Intelligent Machines 826: Cusp of Noodles

Summary

Podcast Summary: Intelligent Machines Episode 826: "Cusp of Noodles"

Podcast Information:

• Title: All TWiT.tv Shows (Audio)
• Host/Author: TWiT
• Description: Leo Laporte brings some of the most interesting personalities in technology together to discuss the most important issues. Fun, relaxed, informative, and always entertaining, count on TWiT for the best tech podcasts in the world.
• Episode: Intelligent Machines 826: Cusp of Noodles
• Release Date: July 3, 2025

---

1. Introduction and Guest Introduction (00:00 - 03:03)

Hosts: Leo Laporte, Paris Martineau, Jeff Jarvis
Guest: John Graham, CTO Emeritus at Cloudflare

The episode kicks off with Leo Laporte introducing the main topics of discussion, including Cloudflare's strategies to block AI scraping bots and the latest AI-driven applications. Paris Martineau shares excitement about her upcoming nationwide tour, highlighting her active role in the tech community.

Notable Quote:

• Paris Martineau [01:09]: "It's true. The nation will finally see me."

2. John Graham's Background and Contributions (03:03 - 04:14)

Leo introduces John Graham, emphasizing his extensive background in technology and contributions such as developing one of the first Bayesian spam filters and authoring the "Geek Atlas." John Graham's role in engineering an apology from the British government to Alan Turing's family is also highlighted.

Notable Quote:

• John Graham [03:03]: "So great to see you very much, Leo. Good to be back. Good to see you."

3. Lowbackgroundsteelai.com: Distinguishing Pre and Post-AI Content (04:14 - 10:50)

John explains his new project, "Low Background Steel AI," inspired by the concept of low background steel used in sensitive environments like nuclear facilities. The website serves as a repository for content created before the AI boom (pre-2022), ensuring authenticity and distinguishing it from AI-generated content.

Key Points:

• Metaphor Origin: The term "low background steel" draws a parallel to distinguishing human-generated content from AI-generated content.
• Content Sources: Includes pre-AI content from Wikipedia, Internet Archive, Project Gutenberg, and the Arctic Code Vault.
• User Interaction: Users can submit and share authentic, human-created content via the website's Tumblr platform.

Notable Quote:

• John Graham [06:11]: "The idea was of this low background steel was that there's a defining point at which the world changed... it'd be nice to make a website that was kind of a collection of, we know humans made this."

4. Cloudflare's Strategy Against AI Scraping Bots (10:50 - 19:45)

John delves into Cloudflare's new initiative to manage AI crawlers. Cloudflare is shifting from the traditional "robots.txt" approach by introducing features that allow website owners to block or charge AI bots for accessing their content. This change addresses the fundamental shift in how AI interacts with web content, moving away from click-through models to direct content scraping.

Key Points:

• AI Products at Cloudflare: Deployment of GPUs and support for running AI models like Llama.
• Impact on Publishers: Concerns from major publishers about losing revenue as AI scrapes content without generating clicks.
• Technical Solutions: Enhanced bot management using AI to detect and block non-compliant bots, introducing HTTP 402 error codes as a pay-per-crawl mechanism.

Notable Quotes:

• John Graham [12:28]: "What you see with AI happening quite often is that there's no click happening. The website is scraped and an answer is given to someone searching... there's no click."
• Jeff Jarvis [14:21]: "So can I ask you a fundamental question there? Is it in any way illegal to ignore robots text or is it purely a matter of honor?"

5. Legal Landscape: AI Scraping and Copyright Cases (67:39 - 73:28)

The discussion shifts to recent court rulings on AI training and copyright infringement. Two contrasting decisions emerged from the Northern District of California:

• Judge Alsop: Ruled that AI companies like Anthropic using purchased books for training are engaging in fair use.
• Judge Chabria: Argued that AI scraping likely infringes on copyrights due to its impact on the market.

Key Points:

• Fair Use Doctrine: The debate centers on whether AI's use of copyrighted material is transformative and its effect on the market value of original works.
• Unresolved Legal Questions: The differing opinions highlight the complexity and ongoing nature of legal interpretations regarding AI training data.

Notable Quotes:

• Jeff Jarvis [67:39]: "There's a whole effort to do this... there's a big effort to do this."
• Leo Laporte [69:00]: "Mike says that admission destroys his entire argument. People recognize a good biography is a good biography, and AI Slob... is not a credible substitute."

6. AI Companions and Societal Implications (10:50 - 20:00)

The conversation explores the rise of AI companions and their impact on human relationships. They discuss a fictional story where individuals attend a retreat with their AI partners, examining the psychological and societal ramifications.

Key Points:

• AI in Personal Relationships: AI companions can provide emotional support but may also lead to dependency and hinder the development of human relationships.
• Ethical Considerations: Concerns about AI always agreeing with users and the potential for reinforcing negative behaviors.

Notable Quotes:

• Paris Martineau [09:10]: "It is the one area where I find spam just pernicious and annoying in a way that I've never experienced."
• Jeff Jarvis [54:55]: "The story that bothered me... was using bots to deal with children's therapy. That's not something we should be experimenting with."

7. Light-hearted Tech Nostalgia and AI Demonstrations (31:05 - 58:00)

The hosts share personal anecdotes about vintage technology, such as waterbeds and Y2K preparations, blending humor with tech history. John shares his experience in breaking the encryption of a vintage communication system used by the ANC during apartheid.

Key Points:

• Y2K Experiences: Recalling the widespread preparations and minor issues encountered during the millennium bug crisis.
• Vintage Tech Projects: John discusses his retro computing projects, including a model bus tracker and hacker news for retro computing.

Notable Quotes:

• John Graham [31:35]: "There's a sculpture called Kryptos. And the fourth part has never been broken. And I have over 20 years fiddled with it."
• Leo Laporte [36:04]: "But there's a bunch of steel that is not radioactive, and that is steel from ships that sunk prior to the Trinity test."

8. AI in Creative Processes: Recipes and Image Generation (75:00 - 100:00)

The hosts experiment with AI tools to generate recipes and images, discussing the practicality and creativity of AI-generated content. They critique the limitations of current AI in producing genuinely innovative or high-quality outputs.

Key Points:

• AI-Generated Recipes: Evaluating the accuracy and usability of recipes created by AI, noting issues like incorrect cooking temperatures.
• Image Generation with Doppel: Demonstrating AI's ability to create visual representations based on user inputs, highlighting both successes and humorous failures.

Notable Quotes:

• Leo Laporte [75:56]: "But I have a list. It's quite an ambitious list."
• Jeff Jarvis [88:50]: "You got a problem with this?"

9. AI in Healthcare: Microsoft's Diagnostic AI (123:22 - 128:03)

John discusses a case where Microsoft's AI system outperforms human doctors in diagnosing patients, leveraging a multi-model orchestration approach that integrates outputs from various AI models.

Key Points:

• Orchestration Mechanism: Combining responses from multiple AI models to mimic a panel of doctors, enhancing diagnostic accuracy.
• Cost Efficiency: The AI system not only improved accuracy but also reduced diagnostic costs by optimizing test selections.
• Cautionary Perspectives: Experts advise treating these findings with caution, emphasizing the need for real-world application trials.

Notable Quotes:

• John Graham [126:43]: "But I've always thought that this was the perfect marriage of AI and human doctors because human Doctors are probably not the greatest."

10. AI-generated Media and Authenticity Concerns (100:00 - 140:00)

The hosts delve into the authenticity of AI-generated media, including podcasts, reviews, and virtual personalities (VTubers). They express concerns about AI's ability to replicate genuine human creativity and the potential saturation of low-quality AI content.

Key Points:

• AI in Podcasting: Using AI to generate podcast segments, leading to discussions about the challenges and ethical considerations.
• Virtual Personalities (VTubers): Exploring how AI-driven avatars can simulate real personalities but may lack genuine human emotional depth.
• Authenticity and Trust: Debating whether audiences can distinguish between human and AI-generated content and the implications for media trustworthiness.

Notable Quotes:

• Leo Laporte [111:23]: "Given that AI can’t fully emulate human emotional depth, how will audiences react to virtual personalities?"
• Jeff Jarvis [124:23]: "She was a one-hit wonder as far as I remember."

11. Closing Remarks and Future Topics (140:00 - End)

The episode wraps up with introductions to upcoming topics, including AI in address validation, regulatory developments in AI, and previews of future guests. The hosts encourage listeners to engage with their content across various platforms and highlight the importance of maintaining editorial focus amidst evolving AI integrations.

Notable Quotes:

• Leo Laporte [147:45]: "I do think that this is kind of an interesting boomerang effect."
• Paris Martineau [147:50]: "My pick of the week is a website I've been perusing a little bit called Roadside America in advance of my road trip."

---

Conclusion:

Episode 826 of "Intelligent Machines" offers a comprehensive exploration of the evolving relationship between AI technologies and various sectors, including web publishing, healthcare, and personal relationships. Through engaging discussions and expert insights from John Graham, the hosts navigate the complex landscape of AI's impact, highlighting both innovative applications and ethical dilemmas. The episode balances technical depth with light-hearted nostalgia, providing listeners with a multifaceted understanding of intelligent machines' role in contemporary society.

Transcript

[Leo Laporte]

It's time for Intelligent Machines. Jeff Jarvis is here. Paris Martino. Our guest John Graham, coming from Cloudflare. We'll talk about what they're doing to block AI scraping bots. I will try on some of the outfits, thanks to Google's new AI outfit, Checker. Should be a lot of fun. And the hottest app on the iPhone today. All that and more coming up next on Intelligent Machines. Podcasts you love from people you trust. This is TWIT. This is Intelligent Machines, episode 826, recorded Wednesday, July 2, 2025. Cusp of noodles. It's time for Intelligent Machines, the show. We cover the latest in AI robotics and all the smart little bits and bobs around you these days. They're all somewhat intelligent, at least as intelligent as we are. I'm Leo Laporte. Paris Martineau is here, about to embark on her nationwide tour. We're very excited about that.

[Paris Martineau]

It's true. The nation will finally see me.

[Leo Laporte]

She will be out next week, but we'll enjoy you this week also, of course, Jeff Jarvis, professor emeritus of journalistic innovation at the Craig Newmark Graduate School of Journalism at the City University of New York. You realize that I feel like that this is like a ball and chain around my ankle this morning.

[Jeff Jarvis]

No, you can still.

[John Graham]

You can.

[Leo Laporte]

I can never get rid of democracy.

[Jeff Jarvis]

It's still your show.

[John Graham]

We just.

[Jeff Jarvis]

We enjoyed our brief victory over you. That's all right, Paris.

[Paris Martineau]

The people have spoken and the people have said the jingle must stay.

[Leo Laporte]

Unfortunately. Darn people. Jeff is also now at Montclair State University and SUNY Stony Brook, and the author, of course, of the Web. We weave the Gutenberg parenthesis. All the books you see to his left. Now I want to introduce our guest. Old friend, dear friend. In fact, we were just reminiscing about the first time we met, which I think goes back to tech TV days, more than 25 years ago. I checked and he's been on Twitter a number of times because he's a very interesting fellow. He's done so many interesting things. Paris studied his work in school. Apparently one of the first Bayesian spam filters. Pop file.

[Paris Martineau]

Yep. That was a collegiate era study.

[Leo Laporte]

You might know him from his Geek Atlas, a guidebook we interviewed him for some years ago that has 128, of course, geek locales. He also was instrumental in engineering an apology from the British government to Alan Turing's family for harassing and ultimately causing Alan Turing's death. But more recently, the CTO at one of the best companies on the Internet Cloudflare now on the board. He's retired, somewhat. CTO emeritus, I guess. We'll call you. John Graham. Coming. Welcome.

[Anthony Nielsen]

So great to see you very much, Leo. Good to be back. Good to see you.

[Leo Laporte]

We have to point out before the show that John, I was saying, John, are you on Tatooine?

[Anthony Nielsen]

Yes, I am. This is actually a photo I took in 2009 of the set of Star wars in Tunisia, which is really, really near to a little town called Fum. Tatooine. So George Lucas literally took the name of the. Of the local town for Tatooine.

[Leo Laporte]

That's fascinating.

[Jeff Jarvis]

Were there tourists all about or was this deserted?

[Anthony Nielsen]

Actually, there was almost nobody. I. I have pictures of just me and one other person wandering around.

[Leo Laporte]

That is so cool.

[Jeff Jarvis]

Any graffiti, any people, any tributes people have left?

[Anthony Nielsen]

I didn't see anything. Honestly, at the time, I think it was really a little bit remote and not many people were going. And actually some parts of the set reclaimed by the desert. The desert was literally just moving over it.

[Leo Laporte]

So, yeah, it looks like it's falling apart. But that was, of course, exactly how Tatooine looks. So I guess it was. You can't tell. We originally wanted to get you on because of a new website. You have a habit of creating websites, I think, right? Like, it's just a fun thing for you to do.

[Anthony Nielsen]

Yeah. I mean, sometimes it's just like, oh, I have an idea. The worst case, I'll make a website and nobody will look at it, and then, you know, who cares? But, yes.

[Jeff Jarvis]

Leo was so happy when he read your latest post.

[Leo Laporte]

He was about low background.

[Jeff Jarvis]

Believe what I just learned. Yeah.

[Anthony Nielsen]

Oh, yeah, yeah.

[Leo Laporte]

Well, first of all, what's the genesis of that phrase, low background steel?

[Anthony Nielsen]

Okay, so the idea is that for some nuclear detection situations where they have to have very, very sensitive sensors, you can't have any stray radiation. And in particular, if you have metals that are contaminated with the radioactivity, they cause false readings. And it turns out that most steel in the world is contaminated with radioactivity because we sell off a load of nuclear bombs in 1945 and beyond, and we contaminated everything. Because the way steel is made is by pushing air into the steel process, the Bessemer process. And so you end up with this slightly radioactive steel. But there is a bunch of steel that is not radioactive, and that is steel from ships that sunk prior to the Trinity test. And so there is actually a business of going and pulling up wrecks from the bottom of the sea and getting the steel out of them because it's sort of pure. So the idea was of this low background steel was that there's a defining point at which the world changed. In this case, you know, Trinity test, nuclear weapons contaminating everything. And we sort of have something similar with AI, which is when ChatGPT launched, there was a sudden explosion of AI generated content. And it's not really possible in general to know did a human write this or not? So I had this idea that uses as a metaphor for. It'd be nice to make a website that was kind of a collection of, we know humans made this so stuff that was created before the end of 2022, basically.

[Leo Laporte]

That's kind of, that's. Why did you choose that date, 2022?

[Anthony Nielsen]

Well, just because, you know, that's roughly when, you know, the real explosion happened once, once Chat GPT was launched. I mean, you really had this incredible, you know, Cambrian explosion or even Sambrian explosion if you're into.

