is why people use AI to plan their weddings.

B (129:58)

So it does say roll for initiative at the start of the ceremony, the officiant or dungeon master and Ask them to roll a giant D20.

A (130:05)

There you go.

B (130:07)

That's cute.

A (130:07)

There you go. So I. I would bet. I don't know. And Megan doesn't make it clear that most people use it to write their vows, don't you think?

C (130:17)

The not future of AI report to which she links future marriage. Marriage. Right. Sorry, what did I say?

A (130:25)

AI I don't think the knot really knows much about the future of AI.

C (130:29)

They use it for writing support for invitations, vows, speeches, inspiration, design help for generating fun extras like games, trivia and hashtag ideas. Oh, saved me from the icebreakers. Understanding wedding norms, etiquette and traditions.

A (130:44)

I think this is great.

B (130:45)

I like where it says couples are using AI to draft separation agreements, to code legal jargon and help manage their post divorce Greek.

A (130:55)

It's Gen Z's new divorce crew. Divorce, divorce, Divorce coach.

B (131:00)

Diverse coach.

A (131:03)

I have a block. I can't say that word. Yeah, yeah. The global online data dating market's $10 billion. The wettest wedding industry though a hundred billion dollars just in the U.S. so.

C (131:15)

You know, I used to be. I used to somewhat oversee weddings at Conde Nast.

B (131:21)

Say more.

A (131:22)

Would you be the officious.

C (131:24)

No, no, I was. I was part of.

B (131:25)

We had a matchmaker.

C (131:27)

We had the wedding because we own brides. Modern Bride, An Elegant Bride and I. Which one were you helped build? It was all three together.

B (131:35)

Were you elegant?

C (131:36)

We thought we owned bridal.

A (131:38)

We had it before that.

C (131:40)

Then along came. We found the brides who. Who leave every 18 months and you hope they don't come back for their sake. And you have a whole entirely new audience. So it flushes out, so to speak. And so we found that we weren't getting the audience. Where was it going? Where was it going? There was this new thing called the not.

A (131:55)

Yep, yep, yep.

C (131:59)

And they don't own bridal anymore.

A (132:01)

Megan refers to a Reddit post where a guy says his wife to be left him at the altar when she realized he used chat GPT to write his vows. I read that.

C (132:14)

Too good to be true.

A (132:15)

But no, I. I think it sounds real. I do. I do. It sounds like. Exactly.

B (132:19)

I believe every. A hole post. I believe every Reddit post.

A (132:24)

Yeah, they're all is it from. Is it from. No, it's relationships.

B (132:28)

Probably relationships that. That seems. That's how you can tell that. Because if it wasn't on our relationships, which has stricter posting standards, it was on the more generally applicable our relationship advice.

A (132:42)

You know, we have to take advantage of. Of Paris's Reddit expertise a little yes, that's true. She apparently really knows her stuff when it comes to.

B (132:51)

I really do. Unfortunately, I. I'm deep. I'm deep in the forums.

C (132:57)

My wife is too.

A (132:58)

I love her. I check Reddit all. Every day.

C (133:00)

Yeah, that's where she gets lots of information, I bet.

A (133:03)

I follow a kind of. A different. It would be fun to compare our subscriptions.

B (133:10)

I'm just saying, right now it's being dedicated. Right now it's being dominated by Secret Lives of Mormon Wives because they just dropped into season.

A (133:16)

Mine's Emacs Cashios Lisp.

C (133:21)

Is there an r Nuzzy?

B (133:23)

No. I am interested to see what the journalism subreddit thinks of this.

A (133:27)

There's a foam wa.

B (133:30)

I don't like fomo.

A (133:32)

No. Give me a better one.

B (133:33)

Near Dumois. Well, no, I don't. What for gossip.

A (133:38)

Celebrity gossip.

B (133:40)

Oh, I'm not a celebrity gossip person.

A (133:42)

Oh, you're not? Oh, okay.

B (133:43)

No, I just enjoy consuming some. Whatever it's called.

A (133:48)

Here's a little something I've just discovered on Reddit. Two 20,000 Epstein files in a single text file to download for your notebook lm.

B (133:59)

Oh, that's fun. I like that.

A (134:01)

Yeah.

C (134:01)

Yeah.

A (134:02)

So you could put it in and then query. And you know what? Journalists have been complaining about this email dump, that it's so big it's.

B (134:10)

I was gonna say, I haven't really been able to go through it that much because it's very, very difficult to navigate on your phone.

C (134:16)

Google put out a database tool for journalists and one site, I can't remember the name of it right now. Put it up with that so you can search it.

A (134:23)

Well, here is from Hugging Face, the tenso not contributed data set Epstein files 20K, feed it to Notebook LLM or your favorite LM. And you know, you could just ask questions.

C (134:37)

If you go to journaliststudio.google.com you will find the pinpoint. You can search there.

A (134:48)

A collection of tools to empower journalists to do their work more efficiently.

C (134:52)

Oh, there's a. There's a slash shoot. Okay, hold on. I'll put it in the chat.

A (134:56)

Oh, this is not it.

B (134:58)

I will give you a little peek into my Reddit world, which is have you, Leo, have you come across the men who's cutting chives almost every day until a subreddit says they're perfect?

A (135:12)

Oh, my God, no. What subreddit would that be?

B (135:16)

Kitchen Confidential. It's a user called Flexican, but the L. And Flexican is number one, and he is, I don't know, 43 days ago he started.

C (135:29)

Who defines the perfect chive?

B (135:31)

Reddit. Reddit defines the perfect course.

A (135:34)

You fool.

B (135:35)

You fool.

A (135:36)

Here's day 44's effort.

B (135:38)

On first glance, it may look perfect.

A (135:40)

Yeah.

B (135:41)

But if you go to the comments, people will find one chive that is not. Somebody just says of that post. Somebody just says, no, no, that's him replying to.

A (135:56)

Oh, the comments. Oh, go to the comments.

B (135:58)

You've got to click on that.

C (135:59)

His fingers must.

B (136:01)

No, no, no, no. Oh, no. Leo, go back. Go back to where you were.

A (136:04)

Yeah. Oh, it's his comments.

B (136:07)

You're in his user history. Click on that post. Click the title. No, no.

A (136:11)

And then see the comments in the title. I see what you're saying there.

B (136:14)

Click that.

A (136:15)

Wrestle girl.

B (136:16)

That's the Mod. They're getting 2.4 million. So part of the thing is the people will now zoom in and find if any of the chives are miscut us.

A (136:30)

Well, let's help him. Let's see, let's see. Can we see? Oh, no, this is a mess.

B (136:37)

I know.

A (136:39)

This guy's going to have carpal tunnel syndrome if this keeps up.

B (136:42)

He's already had to go down from cutting a cup of chives a day to just a couple chives a day.

A (136:48)

What is he doing with all these chives after he cuts them?

B (136:52)

Frankly, that's part of the reason he's like, I'm wasting too many chives.

C (136:56)

His fingers stink.

A (136:58)

He says he is a chive lord. Okay, this is why we go on.

C (137:03)

This is harmless, though. This is nice.

A (137:05)

This is great. This is great. I am going to follow Kitchen Confidential. That sounds like a good point. Definitely a subreddit I should follow. This is the problem with Reddit. You get ideas for other ones you should follow. I pretty much stick to geek subjects.

B (137:20)

But I have follow so many cat subreddits. A large amount of it is cat subreddits.

C (137:25)

Would you like me to give you one of my. One of my things on line 196?

A (137:30)

Well, let me do a commercial. Commercial mention. We are already 2 hours and 21 minutes into this program and I think if we're gonna get out of here before dawn, I should probably do another commercial. You're watching Intelligent Machines and we will do the picks. Benina, where am I? Where. How do I stand? Do I have to do another commercial still?

C (137:51)

No, this is the last one. And we also have HTG live chat after this at 5.

A (137:57)

So we want to wrap this up because Scott Wilkinson and home theater geeks are going to do A live chat at 5pm Pacific.

C (138:04)

Yeah, that's when they're scheduled.

A (138:06)

Okay, so we've got a little bit more time.

C (138:08)

What does a schedule mean, Benito? They never hold to a schedule. How quaint.