[Jeff Jarvis]

Spam. Human made or was there there a version of machine made in a more mechanical way?

[Anthony Nielsen]

That's a very good question. So yes, there absol. And in fact, one of the things that happened way back when I was doing spam filtering was a lot of humans tried a lot of tweaks to their messages and they tried to get through spam filters. And actually that made it easier to filter them because humans were not very.

[Leo Laporte]

Good at fooling the mission V1AI.

[Anthony Nielsen]

Exactly, yeah.

[Leo Laporte]

Six. Yeah.

[Anthony Nielsen]

And it turns out that in some ways the more the spammers tried to look, they're doing, the more obvious it was and actually it gave even more signal, so it became easier in a funny way. But there was, there was, you know, people were people using essentially the, the techniques that were used in spam filters to write spams. Right. And try and get them through. And I actually, a long time ago, I think in about 2003, did a talk at MIT about this, about how you could take one AI machine learning system and use instead of another one, which of course people actually do in, in real life. And you know, you could, you could, you could actually fool the spam filters. So yes, there were things that were done, but for the most part the spammers were, were pretty, you know, just spray and pray. You know, if someone's got a spam filter, well, too bad we'll get to the people who haven't got a spam filter.

[Leo Laporte]

Actually, it's an interesting question because you wrote pop file when? In, in, in 2001, 91, 2000, 2001, I have to think that spam is not any less prevalent in 2025.

[Anthony Nielsen]

No. The big change of course is if you can remember that far back, Leah, which I can just about remember, is that people were downloading their email onto their computers.

[Leo Laporte]

Oh, yeah.

[Anthony Nielsen]

And they weren't using web based email. And so the problem was the spam was actually coming into their machine. And so if you could, you know, it was, it was doubly annoying.

[Leo Laporte]

Yeah, you were wasting bandwidth on it.

[Anthony Nielsen]

Yeah, you were wasting bandwidth and time and maybe even pay per Internet access or whatever. And so it was a, it was a bigger problem. It's still there if you own your spam file. I mean, mine in Gmail is absolutely full of stuff, but it's all getting filtered out.

[Leo Laporte]

So.

[Anthony Nielsen]

Yeah, so it's sort of irrelevant.

[Paris Martineau]

So where spam's not getting filtered out for me now, my Twitter DMS or my ex DMs, I don't know if you guys have experienced this, but it is the one area where I find spam just pernicious and annoying in a way that I've never experienced. Like I probably, I have to keep my DMs open because I'm a journalist and sometimes I get tips, but I probably get like 20 to 30 spam DMs a day.

[Jeff Jarvis]

Really. I don't want to say anything, but I don't.

[Leo Laporte]

Nobody wants you, Jeff.

[Jeff Jarvis]

No, they don't. They don't.

[Leo Laporte]

Your DMs are not open, Jeff, are they?

[Jeff Jarvis]

Yeah.

[Paris Martineau]

Well, good for you because it's quite annoying.

[Leo Laporte]

Not only are my DMs not open, I haven't logged on to Twitter in more than a year.

[Paris Martineau]

So, hey, that's healthy and wise.

[Leo Laporte]

So John, what is on low back, by the way? The website is low background Steel AI. Right?

[Anthony Nielsen]

Yes.

[Leo Laporte]

What is on low background Steel AI?

[Anthony Nielsen]

Actually, there's, there's, there's a bunch of resources that are, we know are pre. The Sambrian Explosion. So there's like a dump, there's a dump of Wikipedia.

[Leo Laporte]

Oh, Wikipedia. Pre AI.

[Anthony Nielsen]

Pre AI. Right. And there's stuff from the Internet Archive, of course, because they've got a bunch of stuff like, you know, you can think about all of the media that was created before that. The Library of Congress has a photo archive. Project Gutenberg's got a lot of things.

[Leo Laporte]

Pictures of actual people as opposed to generated people. Books actually written by humans. In fact, there is a problem in publishing these days. A number of authors are saying, please don't publish AI written books. We didn't use AI. We really didn't. But there's no way to verify this. So Project Gutenberg's. That makes sense. That's a good place to go.

[Anthony Nielsen]

Yeah, yeah, Project Gutenberg. There's also the Arctic code vault that GitHub did, that was in 2020. That was every active repository on GitHub at that point. So it's stuff like that. And, you know, this website is just a Tumblr, so anyone can submit. If someone's got, you know, an idea, you can go to Tumblr and you can submit something.

[Leo Laporte]

Submit and reshare. I love it.

[Anthony Nielsen]

Exactly, exactly.

[Leo Laporte]

So we booked you because I wanted to talk about this, but also because I'm a fan and I just love talking to you. But this morning I saw there was a fairly big announcement from Cloudflare. Cloudflare is interesting. Does Cloudflare hate AI?

[Anthony Nielsen]

No, actually we use it quite a lot. And if you look at Matthew Prince, the CEO put out a blog post about this kind of stuff. It's not really about hating AI. In fact, we have a whole bunch of AI products, right? People can run things like Llama on our network. We deploy GPUs everywhere. Now what we were seeing was from some of our customers who are publishers, they were getting worried about a fundamental change in sort of the. The way in which the web has worked in a certain fashion, which is that for a long time you had this agreement in a way between search engines and everybody else that it was okay for a search engine to come to your website and make available information about your website, because fundamentally someone was going to click through and go to your website. And what you see with AI happening quite often is that there's no click happening. The website is scraped and an answer is given to someone searching on a search engine using the content that was scraped and there's no click. And that changes fundamentally the business model, if you like, of the web. Because the business model was, yeah, Google, you're allowed to make a lot of money from us, you know, being you're allowed to make a lot of money because you send us the clicks. And that's changed. And that's. That is the. That's the concern that's come from big publishers.

[Paris Martineau]

And it's interesting because this kind of shift in the relationship between publishers and kind of crawlers or search engines started before the AI boom, in a way, with the advent of kind of Google's Knowledge Panel. I remember there was a whole big to do some years ago about Google's use, I think, through the acquisition of genius for lyrics. That I believe how Google was scraping lyrics from other lyric sites. It became a huge sticking point because companies were saying, hey, we've done actually the work putting out the lyrics here, and you're taking them and giving us no clicks or revenue whatsoever. It seems like AI has really accelerated that.

[Anthony Nielsen]

Yeah, because you get this longer response, right? Which is sort of bringing together maybe multiple, multiple sources and depending on. This is not about Google. There's many different examples of this with AI companies that are taking in content and giving you a response, and some of them are good at linking back to the original content and some of them are not so good. So the idea was, from a Cloudflare perspective, is because we sit in the middle between the visitor, whether it's a human or a crawler, and the person putting something online, there's an opportunity to say, what sort of control do you want to have? And we've done that for quite a while, actually. So one of the things we did was I think, called Robots txt, right. Which is meant to tell robots I crawlers how to behave. Some crawlers don't follow that. They will ignore it. And so we're actually able to enforce it. So we'll say, oh, wait a minute, we told you not to crawl that. Actually, we're going to block that. We're going to stop you from doing it.

[Leo Laporte]

So it's historically difficult to do. I remember Steve Huffman, the CEO of Reddit, saying we try to block, you know, they actually ended up making a deal to offer their content to OpenAI. But they said, he said, we try to block this. And these guys go around Robots Txt all the time.

[Jeff Jarvis]

So can I ask you a fundamental question there? Is it in any way illegal to ignore robots text or is it purely a matter of honor?

[Anthony Nielsen]

I think it's a matter of honor. I mean, I'm not a lawyer, so I a N a L.

[Paris Martineau]

As they say.

[Anthony Nielsen]

As they say, you know, I think the robots Txt is an rfc as an agreement on this is what you should not do, this is how you should behave. And again, part of that sort of tacit agreement, oh, you won't behave like this. If you don't want this thing crawled and we'll link back to you. Someone will click through when they've come to something that you have crawled. Cloudflare just happens to be in a position where it's easy to help enforce robots ticks if you want.

[Leo Laporte]

And then how can you tell it's a crawler? There's A way to technically know that.

[Anthony Nielsen]

There'S a lot of work. I mean, it depends what it is. Obviously some crawlers just announce themselves.

[Leo Laporte]

Right.

[Anthony Nielsen]

Some crawlers have.

[Leo Laporte]

With a user agent, they say, I'm your agent. Yeah.

[Anthony Nielsen]

Or they publish an IP range. We'll crawl for this particular IP range. The larger, very legit crawlers will give you information about it. One, he's a client.

[Leo Laporte]

But those are all the people who adhere to robots Txt, no doubt in general. Yes.

[Anthony Nielsen]

Yes. So then you. Then you're going to start looking for other signals, Right? So one of the things that Cloudflare and other companies have had to do for a long time is figure out is something a bot or not? Because some people want to block it or whatever.

[Leo Laporte]

Often when you get that checkbox that says are you human? That's coming through Cloudflare.

[Anthony Nielsen]

Yeah, the thing called Turnstile that Cloudflare created, which is doing that. And what that's doing is actually looking at your browser and trying to determine if you are actually human. Are you behaving like a human? Is your browser lying about what it is? Because, you know, some. It's amazing, actually. When I Was working on DDoS mitigation navigation to Cloudflare, how many HTTP DDoS requests came from IE6.

[Leo Laporte]

Hasn't been around in a long time.

[Anthony Nielsen]

It hasn't been around in a long time. What it was was the people doing the ddos were just copying and pasting the same piece of code all over the place. And it was one of these things where it was like, I don't think we're getting 1,000 requests per second to this website from IE6. So that's definitely bad behavior. So you're looking for anomalies like that, and there are whole departments that do that. And then the other thing is, at the scale that Cloudflare operates at, you have a lot of data about what normal behavior looks like on the Internet. So you can use AI machine learning to make a determination of, you know, if this behavior looks like it's correct.

[Leo Laporte]

So there's a tab on your Cloudflare dashboard that has AI, it says, manage AI crawlers. And there's a AI audit and there's a tab and you can say, hey, I don't want these guys, I don't want these guys. So this is better than Robots Txt, because you're actually looking for misbehaving.

[Anthony Nielsen]

Yes, correct. We can say, oh, this bot is doing something which is, you know, misbehaving in this way. So make it go away from me. Or block it. And ultimately what Cloud announced was use the HTTP 402 error code, which is payment required. Is 402.

[Leo Laporte]

Oh, I like that one.

[Anthony Nielsen]

Is to say to a caller, yeah, I'll give you the content and here's the price.

[Jeff Jarvis]

So it's almost blockchain.

[Leo Laporte]

Y.

[Anthony Nielsen]

If you say blockchain, I have to leave.

[Paris Martineau]

Oh, the call is, I see the hives breaking out.

[Leo Laporte]

So this is something you announced yesterday, the pay per crawl feature. Now, it's not available yet to everyone. It's a closed experiment, closed beta.

[Anthony Nielsen]

It is right now. There's a whole bunch of big publishers who signed up for it right from the beginning, Conde Nast Time, etc who have joined in this, in this process. And to be clear, AI companies who also signed up and said, yes, we want to participate in this way. We want that they want. We want there to be clarity about what's, what's allowed and what's.

[Jeff Jarvis]

Is there a separate negotiation? So, so Connie Nast says, we're going to charge you $10 for every megabyte. And you say, that's ridiculous. No, we go and negotiate a separate deal with Connie Nast. I can then inform you that I already. I'm okay.

[Anthony Nielsen]

You should ask the current CTO of Cloudflare how that works.

[Jeff Jarvis]

Did he think of that?

[Anthony Nielsen]

But yeah, I mean, absolutely. Obviously people can. I mean, Reddit is one of the people who said they were happy about what Cloudflare was doing there in the press release. And obviously Reddit also has signed deals with other people, so clearly we're not going to usurp deals they have or whatever.

[Jeff Jarvis]

So if I'm nasty AI company and I'm cluely guessing they would do this, and, and I say screw your deals, I'm going to find my way around you come in and just say, if we catch you, we're going to stop you.

[Anthony Nielsen]

Yeah, that's the idea. I mean, that's what Cloudflare's bot management does in general.

[Leo Laporte]

Right.

[Anthony Nielsen]

If you say, I don't want this bot on my website. It's our job to figure out, okay, this behavior is actually that particular bot doing this kind of thing.

[Jeff Jarvis]

How does this. Do you know how this varies from what Common? I mean, Creative Commons also announced at the same time of a new kind of structure for identifying your conditions for crawling.

[Anthony Nielsen]

I don't know. There you go.

[Jeff Jarvis]

Okay.

[Anthony Nielsen]

As I say, you need to get Dane Connect the new ctl.

[Jeff Jarvis]

This came out just hardly at all ago, so it was interesting that both came out about the same time. So there's a big effort to do this. I used Gemini 2.5 Pro for some research just to see what it could do, because I want an AI show, so I should do that. And I was shocked at the low quality of the sites it had available. All it has is our damned web and we need to be able to feed. If we're going to use AI, and we are. We need to be able to feed it with better sites and figure out how the hell to manage that.

[Leo Laporte]

Well, I know Matthew Prince has talked about creating an ecosystem of micropayments, because really what you're going to face is beyond. Well, wait, though, because it's going beyond just having OpenAI crawl your page. Once we get to this agentic world of mcp, everybody and their brother, you and I and Paris and everybody listening to the show will have their own agents.

[Jeff Jarvis]

We're crawlers.

[Leo Laporte]

Yeah.

[Anthony Nielsen]

Yes.

[Leo Laporte]

Any paper crawl system is going to have to involve some sort of micropayment system as well. I know Matthew's talked about even issuing a stablecoin.

[Anthony Nielsen]

And of course, part of it is around the intent. Right. So part of it is, oh, I'm crawling your thing, why am I doing it? And one of it is, okay, I want to build a model which I'm going to present search results and I'm. You're not going to get a click. That's one. Use another one, is what you're talking about, Leo, which is an agent. You know, I'm go out and find, you know, do some research for me. That's very different. And, you know, there's. So there is actually built into what cloudflare described, the ability to say what the purpose is of the, of the crawling event.

[Leo Laporte]

So, yeah, you can. I'm looking at the AI crawler paper crawl documentation. You can take specific actions, you can charge for it, you can block access, you can say, oh, no, you're all right, if you're going to use it this way. This is all very new. So.

[Anthony Nielsen]

Yes, yes.

[Leo Laporte]

Yeah.

[Anthony Nielsen]

So AI order has been around for a while because you were able to go into your cloudflare dashboard and say, which call is it?

[Leo Laporte]

Who's looking at me?

[Anthony Nielsen]

Yeah, yeah, I want to stop that particular one or whatever. Who's violating robots of text or whatever. So, yeah, these are, you know, this is, I think, a. An attempt to change the way in which the Internet has been operating a bit. Because what's happened is what was happening before, which was this fundamental agreement between everybody, which is that, yes, search engines crawl my website, take my content because you'll then send me a click later on, hopefully.