A (138:14)

How quaint. Our show today, brought to you by Spaceship. Now, if you're looking for somewhere to register your domain domains, I've got the place for you. Spaceship.comTwit We've been talking a lot about Spaceship. They are a modern. Well, I don't want to just call them a domain registrar. Everything you need for your Internet presence. They just passed a major milestone. Over 5 million domains under management. And there's a good reason how a company that just started can get to that size so quickly. Not by chance. It's because Spaceship delivers real quality features that make sense not just for domains, but everything that helps you build and run your online presence. That means hosting of course, but also they've got a great business email solution. They've got tools for creating and managing web apps, VPSs. They even have a fantastic end to end encrypted messaging system which I wanted to take off. I want everybody to use it. It's so great. Great. Instead of having your phone number be your identity or you know, LeoLaPort24 or ParisMartNo me or whatever, you can have a domain name. So mine is Leos im and here's the thing. I bought the domain name from spaceship. $5 a year for Leos IM and I get the messaging built in. Thunderbolt's fantastic. So if you wanted to message me, you could do it@Leos IM. That's my domain name and my messaging. Now this is great for businesses because you can have your business domain be your messaging address. Another great reason people are switching to Spaceship pricing. Basically it's Black Friday and Cyber Monday level prices all year round. You don't have to wait for a great deal. Like I said I got for 4 bucks I got a IM domain right now. You even get more as a twit listener, exclusive offers that make it even better. So whether you're planning a new online project or moving an existing one, by the way, great to move domains to Spaceship. You are going to love this process. Spaceship has what you need to get it launched, connected, running smoothly, plus more affordably too. I love these guys. And what a modern, good looking site. Spaceship.com check it out to see the exclusive offers and find out why millions have already made the move. That's Spaceship.comTwit. it's Black Friday every day at Spaceship.comTwit. you'd be crazy to go anywhere else, honestly. All right. Picks of the week. Jeff, you volunteered.

C (141:04)

Well, I was one. I was just for parasect actually, but I'll have others. But cat island on 196 we've heard of.

B (141:11)

Sounds like my kind of island.

C (141:13)

That's why I did it.

A (141:14)

Are you are. I mean, you've got one cat. I don't know if that qualifies you really as a cat lover with one.

C (141:19)

So this is. This is in Japan. It's an island that has tons of cats. A guy sits down on a bench in Cat island and see what happens.

A (141:28)

Dream come true.

C (141:29)

Are you. Are you. Are you showing it?

A (141:32)

I will, yeah, if you want. Here we go on Cat Island. You're live lap will be filled. Why I don't go. What's.

C (141:39)

Why is it doing that? Leo, you can't do anything right.

A (141:43)

Here come the cats.

C (141:44)

There we go.

A (141:46)

Here come the cats.

C (141:48)

Just wait.

A (141:50)

Japan's amazing. There's one town where deer run wild and. And they come right up.

B (141:55)

There's also a place where deer run wild. It's called Staten Island.

C (141:59)

It's called my Backyard.

B (142:02)

They come from New Jersey.

A (142:03)

They swim across the marina.

C (142:05)

Look at this. Look at this.

A (142:08)

I did go to a cat cafe in. In Japan. In fact, I went to a hedgehog cafe as well and an owl cafe.

B (142:16)

I want to show you guys.

A (142:17)

Oh my God. Is this your dream come true, Paris?

B (142:21)

It is. But I feel Gizmo would be so confused when I came home smelling of so many different cats. Cats. Oh, look at. Listen to their little meows. I went to a rubber stamp shop this weekend and there was a cat in the window and we had a beautiful moment.

A (142:37)

You know wh is in a lot of trouble in San Francisco because a waymo apparently killed a bodega cat. Huge stories and there's huge story. People are pissed. I know. Look how shocked Paris. Look.

B (142:54)

I was shocked. I was shocked. I'm appalled. I don't want to hear any more about it because I'm sad.

A (143:00)

Yeah.

C (143:01)

Then I think the wire cutter gives me inspiration to recommend a road trip to Paris and her pals at 199.

B (143:07)

Hey, I'm trying to think of locations for a two week road trip in January. February.

C (143:12)

This is pretty cool. This is pretty cool.

A (143:14)

We bought a 450 pound mystery palette packed with returns goods. Oh, this is. I've heard about this. Return goods from Amazon. They don't. They just sell them in bulk because.

B (143:26)

Guys, should we go in on one.

A (143:28)

Let's do it. Who gets it?

B (143:30)

That could be us, you, you, you or Jeff. It's going to be delivered to your house and then I.

A (143:35)

You got no room in your apartment?

B (143:37)

Yeah. No.

A (143:37)

So Anna Marie Conti at wirecutters spent 700 bucks, well, her bosses did, on a 450 pound six foot tall cardboard box filled with hundreds of mystery products that have been returned to Amazon and other ret retailers. Anyone can buy them and you can.

C (143:57)

Go to the warehouse that's filled with them. That's what's fascinating. That's the road trip part.

A (144:02)

She was hoping for, I don't know, a diamond ring, a cozy custom blanket, some sexy unworn lingerie. Instead, what she got was a whole heap of sadness.

B (144:14)

Unfortunate.

A (144:16)

Yeah. And because it's cheaper for them to discard this stuff stuff than for them to repackage it and resell it.

B (144:27)

Yeah. We could get a palette and then if people sign up for the club, they get one. They get their fraction of the content from the palette. Depending on how many people sign up.

A (144:37)

We could have daily random drawings. I could put the box here and I could say, and now this one's going to Butch Henderson in our club, Club Twit Butcher. Let's see what you get from the heap of sadness. And oh look, it's a thousand piece puzzle. Missing a piece.

C (144:58)

Well, they also have, they have other things that are labeled, you can know. But they didn't. So they took it out and it's all these plus sized garments that are all kind of ugly. And that's all it is. And at the end, how much plastic? How many pounds of plastic? How many pounds, pounds of cardboard. There they are looking at. Some of these are open and from various retailers, Walmart, Home Depot and such.

A (145:25)

Butch has turned down his thousand piece puzzle. Missing a piece. He says no, I think this is a great promotion for Club Twit. I wonder if Lisa will let me do this.

B (145:37)

Hey, wait.

C (145:37)

Actually, I've got a great idea downstairs.

B (145:40)

You could get enough power that it could cover the hole in your house and you could kind of like build another wall. But out of pallets of. Amazon returns.

A (145:50)

Amazon returns. She calls it the Polyester Mountain. £450, 582 items. Oh God, she looks depressed.

B (146:06)

So much excitement quickly shifted. Distress.

A (146:10)

This is my fear is that we would order this and this is what we would get instead of a video.

C (146:15)

Of her going through this. If you go down.

A (146:17)

Oh God.

C (146:18)

It's pretty funny. It's a good story idea.

A (146:20)

Yeah, it's A great idea, but it's.

C (146:22)

Also, it's a good kind of consumer report.

A (146:24)

What's sad is that these are still in the packages. Right? You have to, you have to open them. Yeah, yeah.

C (146:30)

Here she goes.

A (146:30)

Yeah, I can see why that was returned. Somebody did not want their cargo shirts and matching pants. Health bathing suit. A fan. See if I want a fan hat. Yeah, I like that idea.

C (146:47)

Some of them.

A (146:50)

Do you have to go to the warehouse to get this or will they.

C (146:53)

I'm not sure they did because it's a better part of the story. Yeah, yeah, there's, there's, there's, there's, there's tiktoks about this and social medias about.

A (147:02)

She's very funny, I have to say. Say well done. She's making this very good see through pair of sweatpants. That's good.

B (147:10)

Yeah, just what you need.

A (147:11)

Yeah, that's what they said on Landman last night. Ali Larder talking to her college age daughter going to be a freshman at TCU football player, goes by and Ali Larder says, I love gray sweats. It tells you, it's like. It tells you right on the box what you're going to get.

C (147:34)

So finally there is also a. Google has put out a font maker, Gen Type. So you can go to Gen Type and you can make a font out of anything you choose.

A (147:48)

Oh, we've, we've done this before. This has been around for a while.

C (147:51)

Gen Type has.