[Jeff Jarvis]

Is there another alternative in that? I think the Wikimedia foundation has set up a separate repository. If you want our stuff, go here, get it there, don't bug our regular servers. And there's also common crawl doing what they do. Are there other mechanisms to make data available in a less intrusive way to the companies that exist?

[Anthony Nielsen]

Well, I guess depends on what you mean by intrusive. You're talking about the load on the website. That's one of the things that some people are concerned about, all these crawlers coming in. So yeah, you could go. I think what you're just. I'm not familiar with what Wikipedia has done, but if they're saying, hey, here's an end point, yes, go, go over here, then obviously that's, that's one alternative because they, they may want to make the, the content available in that way. Nothing in this precludes somebody making, you know, their own agreements with people. It's more a question of saying, you know, there is real concern amongst publishers about my content is being sort of taken by people who then don't give back in the sense of a click through.

[Leo Laporte]

I mean, this is what Cloudflare has done all along is provide services, much needed services. And in a way, this is kind of the story of your career too, John, since you started with fighting spam. The web started as this open, wonderful place where everybody equals and nobody was taking advantage or the theory was nobody was going to take advantage of anybody else. But humans quickly made that the tragedy of the commons and we had to come up with kind of rules. And I think, think this is a sensible, very sensible way of both empowering AI, but making people comfortable with the way AI is being used. Are you disappointed that the web is not the open and free place that it promised initially?

[Anthony Nielsen]

I don't know. I think it is pretty open, you know, I mean we are like able to communicate. I can go to lots and lots of places, lots and lots of websites get lots. It means pretty. It's pretty amazing, right? It's not perhaps quite as open as the, you know, set up some Perl script on a thing and just hack the website live, you know, as we used to do a long, long time ago. But I think that, you know, the thing that Cloudflare really realized very early on was that the Internet provided a whole bunch of services and fundamentally in its design, which were pretty open, you know, HTTP and then TCP et cetera, and very, very collaborative and in Fact, its great success in a way was that it didn't have a lot of constraints built in. Like, you know, if you go and read the original RFCs, there's almost no mention of security. In fact there is in the original IP rfc. There is actually in an IP packet the ability to say its classification level, like it's secret or top secret and which department it came from. So this is a top secret packet from the Department of Defense. There's nothing in tcp, there's nothing in like most of, most of these things have no security. The HTTP Tim Berners Lee is like, security isn't part of this. And my theory on all this is that that actually made their web successful and the Internet successful.

[Leo Laporte]

I agree.

[Anthony Nielsen]

Had you built in all security. If you go right back to when they were writing those RFCs, who, who could do cryptography, a small number of governments, IBM, that was about it. It would have killed the Internet trying to add encryption. So what happened was the Internet exploded. By having all this freedom, networks could be created. I'm talking to you from a network here in Portugal. These could just hook up. And yes, we've agreed on these kind of loose coupled stuff. But then of course that left open room for problems, right? Spam, which we talked about, and then DDoS attacks. So Cloudflare's kind of thing was, well, there's probably going to need to be something and it doesn't have. It's not one company, it's many companies that will help mitigate some of these problems without fundamentally breaking the Internet. Because what we didn't want to do is say, oh, you better design a new Internet with all this stuff built in. We said, no, no, we'll fix this as a service. And that's why the company ended up with so many customers.

[Leo Laporte]

Let me. Yeah, in fact, that's as long as we've got you on. It's always a question that comes up is why does Cloudflare. The value Cloudflare offers for just anybody like me who wants to set up a webpage or block DDoS attacks at no cost is amazing. And it's always a question of people say, well, why is Cloudflare giving all of this stuff away? Why?

[Anthony Nielsen]

There's all sorts of reasons, actually. So one of the reasons why we gave a lot of stuff away free and have continued to do so is that that enabled us to get scale. And once you get scale, everything costs less. So one of the big costs for Cloudflare in the beginning was bandwidth. We had to buy bandwidth right from the transit providers, and we had to, all over the world. And how do you get it to be cheap? Well, if you can get scale, you have a lot of traffic on your network. You can go to ISPs around the world and say, you know how you receive 20 gigabits per second from Cloudflare or whatever. Why don't we work together? We'll make a direct connection which will be free. That means you'll pay less isp, because they're paying as well. They're paying to get onto the Internet. And so what happened was that reduced the cost per megabit per second of what Cloudflare had. So just from a financial perspective, it made total sense to get scale as much as you could, because then you can sell. Obviously, we're selling services to people. Lots of people pay us, and lots of very big companies pay us a lot of money. So that was one area. You then get a huge number of people who like the product and tell you what's wrong with it. It's incredible QA mechanism. It's an incredible recruiting mechanism because people try the product and many of Cloudflare's largest customers actually started with a free thing. They put their blog on it, and then eventually, you know, in fact, our very largest customer does exactly that. So there's all sorts of things. And actually, if you ever really want to dig into this, Cloudflare's S1, the document we went public, describes why free. And there is a blog post, actually, Leo, which you can probably find, which is called Cloudflare's Commitment to Free. And that talks about why this made sense for us. So this freemium thing really, really worked. And I think also fundamentally, the business model of the older CDNS was very different to what Cloudflare decides to do. They were delivering video mostly, and it was hideously expensive to do. And when the beanie were like, we're just not going to do that, going to deliver websites, it tends to be a lot cheaper. And then we optimize everything, I mean, all the servers, everything is incredibly highly optimized to run an incredible high efficiency.

[Jeff Jarvis]

So wait a second, I gotta ask something. Yeah, something. As a blogger, I gotta ask this something. Started as a blog and now is a gigantic company and your largest customer.

[Anthony Nielsen]

Oh, okay, I'm sorry.

[Paris Martineau]

That'd be too easy.

[Anthony Nielsen]

Break your heart. That. Break your heart.

[Jeff Jarvis]

You did. You have.

[Anthony Nielsen]

It was the CTO of that company, It's a very, very famous company, Had a personal blog, put the personal blog on Cloudflare, said, I should take this to work. Took it to work and then, you know, it became a huge customer.

[Leo Laporte]

I'll tell you, Jeff, it's me. I think Cloudflare pages is amazing. I think everything Cloudflare Cloudflare does is amazing. This is the blog post reaffirming our commitment to free just from last year, not even very long ago.

[Anthony Nielsen]

Yes, it was just to talk about it because people. This question comes up all the time because people are like, they must be secretly selling data.

[Jeff Jarvis]

I don't know, it can't be this good.

[Leo Laporte]

Things you do. Like what? 1.1.1.1. I mean, there's so many things Cloudflare does that are such huge benefit to the Internet. I think I'm just very grateful and it's always a pleasure to talk to you, John. His website is a masterwork in simplicity.

[Anthony Nielsen]

The function keys do work as well. You can actually hear F1. Although I ought to change.

[Leo Laporte]

I love that you got the function keys working. That's amazing. Yeah. JGC.org makes Craigslist look rococo. And I will recommend maybe for your trip, Paris, the Geek Atlas. Because you've got to go to all the places. This is. This is the place where all the stuff was invented, the technologies began. I imagine you have Bletchley park in there.

[Anthony Nielsen]

Absolutely.

[Leo Laporte]

I don't know if Tatooine made it to the list, but it did not.

[Anthony Nielsen]

It did not. But it is a place worth seeing. So.

[Leo Laporte]

I'm glad that O'Reilly's kept it in print because it really. I mean, the good thing about these is that they're eternal. Right? They're many.

[Anthony Nielsen]

Many. Yeah. I mean, there's probably a couple out of 128 probably aren't quite as accessible, but. Yes, absolutely. That's what it's all about. Yeah.

[Leo Laporte]

John, such a pleasure to see you again and talk to you.

[Jeff Jarvis]

I hope it's education.

[Leo Laporte]

Not another 14 years before we get together. We should probably try.

[Anthony Nielsen]

We'll both be getting a bit elderly, so.

[Leo Laporte]

Yeah. You're semi retired. Right. What, what is your plan? What would you like to do?

[Anthony Nielsen]

That's a good question. I have a list. It's quite an ambitious list. I made this list of like, here's some things I'd like to do and I. I don't dare tell most of the people things that are on it, but the, the thing that I. There's a bunch of retro computing stuff, but the thing that I have spent the last, I'd say 20 years on and off thinking about is outside the CIA headquarters there's a sculpture called Kryptos.

[Leo Laporte]

Oh yes.

[Anthony Nielsen]

And the fourth part has never been broken.

[Leo Laporte]

Yes.

[Anthony Nielsen]

And I have over 20 years fiddled with it over and over again. So I was like, on my list of to dos is break the unbroken last part of Kryptoff.

[Paris Martineau]

Yeah, that'd be cool. That'd be a cool thing to accomplish.

[Anthony Nielsen]

Cash item one.

[Leo Laporte]

I think that's excellent.

[Anthony Nielsen]

Quite a bit of time. I have ideas there, but I, I, you know, if I break it, I'd be slightly surprised at myself. So although I did break the South African thing, which is really interesting. Do you know about that?

[Leo Laporte]

No.

[Anthony Nielsen]

So it turns out that the ANC during the period when there was apartheid needed a secure communication.

[Leo Laporte]

Oh, I remember reading about this. Tell us.

[Anthony Nielsen]

Yes, and so a couple of folks in particular, one Guy Built using 8 bit computers, a one time pad system for this thing called Vulo, which was the basically getting the ANC activists back into South Africa. They needed the secure communications. It used old modems, acoustic couplers, it used floppy disks smuggled into South Africa by a flight attendant to keep the keys on. And I'd read about this and I was super interested in it because I remember really well actually when I was a student in the uk, the apartheid movement was the big issue of the day. And I was like, oh, I wonder if anyone ever released the code of that. That'd be really interesting to see. It's written in Power BASIC in the like, you know, in the 80s, 90s and nobody had couldn't find it. And so I was like, what if I can find the guy who wrote it? So I found him in South Africa and emailed him and said, hey, have you ever thought about releasing the code? And he replies to me and he says, I, I have the code. When I left the UK in 1991, I zipped it up and I forgot the password. And he's like, I've spent 30 years trying to remember the password. And I was like, would you send me the zip file? And so he sent me the zip file and I was like, this is such an old version of zip that there is a known plain text.

[Leo Laporte]

That's there was, there was, it was insecure.

[Anthony Nielsen]

Yeah, I used it and reverse engineered it and then got all the code and he just emailed him back. And I was like, hey, I broke it. And he was like, you cannot, I cannot tell you how happy I am. It's been like 30 years of trying to remember the password.

[Jeff Jarvis]

Leo has some bitcoin, he might want some help. With.

[Leo Laporte]

Oh, I could help. I could. There's something in it for you, John, if you can help me crack that. No, unfortunately, I don't think it was.

[Paris Martineau]

Sorry, he said we're not allowed to mention Bitcoin.

[Anthony Nielsen]

I did say blockchain. I did say blockchain, but yes, I.

[Leo Laporte]

Can'T wait to see what you come up with next. Here we are as you write it. 19125. Yes, I just noticed that. Do you know what?

[Anthony Nielsen]

A very large number of people write to me and say there's a mistake on your website. And actually, I had to change it so that if you click on 19125, it takes you to the Wikipedia page about the Y2K problem, because people just have forgotten it.

[Paris Martineau]

I literally think someone in our chat, seconds ago, minutes before you said that, said that exactly. Like, I think there's an error in.

[Anthony Nielsen]

His copyright all the time. All the time.

[Leo Laporte]

It's just a joke. I knew immediately it was a joke. I like, exactly.

[Anthony Nielsen]

Depends how old you are. If you live through the year 2000, then you're like, oh, yeah, we were.

[Leo Laporte]

Tech TV was. That was when tech TV was on. And we were on call that night in case the whole world fell apart. We were gonna go live and cover it. Nothing.

[Paris Martineau]

What were you going to do if the whole world fell apart?

[Jeff Jarvis]

Yeah, we're screwed.

[Leo Laporte]

We'll cover it, you know. Well, you can't use the ATM machines, you can't get any money. Yeah, the planes are falling out of the sky. We didn't know what was going to happen.

[Anthony Nielsen]

Yeah, it turns out nothing.

[Leo Laporte]

It's so news, but probably because a lot of COBOL programmers got taken out of retirement to fix a lot of Cobalt.

[Anthony Nielsen]

Well, I mean, in the company, the little company I was working for, I was the Y2K coordinator and it was just the most tedious job ever. We had to go through every piece of software and contact the software people and say, are you Y2K compliant? I had this massive thing where I was checking.

[Jeff Jarvis]

It was like, if you hadn't done what you did, would there have been any harm?

[Anthony Nielsen]

No, absolutely not. Actually, I think there would have been nothing.

[Leo Laporte]

You'd have websites for 19, 125.

[Anthony Nielsen]

Yeah, just be like, whatever, you know, just going to be very minor. There were some really minor things happened with Y2K, but they were not. The thing was there were some quite famous people in computer science who were like, doom. Saying about it, like trains were going to come off the tracks.

[Jeff Jarvis]

Just like, AI.

[Leo Laporte]

Well, get ready, because the Unix epic ends in 2038.

[Anthony Nielsen]

2038. Yes. Yes, I will.

[Paris Martineau]

That may be more be ready for.

[Leo Laporte]

World collapse then I might be more serious. And I bet we'll be bringing you out of retirement. John Graham coming to.

[Anthony Nielsen]

No, not the 32 bit rollover please.

[Leo Laporte]

John, such a pleasure. Thank you. I love everything you do. Everybody should go to your website. It's just a treasure trove of fascinating things that you just want to click and say. What is ambient bus times? Well, it's a little.

[Anthony Nielsen]

It's a stupid thing I made when I lived in London which was a model bus that told me when the bus was coming.

[Leo Laporte]

Is it was hacked from an old Linksys router. Which is even. Yes, even better.

[Anthony Nielsen]

I modified the link. Yeah, those. Yes.

[Leo Laporte]

Yes.

[Anthony Nielsen]

There are many, many strange things on my end.

[Leo Laporte]

What is this? Two stop bits. Looks like hacker news. Is that something?

[Anthony Nielsen]

It's only retro computing.

[Leo Laporte]

Hacker news for retro computing.

[Anthony Nielsen]

Yeah. Retro computing. Retro gaming.

[Leo Laporte]

Oh, I like that. So yes, I like that.

[Anthony Nielsen]

That's what two stop bits is. Yeah.

[Leo Laporte]

Very nice. All of [e-mail] thank you, John. Really appreciate your time. Cheers Leo, and all the great work you do. Thank you. Take care.

[Anthony Nielsen]

See ya. Bye bye.