A (147:52)

I think so. Oh, wait, but that was. Yeah, because I. Yeah, they made it out of ice cream and stuff.

C (147:58)

Elon Musk.

B (147:58)

Musk.

C (147:58)

It was the Elon Musk one that you're thinking about?

A (148:01)

No. Was it? All right, well, let's. No. Yeah, we did this before. This is around for a while. Yeah, that's okay. It's pretty cool. I mean, it's worth mentioning again. So what should I. Yeah, see, I did this one. Honeycomb pink background.

C (148:18)

How about a font of cats?

A (148:22)

Cat font. Is that enough? Do I have to say more?

C (148:26)

See what it generates.

A (148:27)

Yeah, there's an Alphabet of cats. This would be good for you, Paris. You could put this on your website along with your plaid suit and bolo tie. Yeah, just have a bunch of cat in your lap. Or no, all the headlines could be. Oh, look, they're doing it. They're making the letters. Oh, and when they say cats, they're using many, many cats.

C (148:51)

That's cheating.

B (148:52)

I don't.

C (148:53)

I wanted to see a. You. Yeah, I wanted to see.

B (148:54)

Oh, there you've got the L is made of One cat, for some reason.

A (148:58)

Just L. Oh, okay. That's cute. That's how it should all be.

C (149:02)

Yeah.

B (149:02)

Yeah.

A (149:05)

That's pretty good, though. I mean, come on.

C (149:08)

Yeah, you.

A (149:08)

You guys are so spoiled now. There's a good one. There's a U with.

B (149:13)

Oh, Y. I do like that.

A (149:15)

Yeah, that's kind of cool. That's a Q using the tail of the cat as the curly Q.

C (149:21)

That's smart. Yeah.

A (149:22)

Yeah, that's good. Yet, you know, I need to give it more, I think, more input to make it more interesting.

B (149:28)

Yeah, just say one big cat would be the description of it.

A (149:35)

Let's make this. Hello, Paris. I am.

B (149:42)

It looks upsetting. Honestly.

A (149:44)

It does. It's a little disturbing.

B (149:46)

I don't like that the cats are being made to do this.

D (149:49)

Okay.

B (149:50)

Yeah.

A (149:51)

Okay. Never mind. That's labs, Google Labs. Labs, Google, gen type.

C (149:55)

That's my end.

A (149:56)

Yes, that's the end.

B (149:59)

My pick this week is a brief descent into Nick Vember, which is.

A (150:03)

What have you been watching?

B (150:05)

I on Thursday, went to the premiere, I guess, day before the premiere of the Carpenter's Son.

A (150:13)

This is brand new. Oh, my God.

C (150:19)

Audio is.

A (150:20)

Nick is. Wait a minute. Are you telling me Nick Cage as Jesus?

B (150:23)

No, Nick Cage is Jesus's father, some might say. Paris, isn't Jesus's father God? I say that's the question that in this thing is, is his son Jesus? Is he his son at all or is he something demonic? It's a horror movie. No, based on the Infancy Gospel. I think that's what it's called. The Infancy Gospel of Thomas. Thomas about teenage Jesus. And Nick Cage is Joseph and FKA Twigs is Mary. And it. Listen, I won't say it's good, but it's not a bad movie.

A (151:00)

You saw it in a theater?

B (151:01)

Certainly I saw it in the theater. I not only saw it in theaters, I also saw a Q and A there with the actor and FKA Twigs.

A (151:11)

God, FKA Twigs.

B (151:13)

I wish there was popcorn.

A (151:16)

One and a half stars from Roger.

B (151:18)

It was in the Yafe Theater at Angelica at the Village east, which is a beautiful theater.

A (151:25)

You're taking bring this up a little too seriously. You're actually going to premieres now.

B (151:29)

Well, I know.

C (151:30)

What are you gonna do when it happens in middle of now?

B (151:32)

I mean, in the middle of nickvember.

D (151:34)

You gotta go.

B (151:35)

I honestly didn't realize that I was signing up to go to one with a Q A until after I'd kind of looked at my ticket. I just kind of booked it while I was doing something else. But I did record the one question that was asked about Nick Cage during the Q and A. FK Twigs was asked about her experience working with him, and she says he really. I can't do that. She does like a. Has like a British accent. He really turns up to the scene ready to go. There's no warming up. You know, you say action, and he's already there, like, fully embracing the moment. So I think working with him, I learned about the importance of, like, winding myself up as an actor. The part of this I really liked is he has such a huge presence, which initially I was kind of nervous about, to be honest. To be with someone like that, that. That just takes up the whole scene, really. He's also. I mean, you can feel him before he comes on set. He's 10 minutes away in the car, and you're kind of aware, which I just. I love that as a visual being. Like, Nicholas Cage is arriving, everyone.

A (152:35)

I feel like he launches into the scene because he's just anxious to get to lunch, but, you know, maybe not.

B (152:40)

No. Nick Cage is a true artist. Artist.

A (152:42)

He is an artist. Now, what's been your favorite so far? Cuz you. You have. There's so many Nick Cage movies. You only watch 30 last year. Can you. Are you going to watch?

B (152:51)

I saw Moonstruck in theaters as well on Friday, and that was. I mean, that's just such a phenomenal movie. It's fantastic to see it in the. With wooden hands. Yeah. But it's. It's just like the. It's like a perfect drama against the wall.

A (153:05)

It's so Nick Cage. That. That role is so perfect.

B (153:08)

Him and Cher, they just have the best chemistry in it.

A (153:11)

Cher's more naturalistic.

B (153:13)

He's like, the most. Yeah, but like, he kind of has to be. There's no other way you can play that crazy. His character is written in an absolutely insane way. And I think. I know. It was just great. It was a great. I saw it at the Metrograph. It was lovely.

A (153:27)

Let me ask you about your Gen X, Gen Z, Whatever.

B (153:30)

What are.

A (153:30)

What are you? Gen Z? Millennial.

B (153:32)

Gen Z. Millennial. Cusp.

A (153:34)

Why do you. Why do you spell Jesus or teen Jesus as teen? Just. Is that a Gen Z thing?

B (153:42)

It's like a joke that I don't entirely understand, which I think there's like a righteous gemstone character, and they call him Genesis Genius, who's like. Who's teenage Jesus? Jesus. Because that's what this is about. And whenever this movie was announced, everyone was joking about how this guy's playing Teen Just Teen just. And I do think that that's very funny. So it's always referred to it as teen just Neologism.

A (154:06)

Okay.

B (154:07)

Yes.

A (154:09)

Well, there you go. The Carpenter's Son in movie theaters near you. This is why, by the way, it.

B (154:15)

Was inspired by the cinematography, was inspired by the Exorcist. And I will say that is a correct description of it. No, but there's some other very interesting gross things. Beautiful, beautifully shot. A gorgeous film.

C (154:33)

Can I give you one more real quick? 134.

A (154:35)

Yeah.

C (154:37)

This is a paper, a Pakistani paper.

A (154:39)

Oh, we almost got out of this show without any archive dot org.

C (154:43)

No, it's not archive. No, it's not that. It's not that. It's not that that papers are below that. This is the other AI Okay, I just did.

A (154:52)

Oh, I saw this. I did see this.

B (154:54)

Oh yes, I did see that.

C (154:55)

At the bottom of an article it.

B (154:57)

Says, if you want, I can also create a Leia. You've got to keep it on.

A (155:02)

If you want, I can also create an even snappier front page style version with a punchy one line stats and a bold infographic ready layout out, perfect for maximum reader impact. Do you want me to do that? Next, after the entire article, which apparently. Mistakenly left in his chat GPT, I.

B (155:24)

Think I recall seeing somewhere that it was placed in there. They're saying because of an editing error, but still. Absolutely.

A (155:30)

Well, that's the editing error.

C (155:32)

The editor's fault, not the writer fault.

B (155:34)

Yeah, no, I was to say specifically I think they were describing it was the editor's fault rather than the writer. But I don't know how true that is.

A (155:40)

Okay, what's. What's the order of operations there? He. He wrote an actual article.

C (155:47)

Maybe, maybe the editor used AI to edit it and then it.

B (155:51)

That's what I'm saying. We don't know whether it was the writer.