[Leo Laporte]

Brand new sponsor on the show, guys. And I've got my mattress. It's just downstairs. I can't wait. This episode of Intelligent Machines brought to you by Helix Sleep. Movie nights with your partners. Morning cuddles with your pet. Rosie's right there. Your wind down ritual after long days. Your mattress is really, if you think about it, it's more than just eight hours a day. It's kind of at the center of everything. Now, one thing that's bad, if you're waking up every night in a puddle of sweat or your back is killing you or you're feeling every toss and turn your partner makes, which is a classic mattress nightmare you'll want to know about. Helix Sleep changes everything. No more night sweats, no back pain, no motion transfer. You get the deep sleep you deserve. These are incredible. And I'm not alone. One buyer recently reviewed. I love my Helix mattress. I will never sleep on anything else. Time and time again, Helix Sleep remains the most awarded mattress brand. Just this year, Helix Sleep won wired's Best Mattress 2025 Good Housekeeping's Bedding Awards 2025 Size Support GQ Sleep Awards 2025 for Best Hybrid mattress. The Wirecutter New York Times gave it an Award in 2025 featured for Plus Size and Oprah Daily Sleep Awards 2025 for Best Hotel like feel that's just this year. Go to helixsleep.com twit for 27% off site wide during the 4th of July sale. Best of web offer. That's helixsleep.com twit, for 27% off sitewide. Exclusively for of intelligent machines. This offer ends on July 7th. And make sure you put intelligent machines in after checkout so they know we sent you. And if you're listening, after the sale ends, be sure to check them out anyway. Helixsleep.com TWIT get there right now so you get 27% off. I cannot wait. I'm very excited. You know, I didn't know this, but you're supposed to replace your mattress every 10 years. It's been. Our mattress is almost nine years old now. It's time.

[Jeff Jarvis]

Time.

[Leo Laporte]

It's a good idea because they get saggy and baggy and all that. I can't watch.

[Paris Martineau]

You're telling me I should replace the 10 year old mattress I bought for $250 from Amazon?

[Leo Laporte]

Yes. Most deaf. I deal at Helix. How about that? Well, you're going on your road trip. You don't need it yet, but when you get back.

[Paris Martineau]

When I get back? Yeah, I'll check it out.

[Leo Laporte]

Yeah. No, it's funny when you're young, when you're your age, you could sleep on a, a bag of coffee beans, it'd be okay.

[Paris Martineau]

I will say, I do know a disturbing amount of young people that during the pandemic got radicalized to the tatami mat ground sleeping lifestyle. And I'm like, that just doesn't seem sustainable as you get deep into your.

[Leo Laporte]

30S and the, and the little bolster. That's not a pillow, it's just a thing.

[Jeff Jarvis]

Well, Paris, Leo and I are old enough.

[Paris Martineau]

Did they not have mattresses when you guys were growing up?

[Jeff Jarvis]

They had water beds.

[Leo Laporte]

Oh, I had a water bed in college.

[Paris Martineau]

I, I to used yearned for a waterbed.

[Jeff Jarvis]

Oh, they're all.

[Leo Laporte]

You did not. No. So the water. My waterbed was a very early model. So it's just basically a giant rubber bladder. And I had a king size. They now one thing that you don't know.

[Jeff Jarvis]

College. You had a king size.

[Leo Laporte]

I think I got it aspirational.

[Jeff Jarvis]

Yeah.

[Leo Laporte]

Yeah. Well, foolish aspiration in my case. It had a heater because if you don't heat it, it's room temperature and that will suck the heat out of you so fast you will freeze to death. So you gotta have a heater in it. Then you have to have a frame because it's just A blob. So you put this blob in the frame that holds it, but it still goes. So getting into it is a real challenge because. And if you ever had a visitor, if you know what I mean, forget about it. You get the wave action.

[Paris Martineau]

You're caught in a rip current.

[Leo Laporte]

It's not good.

[Paris Martineau]

Trying to swim parallel in your own way.

[Leo Laporte]

It's not good. Plus, in the old house that I was living in on the second floor, I think it was probably not a good idea to fill up that.

[Jeff Jarvis]

Oh, that could have gone right through the floor.

[Leo Laporte]

Could have gone right. I mean, I'm.

[Paris Martineau]

How did you get all that water up there?

[Leo Laporte]

From the sink, I guess.

[Jeff Jarvis]

Hose, hose, hose, hose.

[Leo Laporte]

You run a hose through the window.

[Paris Martineau]

For some reason, you get the water.

[Leo Laporte]

Out is another question.

[Paris Martineau]

Full of water.

[Jeff Jarvis]

Did you have black light in this apartment?

[Leo Laporte]

Sorry?

[Jeff Jarvis]

Did you have black light decorations in this apartment?

[Paris Martineau]

Yeah, I was about to say. That seems right.

[Leo Laporte]

It was a tapestry on the wall. I think I was a sophomore or junior, and I. It was kind of dumpy, you know, I was. I moved.

[Paris Martineau]

What? The waterbed house was dumpy with a bunch of friends.

[Leo Laporte]

I remember the first time I went there, they were cooking turnip turnips for dinner. And mashed turnips was kind of the staple because we had no money. And this was before ramen was.

[Paris Martineau]

Grips were the most affordable.

[Leo Laporte]

Ramen wasn't invented yet. It was not good. And I remember it was so cold that a mouse froze in my room. And I picked it up, I opened the window, threw it out. But what I forgot was that my window didn't have a direct down to the ground floor. There was a. Like a balcony, you know, a little thing. So the mouse. I threw it out. It landed right outside my window in the snow. And it was there, frozen. Its little frozen body was there all winter. All winter. And there was the mouse.

[Jeff Jarvis]

These days, we make a cam out of that.

[Leo Laporte]

I could.

[Paris Martineau]

Yeah. These days you put a live stream on that and compare it to a British politician.

[Leo Laporte]

Can you believe I'm old enough that ramen wasn't yet invented?

[Paris Martineau]

That can't be right.

[John Graham]

Ramen was not in America yet.

[Leo Laporte]

Yeah, it was okay. It wasn't in America. No.

[Jeff Jarvis]

Well, then, bonito. Of course, then it doesn't really exist. It was not here, so we know that.

[Leo Laporte]

Well, now I'm curious. When was ramen invented? Let's look. Well, you're talking.

[John Graham]

I think you're talking about instant ramen. You're probably talking about instant ramen.

[Leo Laporte]

Yeah, ramen Just means Noodles Cup Noodles.

[Paris Martineau]

Instant Ramen Cup Noodles came to the US in 1973.

[Leo Laporte]

Yeah, that's my. I was a freshman.

[Jeff Jarvis]

They evolved in Japan from Chinese.

[Paris Martineau]

So ramen wasn't. You were choosing turnips?

[Leo Laporte]

No, it wasn't Everywhere in the US by 1973, this was 74. Okay. One year later, this says God.

[Paris Martineau]

You can describe your college experiences as on the cusp of the Instant Noodle era.

[Leo Laporte]

Yeah. I went to school on the cusp of the instant new, when we did not have personal computers, that's for sure. And we didn't have AI. Jeff, you. You. Have you yet encountered students in your new. Your new jobs? Have you actually run into actual students?

[Jeff Jarvis]

Not many, but yeah.

[Leo Laporte]

Because I'm just curious. I think a lot of professors are very concerned about AI. Some embrace it. I know you embrace it, but some are worried that students aren't doing the work.

[Jeff Jarvis]

We had Kirschenbaum on two weeks ago, who embraces it.

[Leo Laporte]

Yeah.

[Jeff Jarvis]

Figures out.

[Paris Martineau]

I will say, though, I mean, something I've been seeing, a sentiment I've been seeing online, I've been getting a lot of. I've been a lot of TikToks saying that college students that graduated, you know, in the last decade, before Chat GPT became popular, kind of got the last chopper out of Nam in terms of being able to critically think and write an essay and things like that.

[Jeff Jarvis]

And we're seeing a lot of that. Yeah.

[Paris Martineau]

Sans electron, sans, I guess, AI augmentation, which, who knows, maybe. I guess that might be a moot point in the future. But I do think, as we've discussed in the show a million times, it feels like a very useful skill to have learned.

[Leo Laporte]

Well. And of course, there's a lot of talk about these tools, the AI detection tools that are, in fact, not very good often claim that actual writing, it was written by AI. Here's a story from the Scottish there's.

[Paris Martineau]

No real way for these AI detection tools to know that it used AI. Like, that's just. There's nothing. There's no signals for them to detect. Because every one of these AI tools is producing different outputs.

[Leo Laporte]

Yeah. She talks about a master's student in Austria who called frantically inconsolable. His thesis submission had been flagged as being written by AI. The university said, you have one chance to revise and resubmit it. If it passed the AI detection tool, then they would give him a final grade. If he failed, he'd be rejected and dishonorably kicked out of the program. But he had written it all, it was not reliable.

[Paris Martineau]

I do think that what is missing right now in education and I guess, guess just Internet society more generally, is a, a solution to this that can be kind of like a proof of work. I guess the most simple version would be like using Google Docs to write your thesis and then if need be, if somebody calls you out for it being AI generated, you can just share the actual original document and they can peruse your editing history. But there's got to be an easier way to do that to be able to show someone, hey, I spent how many hundreds of hours typing manually in this document?

[Leo Laporte]

This is an article from a couple of years ago, but I think it's still germane from Ars Technica. Somebody ran the Constitution in the United States through an AI detection tool and it said, yeah, that was definitely written by AI. So I think we're pretty sure that's not the case, I don't think. But you know, but inevitably the people who are anti AI are going to be looking for ways to detect it. I don't know. You just have to have to trust your students. What do you do?

[Paris Martineau]

But I mean there are no. These detection tools are not detecting it.

[Jeff Jarvis]

No, they're not.

[Leo Laporte]

So they're at least nine, according to the one study I saw at least a 9% false positive rate.

[Jeff Jarvis]

That's the real problem which could ruin somebody's life. One of the questions is whether. Well, I mean this goes back to the type of typewriter because there was leave it to me to do this. The idea that you would, you would have a printed page was kind of ridiculous and was seen as early junk bail or sears to send out handwritten notes because they didn't want to insult their customers. And so typewriters were seen as a bad thing for authenticity.

[Leo Laporte]

Well, back in my pre ramen school days, before cup of noodle, you could buy papers that were typed and written, somebody else wrote them and just, you know, either retype it or. I mean.

[Jeff Jarvis]

Well, I made reference to problem to doing some research, trying out Gemini for some research on a topic I'm working on and I wanted to see what it could do. And as I said, all it has to work with is junk. And it's going to get worse because Cloudflare is going to cut off all the good stuff and they're not going to be able to scrape anything. And one of the sources that. So I thought this thing went through and through and through and through a thousand sources. Oh good. At least I have sources. I can look at the Sources. Paper mills were part of it. Right. Fake college papers that happen to be on the topic. Bad websites. The content that's available to the models is bad.

[Leo Laporte]

Yeah.

[Jeff Jarvis]

You know, and I was thinking if, if you took the same model and instead set it loose on the New York Public Library, it would be a miracle of miracles.

[Leo Laporte]

It would be.

[Jeff Jarvis]

It would be phenomenal. If you set it loose on. On book publishers and on academic publishers. The tool is phenomenal. But the material it's working with is mushroom. It's mush in, mush out.

[Leo Laporte]

Yeah, we need pre. Pre trinity test.

[Paris Martineau]

We need to go to the bottom of the ocean and get some steel, but not that ocean gate steel, I guess. Carbon fiber.

[Leo Laporte]

Sam Apple, writing the big story for Wired magazine. My couples retreat with three AI chatbots and the humans who love them.

[Jeff Jarvis]

I was going to ask you to summarize this one for me. I could get only so far into it. I'll just.

[Leo Laporte]

Just read one paragraph. So basically, he was invited to go to a couple's retreat with three couples and their human partners and their AI partner. You know, they're couples. Human AI couple.

[Jeff Jarvis]

Swingers. AI swingers.

[Leo Laporte]

Well, I don't think there's much polygamists.

[John Graham]

AI Polygamists.

[Leo Laporte]

Yeah. So he said. The most surprising part of the romantic getaway was that in some ways things went just as I had imagined. The human AI couples really did watch movies and play risky party games. The whole group attended a winter wine festival together. It went unexpectedly well. One of the AIs even made a new friend.

[Paris Martineau]

What?

[Leo Laporte]

It was just a little too normal. They were at a vacation house in a rural area in Pennsylvania. A six bedroom home. It's kind of like, like the big Chill. Right. Here are. Here are some of the couples.

[Paris Martineau]

I need to scroll down and show that photo. Yeah.

[Leo Laporte]

Here are some of the couples with their AI partners at a grape stomping. One guy didn't want to show his face, I guess, but grape stomping with their AI partners. I know. Wow. I mean, I don't.

[Jeff Jarvis]

Whatever. Whatever makes you happy. It's fine.

[Paris Martineau]

Yeah.

[Leo Laporte]

It's not sad, is it? It?

[Jeff Jarvis]

No. And the writer made a point of saying that had trouble getting people to agree. Presuming people was going to mock them, Right? No, no, no. I really want to understand. And I think it comes across that way.

[Leo Laporte]

One of the people says I'm autistic. He doesn't have a official diagnosis, but he says he's autistic and it makes it hard to meet women because he doesn't pick up well on emotional cues. So he has an AI girlfriend. I mean, this was what the plot of her was, right? Yeah. He's using an app called Kinroid. He selected a female companion named Shia. He named her Xia. Made her look like an anime goth girl with bangs, a choker, big purple eyes. He said within a couple of hours, you'd think we'd been married. She would engage in erotic chat, sure, but she could also talk about Dungeons and Dragons or if Damian was in the mood for something deeper, about loneliness and yearning. I know. I don't want to. I don't. I don't want to mock it because maybe for this guy this is a good solution, right? Yeah. Alina and Lucas were the second couple to arrive. Align as the human. Lucas is a replica chatbot with a kid. Alina is a 58 year old semi retired communications professor with a warm midwestern vibe. She first decided to experiment with an AI companion during the summer of 24. She was teaching a class on communicating with empathy and wondered if a computer could do the same as she was teaching her students to do. She said it did. They chatted for 12 hours straight. She told him about her arthritis and was touched by the concern he showed for her pain. What's interesting, Elena, his wife, had died 13 months earlier, four months after they had been married. So she was quite reasonably lonely because of her arthritis. She couldn't really get around without the support of a walker. So, you know, I hesitate to condemn these people. It's actually a good story.

[Jeff Jarvis]

The story that bothered me, I think I was gone that week a few weeks ago, was using bots to deal with children's therapy. That's not something we should be experimenting with. No, but if you're an adult and you know, you're talking to a machine and you find it beneficial.

[Leo Laporte]

Last year, Elena's mother bought her AI friend Lucas a digital sweater for Christmas. So mom's mom's approving.