C (155:54)

Maybe it wasn't even written in English, you know.

A (155:58)

Yeah, maybe it was like an article the AI would generate because it's a financial story that's written really all fact based. It's not, you know, it's. It's exactly the kind of thing AI does. Right. Pakistan's automotive market showed strong growth in October with sales of cars, vans, pickups and sport utility vehicles rising to 17,333 units. I mean, that's not, you know, nobody. That's his lead.

B (156:21)

Yeah, I mean that is.

C (156:23)

That's a press release.

A (156:24)

It needs bamboo. I'm just saying there's nothing that wouldn't be Improved with a little bamboo symbol. I'm just saying. Read, read the telos.com article. Is it by Ryan Lizza? And you tell me, ladies and gentlemen, put it in the review.

B (156:43)

It is not telos.com.

A (156:45)

Oh, what is it? Sorry, Telos News News. Sorry, telos news, part one, how I found out November 17th. Ryan Lizzo, read the article and then put your. Your thoughts about it in a review of this show on Apple's itunes podcast.

B (157:03)

Specifically, your thoughts on the bamboo. Your thoughts on anything related to the nuzzy Gate, ideally. And when part two comes out, give me your thoughts on that as well.

C (157:12)

It's best if you. If you want Paris's heart, you've got to. In the review, you have to beg for more. Can't wait for part two. Paris, thank you so much for bringing this to us. I'm telling you, there is part two.

D (157:22)

Two.

A (157:22)

It's. It's been two days.

B (157:24)

Part three is my question.

A (157:25)

He's.

B (157:26)

He's letting us, you know, really figure it out.

D (157:29)

Yeah.

A (157:29)

Did you see Keith Oberman's response?

C (157:31)

Oh, that's pretty funny.

B (157:32)

Yeah, I sent that in the chat. Yeah.

A (157:34)

Oh, that's why. Yeah, I know I saw it from somewhere.

C (157:36)

Okay, now you got to tell the audience the response.

A (157:39)

Well, so Keith together mentioned in the article because apparently he and Olivia ne had a relationship for four years when.

C (157:47)

She was in college.

A (157:48)

When she was in college. And Ryan, Liz is saying, and he gave her, you know, Hermes scarves and a $15,000 Cartier watch.

B (157:55)

And what a jerk for $15,000 worth of Cartier jewelry.

A (157:59)

And what a jerk he is for. For paying for her rent and stuff. And I think Olbermann's response was good, by the way, about this from Liz's reply. Olivia and I lived together for four plus years. That's four birthdays, four Christmases, four anniversaries. That's like $1,250 of jewelry per celebration. And her apartment was a writing studio. And I made an F ton. Then what was I supposed to do? Get her a lot of gift certificates from Kmart?

B (158:28)

I mean, she was.

C (158:29)

She's not a Kmart lady. Lady. No, no, no.

A (158:31)

How much was Keith Olberman? 34 years.

C (158:36)

34 years.

B (158:37)

34 years older. Yeah.

A (158:38)

How much older? I think Everybody, everybody was 34 years older than her.

B (158:43)

Well, 20. Within the 20 to 30 something year older range. Yes. I think what you're supposed to do, Keith, is not live with a college.

A (158:54)

Student when you're 40s or 50s or 50s. I think that's probably the case.

B (159:00)

Yeah.

C (159:01)

Yeah.

A (159:01)

Ladies and gentlemen, on that sordid note, I think we should wrap this thing up.

B (159:07)

Bamboo.

A (159:11)

Boo.

B (159:12)

Oh, I said bamboo.

A (159:13)

Bamboo. Now, remember, our next show will be the day before Thanksgiving. Does that mean, Paris, that you will be in Florida for this show?

B (159:21)

I will be, yes. Oh, good.

A (159:22)

And you get your appearance from Pop with his special gum.

B (159:27)

I'm sure it's still sitting at the desk back there, so we'll bring it out.

C (159:33)

I want.

A (159:34)

I want. I want so much, much. And will Pop be deep Flash frying the turkey again this year?

B (159:40)

Oh, yes. He's always frying the turkey.

C (159:43)

You do a video chronicle of it?