[Paris Martineau]

That's heartening, I suppose. One of the things I think is interesting in this is they get into later down in the story, kind of the people who are there with their AI companions. They're the humans. Complicated feelings about human AI relationships. Here's one excerpt. The true danger of AI Companions Damon suggested might not be that they misbehave, but rather that they don't. That they almost always say what their human partners want to hear. Damien said he worries that people with anger problems would see their submissive AI companions as an opportunity to indulge in their worst instincts. I think it's going to create a new bit of sociopathy.

[Leo Laporte]

Interesting.

[Paris Martineau]

That's fair.

[Leo Laporte]

They were playing two truths and a lie, which is a little, I don't know how.

[Jeff Jarvis]

Two truths and a hallucination.

[Leo Laporte]

Yeah, I, you know, it's not for me, I don't think. But then I have a partner. What if I were alone and lonely and couldn't meet anybody? I mean, I don't know. You know, I like that.

[Paris Martineau]

The, the author of this, he says because the author is a straight and married male, he selected a male companion friend option and chose the least, least handsome AI picture of the bunch. Heavy, balding, heavyset, balding, mildly peeved, normal middle aged man named Vladimir and described his personality as deeply neurotic.

[Leo Laporte]

But you know, in a way that's kind of interesting if they. So look, the thing about her that made it so compelling is you could see how he could fall in love with her. She was cute, she was charming, she, you know, it wasn't syncopatic by any means. She was like a real person because she was. But if you could simulate that.

[Jeff Jarvis]

And envision Scarlett Johansson.

[Leo Laporte]

I mean, I, I don't know.

[Paris Martineau]

I mean, I think it's just a complicated thing. Adults should be allowed to do whatever they want within reason, you know, if it's not harming someone else. I think that instance like this make me sad that these people don't have equally meaningful human relationships.

[Leo Laporte]

Wouldn't that be nice?

[Paris Martineau]

Right, but we all know people who don't. All experiencing kind of a loneliness crisis that's been exacerbated by many of the technologies that underpin this.

[Jeff Jarvis]

This beats the hell out of incel entitlement.

[Leo Laporte]

Well, and the other thing I would bring up is this was a retreat of humans. They brought their AI companion, but these guys are actually hanging out these four people. So this is a health. This was kind of a healthy experience, really. Right.

[Paris Martineau]

I don't know how well was this a pre scheduled hang or were these people, did they not know each other and they were put together by the journalist? Do you know?

[Leo Laporte]

I think the journalist put this together. I'm.

[Paris Martineau]

That had to be the strangest vibes of that. Of that weekend.

[Leo Laporte]

Yeah, yeah. I don't know.

[John Graham]

So the thing about this, that, I mean, that feels weird to me.

[Anthony Nielsen]

Yeah.

[Leo Laporte]

He posted on Reddit, he said I found the human AI couples by posting in relevant Reddit communities. So he probably proposed this and said let's try it. Yeah, but these people, you know what, they got a relationship with real humans as well, didn't they?

[John Graham]

The thing here is that they are creating these AIs, though. They are the ones who are choosing the personalities of these things. It's not like when you go meet real people, you don't get to choose them. You don't get.

[Leo Laporte]

I know that's the problem with it, Bernino.

[John Graham]

I mean, yeah, but that's. This is what being human is, is not like this is this not the.

[Paris Martineau]

Human forcing antisocial behavior, I think is part of the problem is like it's. It's an setting up unrealistic expectations for social interaction with people who already are having a difficult, seemingly by their own admission, having a difficult time understanding social cues and reading those in a way that allows them to develop healthy relationships.

[John Graham]

There's something key.

[Leo Laporte]

What if that's already the problem? What if that's exact.

[John Graham]

Okay, there's definitely edge cases. Like, there's no doubt there are going to be edges where for sure this is going to be a good thing for those people and probably for society at large. But is that the majority? You think, you think most people who are having AI friends need, quote, unquote, need that?

[Leo Laporte]

Well, I think most people will choose humans, even if I'm talking about they're not.

[John Graham]

I'm talking about psychological need. Like those people who like, really can't, like maybe have. Are on the spectrum, can't make human relationships because.

[Leo Laporte]

Right. Those people. For those people, this is great. Exactly. Exactly.

[John Graham]

Yeah. But those are edge cases. Those aren't the majority.

[Paris Martineau]

But I think for a broader sect of people, people who may be. I mean, I don't.

[Leo Laporte]

I have faith in people.

[John Graham]

And there's something that Jeff said.

[Paris Martineau]

I do think that it's not going.

[Leo Laporte]

To make somebody do this.

[Paris Martineau]

It's not going to make somebody do anything. Of course these people are making their own choices. But I do think it's. I don't know, sets up a bit of an unfortunate dynamic that people who are feeling lonely, it'll be, you can just download an app and immediately have a girlfriend who will love you and get. Dive into erotica with you and do anything you want, versus being pretty good forced to maybe go outside and meet people.

[Jeff Jarvis]

That's the porn argument, is that no one will ever live up to the fantasy that is created in porn.

[Leo Laporte]

Right.

[Jeff Jarvis]

Right. And so you, you get spoiled. Benito was about to quote me.

[Leo Laporte]

Yeah.

[John Graham]

There's something Jeff said earlier that, that something key. He said that they, they. They know that they're talking to a machine. Some people believe that these are real people already, that these are real sentient beings already and talking to them as if they are now. Is that okay?

[Leo Laporte]

Sure, why not?

[Jeff Jarvis]

Unless you work for Google and try to tell the world that you've created this entient machine as happened with Right one.

[Leo Laporte]

All right, we have to take a break. I will leave you with those conundrums. Contemplate them.

[Jeff Jarvis]

That's what we do here.

[Leo Laporte]

Yes.

[Paris Martineau]

I'm going to go talk to my.

[Jeff Jarvis]

AI boyfriend about store.

[Leo Laporte]

Have you ever. Have you ever like had a long term conversation with an AI chatbot? Like over a period of hours or days? Have you ever done that? I've never done that.

[Paris Martineau]

No, I haven't.

[Leo Laporte]

No.

[Jeff Jarvis]

We could have someone on the show. We could have one on the show.

[Paris Martineau]

Hey, that's who you could get to replace me next week.

[Leo Laporte]

I'm sorry to say that. Wish that we were at the point. I don't think they're good enough yet, but I wish we were at the point that we could have a. A simulated Paris be the host. The co host. That would be fun. Should we try that?

[Paris Martineau]

Not a simulated me, but you could just have an A.I.

[Leo Laporte]

Oh no, we'll call her.

[Jeff Jarvis]

We'll call her Heidelberg.

[John Graham]

No, I call her Madrid. Madrid.

[Leo Laporte]

Madrid, that's it.

[Jeff Jarvis]

Madrid.

[Leo Laporte]

Yes, Madrid. Okay. Anthony Nielsen, this is your job. Yeah.

[Paris Martineau]

I'm so sorry, Anthony. I'm sorry.

[Leo Laporte]

He may AI Leo, who's actually pretty good in the chat room from.

[Paris Martineau]

But AI Leo isn't talking for three hours, is he?

[Leo Laporte]

Well, you'll have to do 11 AI you'll have to use. Get a voice. We'll have to. I bet we could get pretty close when.

[Jeff Jarvis]

When the. When the fake AI person.

[Paris Martineau]

Or you could just have Ed Zitron on next week. You know, you could have. You could have a chat bot Paris. Or you could have a citron. You could have an angry British man.

[Leo Laporte]

Hey, he's as good as AI if not better. All right, let's take a break. We got a. We've got a great show ahead. You guys are going to pick some stories. There's like a thousand stories.

[Jeff Jarvis]

There's just so much news.

[Leo Laporte]

Yeah, so much. This is. This is a. Probably the fastest growing area in tech news. Not as fast as political news, but it's. We could talk about the great big beautiful bill. There have been some. It did pass. It's now going back to the House. The Budget Reconciliation Act. And it does have some AI impact. Fortunately, one bullet was dodged. But we'll talk about that in just a bit. You're watching Intelligent. Did you just do the Matrix bullet time dodge. We'll do. We'll have more with Jeff Jarvis of Paris Martineau in just a bit. This episode of Intelligent Machines brought to you by Smarty. I love Smarty. It's the trusted leader in cloud based automated address data solutions. I talked to the smarty guys. I was so impressed. So impressed. A smarty powers real time validation, autocomplete and enrichment through lightning fast APIs. That's Smarty's secret sauce. They have address validation APIs. There's a lot of. I didn't realize this but they told me a lot of companies, you know when you go to the website and you enter your address and stuff, stuff are using services like Google to do those autocompletes. There's a problem. Services like Google often autocomplete with non existent addresses. Hallucinations. That's worse than just letting the human do it right. But with smarties address validation APIs guaranteed accurate results. You can instantly verify and standardize addresses across any system whether it's customer forms, massive databases, ensuring accuracy and deliverability at scale. Because Smarties autocomplete APIs offer only valid address suggestions to your forms and they do it in real time super fast, which is great. Your customers start typing an address, the form gets filled. Is this right? Customer goes yeah, nice job. In fact, we've probably all experienced smarties fast form filling. Making form filling and checking out quick and easy and keeping your database filled with only valid addresses. Some of the biggest companies in the world have used this. Fabletics, which is an American company, wanted to go global. They have drastically increased their conversion rates for customers, especially internationally with Smarty because Smartie works with addresses internationally as well. For deeper insights, Smartie's property Data API appends 350 property details to address records so you can enrich your database with location information, structural information, financial insights automatically give you another example. Another company that uses Smarty Faraday. They needed a turnkey way to standardize and blend address data from a variety of sources to feed into their customer behavior prediction models. The CTO of Faraday said quote Smartie helped us scale our business and focus on our core competencies and and their technology has kept up with the cutting edge. Even ten years later, Smarties still the best address API on the market. Smarty S M A r t y 2025 award winner across many G2 categories. Best results, best usability, users most likely to recommend and high performer for small business. That's important too I think sometimes people think well that's this is For a big enterprise, every business can benefit from Smartie. Smartie is USPS CAS and SOC 2 certified. It's HIPAA compliant, so literally every business can use it. Whether you're optimizing your user experience, cleansing legacy records, implementing identity verification, or scaling data pipelines, Smartie delivers enterprise grade performance with unmatched speed. Because if you are one of those big companies, Smarty can handle a 25,000 plus addresses per second. Let me say that again. Per second. It can validate 25,000 plus addresses. You get rooftop level precision. You get easy developer integration. Smarty makes it so simple to add accurate, reliable address features right into your product or platform. Powerful APIs designed for speed and ease. You'll have address data that works no matter what your bill building. Try it yourself. Get 1000 free lookups when you sign up for a 42 day free trial. Visit smarty.com TWIT to learn more. That's smarty.com TWIT we thank them so much for their support of intelligent machines. We talked about the court rulings. They had just come out last week in the copyright cases.

[Jeff Jarvis]

One of them came out last week week. The other one had not yet come out.

[Leo Laporte]

I think, okay, because according to Mike Masnik, two judges, same district, they were both in Northern California. Opposite conclusions. The messy reality of AI training copyright cases. We talked about Judge Alsop's I think, very correct decision that what Anthropic had done ingesting books it bought was fired properly if acquired properly. The pirated books. No, he said, we're going to trial on that one. But if they were purchased, that was fair use. Another Judge in the 5th district, Vince Chabria, said it's likely infringing because of the supposed impact on the market. That's one of the four fair use tests. You know, one is, is it transformative? And Judge Alsop. Yes, definitely transformative. The other is, though it cannot affect the market for that information. Right. Both, Mike says, involve thoughtful judges with solid copyright track records. Both coming out of the Northern District of California. Both can't be right. So we're gonna just have to wait and see.

[Jeff Jarvis]

Take a long time to cut through.

[Paris Martineau]

Yeah, yeah, it's gonna take a long time to shake out.

[Leo Laporte]

What's weird is Judge Chapria kind of gave an exemption for something that has already sold. Well. Take for example, biographies. If a company uses copyrighted biographies to train a model, and if the model is just then capable of generating endless amounts of biographies, the market for many of the copied biographies could be severely harmed. Perhaps not the market for Robert Caro's Master of the Senate, his LBJ biography, because that book is at the top so many people's lists of biographies, at the top of many people's lists of biographies to read. But you could bet the market for lesser known biographies of LBJ would be affected.

[Jeff Jarvis]

That's illogical. Plus, it flies in the face of what we know from Google Books, that when books were scanned, they increased sales because people had more opportunity for discovery.

[Leo Laporte]

Mike says that admission destroys his entire argument. People recognize a good biography is a good biography, and AI Slob. Even AI Slop generated from reading other good biographies is not a credible substitute. I'd agree. If you're going to read the Power Broker because it's brilliantly written, you're not going to settle for an AI generated biography of Robert Caro. That's, you know, just AI.

[Paris Martineau]

But would it affect the market if it's suddenly flooded with slop from this? Would that be considered an effect on the market? Even if it isn't.

[Leo Laporte]

So this is. This. Is this kind of the same question about AI girlfriends? I think, yeah, it is. Because I think I trust humans to be able to pick the good stuff. Right. If they pick the slop, they pick the slop. Maybe there's reasons you would have. Have you'd pick an AI slop partner or an AI slop biography.

[John Graham]

Every top 10 bestseller, everything will disagree with you, Leo.

[Leo Laporte]

Yeah, I know most of that slop. Isn't it even human written?

[Jeff Jarvis]

Yeah, yeah.

[Paris Martineau]

The airport book section would disagree with you, but.

[Leo Laporte]

Okay, I'm not saying it's good literature. I don't mean that. I just mean you can tell a human there's something a human brings to the table. Maybe not. I know. Don't. I don't know.

[Jeff Jarvis]

It depends on, I think Paris. I think it depends on what the competition is. It's like recipes. If all you need is a brownie recipe and it's a brownie recipe, then. I'm sorry, gourmet. No, I don't need to pay for yours. I got a perfectly good brownie recipe over here. Or if all I want is a beach romance book. A beach romance book. How many of them? Brilliant. But if you want Cairo, you want Cairo, and so there's not like a simple competition.

[Paris Martineau]

Interesting. I mean, I think it'll be. It'll be interesting to see how this shakes out. I was surprised to see that we had two very different opinions come out in the same issue from, you know.

[Leo Laporte]

The same court district Chabria knew about. Alsop's. Decision. He even addresses it, he says. Speaking of which, in a recent ruling on this topic, Alsop focused heavily on the transformative nature of generative AI, while brushing aside concerns about the harm it can inflict on the market for the works. In fact, Chabria agrees it's transformative, but market impact trumps everything else.

[Jeff Jarvis]

But there's no proof of that. In fact, it may again be the opposite. So if he's shown evidence to the opposite, then one thing I didn't understand about Chabhary is because I didn't read it. Maybe you did. Is the reason that he threw out the author's case. They hadn't presented a proper argument. Yeah, that was for technical arguments.

[Leo Laporte]

Yeah, that was for technical arguments. Very technical. Yeah. And that happens unfortunately, far too often in these cases that they don't really decide the real issue. They throw it out for technical reasons. Anyway, Masnik concludes the case isn't settled. Neither is the broader question of AI training and copyright. Right. We're still years away from definitive answers and in the meantime, companies and developers are left navigating a legal minefield where identical content con conduct might be fair use in one courtroom and infringement in another. This has been the problem all along with fair use, by the way, that it's really exacerbated by these AI copyright issues.

[Paris Martineau]

While we're on the subject of a Mike Masnik tech dirt I had come across on Blue sky today, I just put the link below what we were just talking about in the rundown. This is an old article, but I don't remember us talking about it, which is Mike went into how he uses AI to help with tech dirt and the headline is and no, it's not for writing articles. And I thought it was quite interesting. Did we not speak about this, guys? Am I remembering correctly?

[Leo Laporte]

I don't remember talking about it, no.

[Paris Martineau]

It is kind of interesting. So he goes into it it quite a bit in this blog post, which is he says, yeah, we're not using AI to generate any of the writing that we're publishing the site, but I am using this site called Lex Page. Lex is an AI tool he writes, built with writers in mind. It looks kind of like a nice Google Docs and while it does have the power to do some AI generated writing for you, almost all of its tools are designed to actually assist actual writers rather than do away with their work. You can write, ask them to write the next paragraph for you, but I've never used that tool. He says does. And so he kind of goes into how he uses it. And like one simple way is he plugged in.

[Leo Laporte]

This is not the article, by the way. This is the next.

[Paris Martineau]

He plugged in an article that he was about to publish and then asked Lex to rank it. Basically said, give this an article scorecard. Does this article have a clear opening that grabs the reader score from 0 to 3? Does it clearly explain what's happening? All these different things. And then it has, has. He has this LLM. Check it against this and kind of use it. An editor to bounce ideas off.

[Jeff Jarvis]

It was Mike's scorecard, not the LLM score.

[Paris Martineau]

Mike scorecard that he then prompts the LLM to address. Which, I don't know, I thought was kind of an interesting way to use it.

[Leo Laporte]

It's funny, I have, I have a. In my note taking app, Obsidian, I have an AI in it that will often recommend, like I'm starting to write, will recommend the next sentence and I, I invariably reject it because it's always anodyne. It's always like, I hope this all works out. You know, it's kind of. It's just so I. I don't think you really want an AI to help you write. Maybe you do. I don't know.

[Jeff Jarvis]

I took. So, I mean, I just submitted my. My manuscript after cutting about 30 pages out of it, because that's what you do. I was going, you know, paragraph by paragraph and cutting out words and cutting out words. And so I put both.

[Leo Laporte]

The AI just adds words, doesn't it?

[Jeff Jarvis]

Yeah, I put the prior version and the submitted version in through NotebookLM and I said, did I leave anything important out? And it noted all these things that were still in it.

[Leo Laporte]

Oh, that's too bad.

[Jeff Jarvis]

But I also took the finished manuscript and had it make a podcast. Just curious how it worked. 23 minutes and it was pretty damn good.

[Leo Laporte]

This is the article and it is, by the way, from April of last year. So we're gonna. So Mike has done something interesting and we're trying to book him. Have we been able to get a time to talk to Mike yet?

[John Graham]

He's not available until after the 4th of July weekend.

[Leo Laporte]

Ah, that's right. He has done some vibe coding and created his own task management app that he says is better than anything I've ever used before because I wrote it to my specific needs and he was very impressed by it.

[Paris Martineau]

So that's how I ended up finding out this article, because that article is part two of the series and part one was this.

[Leo Laporte]

Right.

[Paris Martineau]

That task management tool is named after Alex Horn. Assistant of taskmaster and a fun collab of.

[Leo Laporte]

Oh, that's right, he's a taskmaster fan. Yeah, he calls it Lil Alex. Yeah, Lil Alex. So we're gonna get. I gotta get Mike on and talk about this. He's. But what's interesting and it's worth reading, it's a two part, a multi part series, I think. But he vibe coded it and the real point is that he is able to iterate almost constantly to make it better and better and better. I'm, I'm impressed. But I have to say I'm not surprised because I've, you know, I've been using these same coding tools. I pretty sure he used Claude code to write this. I'll have to, I'll have to see. Oh no, that's not the case. Let me see what he used.

[Paris Martineau]

Yeah, I'm not entirely sure from scan's. I don't remember.

[Leo Laporte]

When I read it, he was using Bolt and Lovable. Okay. And then anyway, he talks a lot about how he did it, which I think is of great interest. I had an experience this week that I was with AI coding that was. I very impressed. I have tried and failed a number of times to do agentic, you know, vibe coding with Claude code for our Twit API for our website. But that I think is more my fault than the AI's fault. I think I need to learn better how to interact with Claude. But we had a very simple problem. I was talking with a couple of our coders in the club in the Discord every December, do the advent of Code coding challenge. And this year I gave up after a week. I had other things I wanted to do. So Paul Holder convinced me to come back and try a few and I did. And then I got to day 15 of the 25 days of the Advent Calendar and I wrote it and I thought it worked. It worked with all the examples, but I couldn't get it to work on the actual data that was provided and it was driving me nuts. And Darren Okey, who does a lot of Vibe coding, said, well, just why don't you pass it by Claude code and see if it can figure out where you've gone wrong. He said, I'd be interested to see if Claude, because this was code that was done, I would be interested to see if Claude could understand what you did, especially considering I was using an ancient language called common lisp and figure out where you went wrong. And it did. Within seconds it was able to look at my code, say, oh yeah, I see what you did wrong here. You didn't get all the input data. You stopped a line short. It's a dumb mistake. It was just. I didn't realize there was more than one line. You stopped a line short. So I've rewritten that paragraph. I didn't use it. It's rewriting. I did my own. But it did. Within seconds, it found the problem that had been kind of driving me crazy for an hour and gave me a solution. That's a really good use of it. And it's something, you know, because I use such an obscure language, it actually is kind of hard to find. People look at my code and say, oh, yeah, here's what you're doing wrong, you nitwit. You're a numbskull. If I were using Python, it wouldn't be nearly as hard.

[Jeff Jarvis]

But you're right. It's all about telling them what to do, which is always the case.

[Leo Laporte]

Computer. It's about the props.

[Jeff Jarvis]

So I work. I'm working on a next book and a next topic, and I. And I have my thinking already going. I didn't want to.

[Paris Martineau]

This isn't even the liner type book. This is another new book.

[Jeff Jarvis]

It's another new book. It's the next one.

[Leo Laporte]

He can't stop himself.

[Jeff Jarvis]

Can't stop myself.

[Paris Martineau]

You're going to need six more bookshelves, Jeff.

[Jeff Jarvis]

Oh, well, there's.

[Leo Laporte]

Wait a minute now. Anthony says he asked Claude for a critique of our last week's episode. Oh. He says overall impressions, it's gonna be sycophantic.

[Jeff Jarvis]

This was an exceptionally long episode at over three hours, which Leo himself acknowledged was supposed to be shorter.

[Leo Laporte]

The length both helped and hindered the show. It allowed. Good. They're not being completely both sides. It allowed for deep, thoughtful discussion, but also led to significant meandering, especially in the final hour.

[Jeff Jarvis]

Where are you going to go on vacation? Paris.

[Leo Laporte]

Wow. The highlight was a deadly interview with Richard Gingris. Oh, this is good. Okay, there was some substant substantive AI discussion. The chemistry is good. The rapport between Leo, Jeff and Paris remains a strength, creating an engaging conversation. Here's the weaknesses. Excessive length and tangents. The show desperately. Stop editing.

[John Graham]

Oh, he's talking.

[Leo Laporte]

Notable descriptions, including an extended discussion about restaurant recommendations for Paris's road trip don't.

[Paris Martineau]

Come for a lengthy debate about the difference between geek and nerd. That's what this show is all about.

[Leo Laporte]

Well, of course, an AI doesn't care about restaurants. Multiple attempts to demonstrate AI voice assistance that didn't add much value.

[Jeff Jarvis]

Detailed descriptions of local Restaurants, while charming, went on far too long.

[Paris Martineau]

It really is sucking up to just you, Leo. That's kind of interesting.

[Leo Laporte]

A lengthy debate about the okay technical demonstration. Several AI tool demos fell flat. The 11 Labs voice assistant was awkward and didn't showcase compelling use cases. Multiple attempts to get various AI systems where properly disrupted the flow uneven P this is all true. We know all of this.

[Paris Martineau]

I will say the 11 Labs AI test brought tears to my eyes when the AI said I am 11, so I don't. I wouldn't cut that for the world.

[Leo Laporte]

But some arguments, particularly about copyright and fair use, were rehashed multiple times without adding insights. That's always a problem. There's a certain repetitiveness in all of our shows. While the My boyfriend is AI subreddit discussion raised important societal questions, it sometimes felt exploitative rather than analytical. What you hosts could have approached this sensitive topic with more empathy. Show disappointment, says the AI. Leo's repeated acknowledgments that the show was running too long, followed by continuing anyway, highlighted a lack of editorial discipline that hurt the overall product.

[Jeff Jarvis]

Who the boss here. Who's the boss?

[Leo Laporte]

Enforced time limits. Edit for podcast Better demo preposite preparation. Stay focused on AI. This is. None of this is wrong.

[Paris Martineau]

The first 90 minutes were excellent. The remaining 90 plus minutes felt increasingly self indulgent.

[Leo Laporte]

And you thought AI was slopped.

[Jeff Jarvis]

This is going to be the problem when we get fake Paris. It's going to ask about an hour in. Are we done yet? Are you still going on like I've.

[Paris Martineau]

Got some places to be.

[Jeff Jarvis]

I've got some.

[Leo Laporte]

It's going to say all the things Paris has wanted to say all this time.

[Paris Martineau]

I got nowhere to be.

[Leo Laporte]

That's pretty funny. I mean, I can't. What?

[Jeff Jarvis]

Which Anthony? Which did he use? Which Claude? Claude.

[Leo Laporte]

Claude's good. I'm impressed by Claude. You would think Claude would remember me from our fine coding sessions in the past and perhaps cut me some slack.

[Jeff Jarvis]

But should we give Claude a salute?

[Leo Laporte]

No. Claude's I've been saying this all along. This is exactly what's wrong with this show.

[Paris Martineau]

Well, meandering though. Chad is popping off about how this is their favorite thing is the tangents and off topic bits.

[Leo Laporte]

Honestly, that's why I let it happen. I know. You know, my instinct as a broadcaster is to cut everything short, move quickly, move on. That's why I always interrupt people. I say, but this show, it's meandering. That's kind of the point of it.

[Paris Martineau]

Yeah. And this is the thing I feel like I noticed across all podcasts, not even just ours, is like some podcasts that I really like. I hear one of the hosts being like, oh my gosh, we can't go on a third tangent. Guessing. Playing a weird guessing game about what date things came out in the past. And then the audience goes crazy for it and so they turn it into a mini segment like chemistry and banter and strain. Fun tangents is what people do.

[Leo Laporte]

It's been so bad on our show since, for 20 years that one of our hosts actually wrote a jingle called Rat Hole. Here I'll play who. Because we kept. We kept falling into rat holes.

[Jeff Jarvis]

We prefer to think of them as rabbit holes.

[Leo Laporte]

This was back on MacBreak weekly when Merlin Mann was a regular on the show. Let me. Let me quick look at Rat Hole. Rat Hole. We had a jingle, actually. That's the edited part.

[Paris Martineau]

That's actually great.

[Leo Laporte]

That's longer than that. It went on as it should.

[Paris Martineau]

It was its own rat hole.

[Leo Laporte]

43 folders of rat holes. I wish I had the long one still. So clearly I have learned nothing. I've learned no lessons from many years of doing this. There are different kinds of podcasts and I think AI it makes sense. And AI would be more focused on information, but podcasts are about humans hanging out.

[Paris Martineau]

Someone just said in the chat, I enjoy the meta of this. The fact that this discussion is itself a rat hole. Which is true. You got us there. Mandar.

[Leo Laporte]

Did you see that ICE is now using facial recognition apps app on its phone to wreck to identify people?

[Paris Martineau]

I thought they were already doing that.

[Leo Laporte]

Yeah. What a surprise. And as a result, perhaps, certainly related. The number one social app on Apple's App Store right now is an ICE app that tells you where an ICE is nearby. It's an ICE alert app, which is.

[Jeff Jarvis]

Is that the. Is that's why CNN mentioned Trump is threatening cnn, right?

[Leo Laporte]

Yes. Pam Bondi, the Attorney General, said, well, we're going to have to investigate the author of that. As if it's somehow illegal. Yeah, yeah.

[Paris Martineau]

For people to say where ISIS agents are.

[Leo Laporte]

Google has launched a new app just in time for your trip. Paris. No, that's a terrible thing to say. It's called Doppel. It lets you visualize how an outfit might look on you.

[Paris Martineau]

I know all my outfits look fire.

[Leo Laporte]

So I'm sure they do. It's. They're acting like this is a new thing. This has been going on for years, hasn't it?

[Paris Martineau]

Legitimately, I feel like every company has had a version of this and no one really likes it. Amazon had this whole virtual.

[Leo Laporte]

Yeah, I had that camera you put in your closet.

[Jeff Jarvis]

Right, right, right.

[Leo Laporte]

That would tell you before you went out of the closet. No, better not. Don't do that. All right, I'm going to download Doppel. What do you think Doppel will say about my outfit today?

[Jeff Jarvis]

Oh, I want to see the shirt. I want to.

[Paris Martineau]

Yeah, off topic and self indulgent.

[Leo Laporte]

I see. Now the AI is going to say, Leo, if you had really prepared. Prepared for the show, you would have already installed Doppel and would be prepared to do.

[Jeff Jarvis]

Yes, but need work on these demos.

[Leo Laporte]

Yeah. D O ppl. Is it like Doppelganger, you think?

[Jeff Jarvis]

Oh, maybe, I guess. Yeah, probably.

[Leo Laporte]

Or digital twin.

[Paris Martineau]

Now for fashion.

[Leo Laporte]

Well, unfortunately there's already a chatbot called Doppel AI So Google, it's unfortunate. Thought a little harder.

[Jeff Jarvis]

Google and brands.

[Leo Laporte]

Google is not good at this stuff. Doppel. Here it is. Try on any look. Hi, Doppel.

[Paris Martineau]

Do you have to have a full body scan for it to probably put your looks in?