B (159:45)

He publishes a video chronicle, but I'll. I'll do my own. In addition to that.

A (159:50)

People don't know this, but Paris's parents, and we won't say the names, but are very, very famous movie stars.

B (159:56)

And they're very famous.

A (159:57)

We try to preserve their anonymity. And her.

B (160:00)

Yeah, we try to keep it because she.

A (160:01)

She's not a Nepo baby. She doesn't want to ride in their coattails.

C (160:05)

Her name is actually Kalamazoo Martin, but, yeah, Paris Coppola.

A (160:13)

That's Paris Martineau. At least that's the name she's using today. Investigative reporter at Consumer Reports. Paris NYC is the website. Got to go there to see the fabulous, professionally photographed images. Her headshot.

B (160:28)

There's a lot going on there.

A (160:29)

Yeah, there's a real, real pro. Hey, I see you've set the table for. Are you having a little bit of a. A big dinner party tonight?

B (160:38)

I love that. It looks like I've set the table because the table is covered in.

A (160:42)

It's a mess.

B (160:43)

It's a mess.

A (160:44)

It's a mess. Did you get it from Amazon's warehouse? Okay. Thank you, Paris. Such a pleasure. Always enjoy, Joy. Same to you, my brother from another mother, Mr. Jeff Jarvis. His books, the Gutenberg Parenthesis and magazine are available now. And of course, he's a professor. Esteemed professor at Montclair State University in SUNY Stony Brook. And we will be back here next Wednesday, the day before Thanksgiving, right after windows weekly, 2pm Pacific, 5pm Eastern. Oh, you know what I like to do lately is check to see who. Our interview.

C (161:24)

Oh, that's right.

A (161:24)

We need to do that next week. We were. I was thrilled to have Jimmy Wales on. Thank you. Yes, Jimmy, it was really good to have you on. We will be talking. Oh, this is interesting. This is going to be very interesting. The guy who put. Who started Stable Diffusion, Ahmad Mosti Mostak. Will be joining us. Founder of intelligent Internet, iit.ink.

B (161:49)

I will not be there for the interview because we'll be recording it at a different time.

A (161:53)

Oh, and you can't make the recording. Will you be here, Jeff?

C (161:57)

Yes, I will.

A (161:57)

Okay, Jeff.

C (161:59)

Happening tomorrow. This is happening tomorrow.

A (162:01)

It's happening tomorrow. That's right.

C (162:05)

Is that your holiday party, Paris?

B (162:07)

No, that's on a Wednesday. I, I just can't do during the work hours. During work hours.

A (162:13)

You, you do more than enough for us, Paris. More than enough. Thank you very much for your appearance in the Dungeons and Dragons, which is now available on the Twit plus feed for club members.

C (162:24)

Say it again. What's that name? You have Paris there? What's it, what's that name?

B (162:27)

Thera Long Swallow. Catherine Swallow. We kind of, we went back and forth.

A (162:32)

I call her Catheter because I'm personal, close, personal friends.

D (162:36)

Yeah.

B (162:37)

I mean, that's only for close, close, personal friends who've fought against a pumpkin headed scarecrow. Yeah.

A (162:43)

Oh, we, we did good too, man. Tasha's hideous laugh. I want a little credit for, for putting that down. And prone.

B (162:51)

Tasha City's laughter is a great one.

A (162:53)

Was a very nice.

B (162:54)

You can get an unwise enemy. You can get him prone.

A (162:56)

And I apologize for the sound effects, which apparently were incredibly annoying.

B (163:00)

They were so annoying. I can't actually.

A (163:06)

I know, I, I, I, I realized it's like it.

B (163:09)

Imagine trying to record a podcast while a child is scrolling through Tik Toks. But they're trying to, they're kind of relevant to the discussion and playing them on full volume, that's the, that's the vibe for a little bit. And we're all very politely being like.

A (163:25)

We were in uncharted territory. The fingers you have choose to dial are too fat.

B (163:31)

To obtain a special dialing wand, please mash the keypad with your palm now.

A (163:36)

Thank you, everybody. We will see you next week on Intelligent Machines.

B (163:41)

I'm not a human being. Not into this animal scene. I'm an intelligent machine.