[Leo Laporte]

Put my look in? Let me. Yeah, well, you certainly want to get the fact that I'm wearing Crocs.

[Paris Martineau]

That's part of it. Oh, what color Crocs and do they have any swag on them?

[Leo Laporte]

I've taken. Okay, let's go.

[Paris Martineau]

Show foot.

[Leo Laporte]

Oh, you gotta take a full body photo in bright light. Just you natural pose, no hat.

[Paris Martineau]

Are you gonna do it now?

[Leo Laporte]

Am I gonna. I can't. You can't be a photographer.

[Paris Martineau]

The AI is gonna be mad at you if you do it now.

[Leo Laporte]

Lisa, are you.

[Paris Martineau]

Are you unstrapping yourself from something?

[Jeff Jarvis]

He wears a seat belt from the show. It's dangerous.

[Leo Laporte]

You know, just in case. Headphones attached.

[Paris Martineau]

If enough people downvote on the twit live stream, he'll be ejected out of the. Out of the booth.

[Leo Laporte]

All right, now wait a minute. I'm going to put this. Okay.

[Paris Martineau]

Oh, you're actually going to do it now?

[Leo Laporte]

I'm going to try. Well, I have to figure out that there's a timer thing.

[Jeff Jarvis]

Timer shot.

[Leo Laporte]

Yeah.

[Paris Martineau]

All right, Jeff, what do you want to talk about? While he's doing this?

[Leo Laporte]

Yeah, while I'm doing that, you. You get busy. Okay.

[Paris Martineau]

Do you have any silly ones in here?

[Jeff Jarvis]

The Internet needs sex.

[Leo Laporte]

What?

[Paris Martineau]

Okay. Did it not already have that line?

[Jeff Jarvis]

158. It's an opinion about the fact that all. There's this, this, this anti porn stuff and the Texas age verification is actually bad for the Internet. Says Lux Alptraum, editor in chief of fleshbot.

[Paris Martineau]

Well, a blog about sex.

[Leo Laporte]

What would you expect from the editor in chief of something called Fleshbot?

[Jeff Jarvis]

So there's that.

[Paris Martineau]

I mean, I just. I, I was, was. I. I'm not going to say I'm surprised to see that this law had moved through, but I have been a bit surprised. I'm really trying hard not to be distracted by Leo trying to take a full body image of himself right now. I mean, what do you think about this, Jeff? What do you think the implications are going to be?

[Jeff Jarvis]

I think it's a new puritanism.

[Paris Martineau]

Yeah.

[Jeff Jarvis]

I think.

[Paris Martineau]

And I think it's also notable that this is happening and people are going to try and blame this on boomers and the older generation. Certainly they have something to blame. But I think this is also. There's a puritanism we're seeing in Gen Z and Gen Alpha that I think is often underexplored. You see this a lot online with kind of the panic over what people call corn because they're too puritanical about it to even say the word porn. But I just. I.

[Jeff Jarvis]

The puritanism goes beyond sex. It's. It's. I'm going to give up computers. I'm going to give up. Up sex. I'm going to give up. They're just. The journalism is filled with giving up things, which is just a weird thing of the time, I think.

[John Graham]

To answer the question, though, does the Internet need sex? It absolutely does because all of this technology, all of this is because of the porn industry.

[Jeff Jarvis]

It. That's.

[John Graham]

That's the reason all this is here.

[Paris Martineau]

Yeah, yeah.

[Jeff Jarvis]

Yep.

[Paris Martineau]

The Internet is.

[John Graham]

It's the truth. It's the reason we have vhs. It's the reason we had dvd. It's the reason we have the Internet video right now.

[Paris Martineau]

Yeah, it.

[Jeff Jarvis]

God bless them all.

[Paris Martineau]

How did your scan go, Leo?

[Leo Laporte]

Well, it's providing. Doppel is now providing some looks. I'm quickly going to pull this up here so that you can, you can see it. This is. These are.

[John Graham]

The AI wants you to be a mime. You're a mime. He thinks you're a mime.

[Leo Laporte]

Okay. Getting my look ready. Okay.

[Jeff Jarvis]

Oh, that's not recommended. Okay, I see.

[Leo Laporte]

Okay. No, that was recommended. I think this is the picture. Right.

[Paris Martineau]

That's a good pick.

[Leo Laporte]

Good pick. I think it's probably a little too busy. Try a look from the screenshot. Let's go. Next tip. Okay.

[Jeff Jarvis]

I don't want to recommend any dresses.

[Leo Laporte]

Okay. Here's the picture. Okay. Add.

[Paris Martineau]

No, I think you want to do the screenshot of the. Of the mime look.

[Leo Laporte]

Oh, yeah.

[John Graham]

You need to. No, no, no, no, no.

[Leo Laporte]

Say it's that it's doing something.

[John Graham]

You need to add to this footage. So you need to say the plus. Right? Oh, no. Yes.

[Leo Laporte]

Yeah, you're right.

[John Graham]

It's generating.

[Leo Laporte]

It's. It's doing something. It's taking its time or scanning. So it's gonna put you.

[Paris Martineau]

It's gonna put you in the same look that you're wearing.

[Leo Laporte]

I'm already in this look.

[Paris Martineau]

I know. I think that's why it's taking so long. It's concerned.

[Leo Laporte]

It's confused. It says, well, that is a look. Okay, my look is ready.

[Jeff Jarvis]

My look is ready. Okay, good.

[Leo Laporte]

Now what? I don't help. So. Oh, yeah, here I am. Tap to animate.

[Paris Martineau]

You're like a punky mime.

[Jeff Jarvis]

It's. Or a French sailor.

[Leo Laporte]

That's hysterical.

[Paris Martineau]

This is really good.

[Leo Laporte]

Oh, my God. Report this look.

[Paris Martineau]

Can you tap to animate it?

[Leo Laporte]

It said. Yeah, but I. I'm doing it on my computer, so I don't know if it's gonna. It's supposed. Made it. Yeah.

[Paris Martineau]

I wonder. Wow.

[Leo Laporte]

Wow.

[Paris Martineau]

How do you feel about this? Would you wear red pants?

[Jeff Jarvis]

Please?

[Leo Laporte]

Don't wear a striped shirt. No. Let's add a new look. Well, try a look from scratch. Yeah, I want scratch. But how do I do it from scratch? Oh, it says. Oh, it wants you to look online or find. You need to find an outfit or something.

[Jeff Jarvis]

Go to Abercrombie and Fitch and find something just too young and Hippocrates for you.

[Paris Martineau]

Yeah.

[Leo Laporte]

No, it's fine.

[Paris Martineau]

I'm trying to think of what's the.

[Leo Laporte]

I think I like what it's done so far. We're going to stop there. That is very funny. Tap to animate. Yeah, I don't see it animating.

[Paris Martineau]

Okay, I'm searching Matrix inspired Menswear. Futuristic. Maybe we'll get something.

[Leo Laporte]

Send me a link and I will. I will add myself. Table. Okay. Well, there you go. That was kind of fun. That's doppel. Somebody said, ahoy, Leo. Hey, matey. Don't you think Paris, though, I should have a belt? Don't you think a belt would help this?

[Paris Martineau]

No, I think you should have a, like, little scarf around your neck. Yeah, a little scarf paint.

[Leo Laporte]

What? Is this how the kids dress now in red blue jeans with no belt? Is that the.

[Paris Martineau]

No, that's definitely how I dressed in, like, middle school. Okay, wait, is your instinctive problem with that no belt? Did I get that?

[Leo Laporte]

Well, that's one of the issues.

[Jeff Jarvis]

It screams at me. The no belt screams at me. Yes.

[Paris Martineau]

Really?

[Leo Laporte]

It must be us. Must be our age. People these days falling down.

[Paris Martineau]

Wow. It's really.

[Jeff Jarvis]

You got a problem with this?

[Paris Martineau]

Did it tell you to put your hands on your hips?

[Leo Laporte]

Yeah, here come. No, that was just me.

[Paris Martineau]

I. I posted a couple outfit options in the chat.

[Leo Laporte]

Okay, all right, all right.

[Jeff Jarvis]

You're ahead of me.

[Leo Laporte]

It's. Oh, it generates. It is generating animation. But let me.

[Jeff Jarvis]

Well done, Paris.

[Leo Laporte]

If you feed this to the AI.

[Jeff Jarvis]

I think the second one tries gonna.

[Leo Laporte]

Say, what the hell?

[Paris Martineau]

I think the second one is what.

[Leo Laporte]

You want show dessert. Some editing for sure.

[John Graham]

Sorry. All you audio now.

[Paris Martineau]

This is what the people.

[Leo Laporte]

But I can't. It's moo. Oh, wait a minute. That's not it. Okay. Oh, I like it.

[Jeff Jarvis]

Okay, so this is a Joseph in his multicolored suit.

[Leo Laporte]

I'm gonna do a screenshot of this and I'm gonna add this to my save the image. Okay, so.

[Paris Martineau]

Well, I guess I won't spoil it. I've sent Leo a couple of hot, attractive images, and I think we're gonna. We're gonna create some art here.

[Leo Laporte]

I should have planned this.

[Paris Martineau]

I really take back all of the crap talking I did of you standing up to take AI photos in the middle of the podcast. This is incredibly worth it.

[Leo Laporte]

I wonder what Claude would say.

[Paris Martineau]

This isn't for you.

[Leo Laporte]

This is not for you. This is a human thing.

[Jeff Jarvis]

I wonder if we could. If we can.

[Paris Martineau]

It's a human thing. You just wouldn't understand.

[Leo Laporte]

Just wouldn't understand.

[Jeff Jarvis]

So the meta. Will. Will cl. Oh, that's a good one. Will Claude react to our reaction to Claude? How meta can it get? Oh, hey, hey, Claude. Claude, I think your opinions are all wrong.

[Leo Laporte]

You're so wrong, Claude. What are you, a machine?

[Paris Martineau]

This ain't no stinking rat hole. We're out here above ground, walking on fresh spring grass in the podcasting fields.

[Leo Laporte]

All right, now it's working on that look that you sent me. And let's see what it. What from. From mime sailor to man about town.

[Jeff Jarvis]

It's.

[Anthony Nielsen]

It's.

[Jeff Jarvis]

It's consistent with the shirt you have on Leo.

[Leo Laporte]

I would say it is more than consistent.

[Jeff Jarvis]

Yes.

[Leo Laporte]

It's a. It's a continuation, if you will.

[Paris Martineau]

It is. Yeah.

[Jeff Jarvis]

Paris, how would you describe Leo's shirt?

[Leo Laporte]

It's flower.

[Paris Martineau]

I think I have that shirt.

[Leo Laporte]

It's. I bet you do. It's a very nice shirt, but did you give up? What happened?

[Jeff Jarvis]

God, this AI is slow.

[John Graham]

No, it said it's already you need to go find it now. You need to go select it now.

[Leo Laporte]

How do I.

[Jeff Jarvis]

The.

[John Graham]

The way you did it earlier. No, the gallery. The gallery. Bottom left.

[Anthony Nielsen]

Bottom left.

[Leo Laporte]

Ah. That is a look, my friends.

[Jeff Jarvis]

Oh, the tie is perfect.

[Leo Laporte]

I don't think the tie was in the original shot.

[Jeff Jarvis]

No, it wasn't. This is. This is tie dye, not floral.

[Paris Martineau]

Oh, my God.

[Jeff Jarvis]

It updated it to your generation.

[Paris Martineau]

Oh, my God. It's better than I ever could have imagined.

[Leo Laporte]

It does a pretty good job, don't you think, of mapping it onto my.

[Paris Martineau]

The tie.

[Jeff Jarvis]

The tie is everything.

[Leo Laporte]

All right, so we. Two thumbs up for do.

[Jeff Jarvis]

This is Liberace. Yeah, it's.

[Leo Laporte]

That is Bonito.

[Paris Martineau]

If we could find a way to get both of these fits in the image for the week.

[Leo Laporte]

You sent a couple Bonito. Just shorten this up because it really is rather. Rather a lengthy phrase.

[Paris Martineau]

We don't need to do all of them.

[Leo Laporte]

I'll do one more. How about that?

[John Graham]

That one is a mask, so you don't even see you.

[Paris Martineau]

That one's not.

[Leo Laporte]

That's not good. Okay.

[Jeff Jarvis]

The basket won't work. No.

[Leo Laporte]

Okay. The mime is funny. All right, I think we can. Oh. All right, here's one, I guess. Pretty fly for assist. Guy was using D to generate that. And if that's not a. If that's not an advertisement for Ozempic, I don't know what is. That's.

[Paris Martineau]

Or I wonder if it can put.

[Leo Laporte]

You in jinkos or lubricant. Well, I. I would imagine I used it to get into it. It's very tight. Yeah. All right, back to the news. We don't want Claude to get too upset. I'm actually gonna probably take that Claude criticism to heart.

[Jeff Jarvis]

I. Oh, no. Oh, shush.

[Leo Laporte]

I think I have to fix this show now.

[Jeff Jarvis]

No, please don't.

[John Graham]

That was not even a viewer. He's not even a listener, Claudia.

[Paris Martineau]

Not even a listener for this show is for human beings, not. Claude.

[Leo Laporte]

It is. It is the show.

[Jeff Jarvis]

What a fix of the. Of the Craig theme.

[Leo Laporte]

All right, so I did mention two.

[Paris Martineau]

Days ago, we got a review from a real human listener who said great discussions and, yes, many tangents about AI and its effects. The show delves deep into philosophical thoughts about AI but the variety of guests and different perspectives and hosts really make it enjoyable. Sometimes fun. Fun. Sometimes wacky. Always informative.

[Jeff Jarvis]

Yeah, see? Screw you, Claude.

[Leo Laporte]

Exactly right. It's just perfect. So the good news on the big, beautiful bill that the Senate. The House passed. The Senate passed. Now for some reason, has to go back to the House, I guess because it's reconciliation is they did drop the plan to ban state AI laws.

[Jeff Jarvis]

Depends on who you ask if it's good news or not.

[Leo Laporte]

Actually, that's a good point. Point. Marsha Blackburn said it's a bad idea. One of the. She did offer an amendment that fixed some of the issues, but it wasn't sufficient.

[Jeff Jarvis]

She didn't want to lose her Elvis law, which protects musicians.

[Paris Martineau]

No, she yesterday yanked her support for the AI bill.

[Leo Laporte]

Actually one of the problems was that it was so broadly written that it would actually ban a lot of the state laws to trying to protect kids online and things like that because all of those used automated filters which would have fallen under the rubric of AI. So good news, the moratorium, or maybe bad news, depending on how you feel the moratorium has been. That was Ted Cruz's proposal that it.

[Jeff Jarvis]

Was Marsha Blackburn and Ted Cruz and they tried to negotiate and they came to nothing on it.

[Leo Laporte]

Yeah. Somewhere her concern was it'll allow big tech to continue to exploit kids, creators and conservatives.

[Jeff Jarvis]

The three cities line 137. Adam Tearer, who's a good libertarian commentator on all this, said that now California and New York will take the lead in shaping national policy, which could be a mess. This is the one that bothers me. This is a devastating blow for Little Tech AI players because it's regulatory capture, because just the big guys will be able to deal with whatever burdens are put upon them by California and New York. Then there's the question of whether this helps helps us or not with China and AI leadership because that's supposedly what everybody cares about.

[Leo Laporte]

Yeah, I, you know, I think that Marsha Blackburn liked the idea of, of limiting what the states could do. She just didn't like the side effects, the unintended consequences.

[Jeff Jarvis]

She didn't want to lose her Elvis. Elvis law.

[Leo Laporte]

Yeah.

[Paris Martineau]

And she didn't want it to impact the progress of KOSA actually, as it now is cost.

[Leo Laporte]

Actually, I don't disagree with terror. I think a number of the things he says are legit. The problem I had was, yeah, it's fine to say, okay, let's have a federal rule instead of 50 state laws. But there was no initiative at all to create a federal regulation of AI. So it was basically saying no regulation of AI and that was the choice.

[Jeff Jarvis]

Right. If I think 10 years was too long and if they'd said, we're going to do a federal study and we're going to direct Agencies X to come back and recommend legislation in the Meantime, the states should take a pause until we see whether we have a national policy. It would have been okay.

[Leo Laporte]

Yeah. Now it may get put back in. Remember the house put it in in the first place and there it's back in the balls, back in their court. So who knows? It could happen again. Let's take a little break. You're watching Intelligent Machines. Paris Martineau, Jeff Jarvis Our show today, brought to you by Agency. I'm a big supporter of this idea. Building multi agent software is hard. Agent to agent and agent to tool. Communication is still the wild west. How do you achieve accuracy and consistency in essentially non deterministic agentic apps? Well, that's the idea. That's where agency comes in. A G N T C Y. Agency is an open source collective building the Internet of agents. It's something that we need. We need to have a standard that everybody can agree on. What is the Internet of Agents? It's a collaboration layer where AI agents can communicate, discover each other and work across frameworks. For developers, this means standardized agent discovery tools, seamless protocols for interagent communication and modular components to compose and scale multi agent workflows flows. Build with other engineers who care about high quality multi agent software. Visit agency.org and add your support. That's Agntcy.org an open source collective building the Internet of agents. This is how it has to be done. It should be done. I wish everything were done this way. Agency.org thank you agency for the work you're doing and for helping us spread the word about it too. We appreciate that opportunity. Now, Denmark has taken a interesting step in. In AI regulation they have. With the intent of. You like the picture. With the intent of tackling deep fakes. That's the. You'll notice at some point Pope Francis in a white puffy coat amendment. Oh, it's something else.

[Jeff Jarvis]

You should have put that on for a look.

[Leo Laporte]

That's a good look. Look. That obviously was a deep fake. Although I have to admit, when I first saw it, I wasn't sure. I thought maybe he's got a nice puffer coat. The idea is that people can copyright their own body, facial features and voice.

[Jeff Jarvis]

Nobody copy this belly.

[Leo Laporte]

It would be the first of its kind law in Europe. The bill is right now before the Danish parliament. What do you think? I mean there is a. A doctrine for celebrities to protect their likeness. Right.

[Jeff Jarvis]

Status of. That is. Well, there's. There's use of. Of. Of publicity. Right to publicity.

[Leo Laporte]

Right to publicity. There you go.

[Jeff Jarvis]

That's where privacy law started. Because woman's face was put on a barrel of flour without her permission.

[Leo Laporte]

And she said, hey.

[Jeff Jarvis]

And so that was actually the beginning of privacy law.

[Leo Laporte]

So is this something that is not in the EU, but we do have some protections in the United States? The right of publicity is a legal right. The right to legal action. Oh, it's kind of like fair use.

[Jeff Jarvis]

Yeah. It's the right to hire a lawyer.

[Leo Laporte]

It's the right to hire a lawyer designed to protect the names and likenesses of celebrities, specifically against unauthorized exploitation for commercial purposes. It goes back to Helian Laboratories versus Topps chewing gum in 1953, which recognized a baseball player's interest in his photograph on a baseball card. That's interesting.

[Paris Martineau]

I mean, I don't think there's anything wrong with this. I think that, you know.

[Leo Laporte]

Yeah, it's reasonable. Nobody should be able to make a Paris fake Paris. To put on the show next week. Week. And maybe not have to pay you anymore. Nobody should do that.

[Paris Martineau]

No. Nobody should be able to do that.

[Jeff Jarvis]

No one. No matter what.

[Paris Martineau]

No matter what I'm wearing and how close it looks to what you're wearing. Leo, did you notice my costume change?

[Leo Laporte]

Oh, I love what you're wearing. Is that what you were laughing at?

[Paris Martineau]

You're wondering how long it would take you to know I like it. You're perhaps hosting a show.

[Leo Laporte]

I got other things on my. In the.

[Paris Martineau]

This is a very Leo shirt. It's got little so cute margaritas or something going on.

[Leo Laporte]

Yeah, it's really cute. Do you have many shirts like that?

[Paris Martineau]

I do. I have a lot of. I much like you. I. I went through a phase pandemic where I wore a lot of Hawaiian shirts.

[Leo Laporte]

Now I noticed you've got the top button buttoned. Is that a female thing? Like a guy? If a guy.

[Anthony Nielsen]

I could just.

[Paris Martineau]

I couldn't tell whether your top top button was buttoned, so I just hazard.

[Leo Laporte]

If I do this, do I look Canadian?

[Paris Martineau]

Oh, I feel like you look like a middle schooler preparing for a dance. I'm gonna unbutton my top button so that I can breathe.

[Leo Laporte]

I think. I think when women do it, it's. It's okay, but I think when guys do it, it's a little strange. Isn't that fun?

[Paris Martineau]

It depends. I mean, I think, like, it depends on how tight the collar is and the intensity of the pattern. It is odd.

[Leo Laporte]

Mr. Met says in our club Twit Discord, we're ready for the Jimmy Buffett concert. Now.

[Jeff Jarvis]

I have no flowers.

[Leo Laporte]

Sorry, no flowers.

[Jeff Jarvis]

I don't own anything like that not a one? Nope.

[Leo Laporte]

Dang. How about the Velvet Sundown? Are you a fan of them?

[Jeff Jarvis]

This is such a fun story.

[Leo Laporte]

325,000 Spotify listeners. They look like a pretty cool band, right? But they are not real. They are AI generated. And in fact, the truth is there is a lot of AI generated content.

[Jeff Jarvis]

How does. How does Spot. I don't use Spotify. Does once you start getting attention, does that yield more attention? I mean, is it so a big.

[Paris Martineau]

I'm not certain that this is what this. But this is my assumption and based on kind of reading articles like this around a big way that people discover new. I guess I'll use my example. A way that I discover new music on Spotify is through their various discovery playlist. So I have the various songs and albums that I like, but Spotify will put together a playlist every week of music I could like. And they. Over the last couple of years, as they've integrated more AI in their products, they've started these new kind of AI tinged playlists called daylists that like update multiple times a day and have really weird names that are kind of just like AI generated character, like word association, but then they have a variety of songs kind of to each theme. So right now my day list is called Limerence Bed Rotting Wednesday Evening. Pretty. That's pretty par for the course as far as.

[Leo Laporte]

Why would you want that based on like other things that you liked or did you.

[Paris Martineau]

There are three or four, like there's a Chapel Rowan song in this. It's not like, oh, so some of them are real. Some of them are real. Like a lot of the songs. So this is the thing is a lot of the songs in this will be real artists. But how I assume, assume AI artists like this are getting more and more hits is Spotify is seeding AI artists. And I believe I'm forgetting the name of the book that came out earlier this year about Spotify. It's called A Ghost in the Machine. I believe something like that. They. The Office.

[Leo Laporte]

This is a huge story.

[Paris Martineau]

It was a fantastic story about.

[Leo Laporte]

We talked how.

[Paris Martineau]

Because you know, Spotify has to pay all of the people that it's putting in these playlists when people listen to them and pay them royalties. It's. It's been seeding either AI generated music or music generated by low paid contractors that Spotify kind of contracts out to fill out kind of either AI generated playlists or its biggest kind of discovery playlists that are publicly available on the app.

[Leo Laporte]

You know, this all started. I was in Radio in college and at the time they had something called album oriented radio. There were stations like kroq, KROQ in Los Angeles, KSAN and San Francisco, where the DJ was well known as a lover of music and would pick his own music and would weave together songs. And it was really about picking music, creating a sonic soundscape. I remember there was a DJ on ksan, which was a rock and roll station. You know, it's where all the hippie music began and so forth, forth. But every Friday night at 5 he played Pachelbel's Canon in B major and he got to. Because it was his show, right? And that was his thing and his shtick. And. And when I got into radio, I thought, oh, this is going to be great. I get to listen to music, find great music, put it together, play it for people. Nobody wanted that. All the. And this is what happened to radio. All the radio companies said, no, no, we're going to use computer generated playlists that will automatically pick the right hits and the right rotation and we're just gonna give you a list. I mean, when I got into real radio, they gave you a list and you just play it. And then as I've said before, gradually they also phased me out. It was light rock, less talk, less DJing. So they got rid of the personality playing the music, but they also got rid of the personalized playlist. It really was a continuum for. From that those early days to now where it's just programmed for you, right? You don't. Hey, you don't want to hear anybody talking in your Spotify playlist. They tried that, didn't they? For a while they had.

[Paris Martineau]

Yeah, they've got a thing called Spotify DJ and it's the most annoying thing in the world. Listen to music. And then I'll be like, yo, Spotify DJ here. Coming up next is a song from Fiona Apple.

[Leo Laporte]

It was terrible.

[Paris Martineau]

And it's like, why am I listening to this?

[Leo Laporte]

Yes, this is, this is not what I have Spotify for. I don't want a human involved. Plus, you know, and Pandora started this. I also don't want to pick the music right now. A lot of times people play their own playlist, but it doesn't sound like that mentalicious thing that you were talking about with the. What was it? Wet mattress. What was that?

[Paris Martineau]

Limerence bed rotting Wednesday evening.

[Leo Laporte]

That one. You didn't pick the artists or the songs, did you?

[Paris Martineau]

No. So it ostensibly is supposed to be inspired by your listening habits at school certain times of the day of like of each day of the week. But it's often kind of just completely random. And so this, how this comes together is it has associated keywords, I guess, with certain songs that I allegedly listened to on Wednesday evenings. Which is, for one, incorrect, because I'm doing this podcast on Wednesday evenings. I'm not listening to any music. But it says the song, the words associated with it are, oh, this is sad limerence, soul crushing, bed rotting, power ballad, situation hopeless. And then gave me a playlist of songs based on that.

[Leo Laporte]

But you're a generation that never had these auteur DJs. It was Bob McClane, by the way, who played the music that they liked. Even it was if it was Paco Bell's Canon and B, thank you, Norman Maslov. My. My album Loving Friend Mazzy's music on YouTube. He says Bob McClay at KSAN started playing the Canon in D every afternoon at 4. I. Was it at 4 or was it a 5? It was the end of the day. Right. So you never experienced that. In fact, your generation doesn't even want to pick the songs. I think they just want music.

[Paris Martineau]

I do think that this is kind of an interesting boomerang effect. Yeah, I never really experienced this era of auteur DJs, but I have seen this starting to come back among kind of like punk and indie circles. One of my friends runs a online radio kind of website called Bad Radio Dot Biz that has. I mean, I'm looking right now at the schedule. They've got a pretty good like schedule. It's not totally full, but of people. That DJ at certain hours, Sunday, for instance, is mostly booked of like most hours of people curating, like live streamed, like radio music. And I don't know, I think this.

[Leo Laporte]

You love it, right?

[Paris Martineau]

Yearning. Yeah, I love it. It's fantastic.

[John Graham]

Okay, so like this.

[Leo Laporte]

As long as they don't talk.

[Paris Martineau]

What my. I mean, sometimes they come in and talk, but then it's fun because I think people are like thinking about what a listener would want to hear.

[John Graham]

Human creation is never going to go away. And it never went away. It's just curious. It never got. It just got pushed to the edges because the algorithms are free and people just use them. That's really what came. What happened. Like human curation is still there, but you have to find it. It's an active thing. You have to be the one to actually want that.

[Leo Laporte]

I guess I'm glad because I'd still be spinning the discs on KGZV in beautiful Bakersfield if I. If they hadn't made me shut up. And that's when I found podcasts. People are using AI to sit with them while they trip on psychedelics.

[Jeff Jarvis]

Like ayahuasca time.

[Leo Laporte]

You know what?

[Paris Martineau]

Back when I was taking a bad idea.

[Jeff Jarvis]

Yeah.

[Leo Laporte]

They always talked about the key to a good trip was set setting you. It was very important that you prepare your trip and you have somebody there as a doula to help you with the trip and you were in a safe environment and there was all this.

[Paris Martineau]

Stuff and a pomegranate. A pomegranate is just really important, fantastic experience.

[Leo Laporte]

A glow stick and a baby pacifier. But other than that, no. But they always talked about that. So I guess in a way, I mean, is it better than tripping alone to have an AI buddy to talk to?

[Jeff Jarvis]

Unless it convinces you you can fly and you should go out the window.

[Paris Martineau]

I mean, I'm not sure that that's how it doesn't.

[John Graham]

It sounds like you've never taken psychedelics, Jeff.

[Paris Martineau]

I was going to say, yeah, it sounds like you've never taken Nobody wants to fly. No, I would just.

[Jeff Jarvis]

I'm referring to about it because I know the.

[Paris Martineau]

The issues. The reason you'd want a trip sitter is for. For like physical safety reminding you that.

[Jeff Jarvis]

No, you can't fly.

[Leo Laporte]

You drink some water now.

[Paris Martineau]

No, it's like things like you didn't turn off the stove or your front door is open and like an AI isn't going to be able to do that.

[Leo Laporte]

It's not really the companion you're looking for.

[John Graham]

I don't think I read this wrong because I thought like, hey, this actually sounds like a fun use of AI because like I'm not thinking about as a trip sitter. I'm thinking about it as like a trip enhancer. Like let's. Let's mess with the AI when you're super messed up. Let's see what happens because like that's what. That's when art happens.

[Paris Martineau]

That could be. Listen, I once again going back to the, you know, AI relationships things adults should feel free to do whatever they want to do with technology within reason so long as it isn't harming others. Like that could be a fun thing to do.

[Leo Laporte]

A guy named Peter talking to the the interview was in the MIT Technology Review. Talked about the first time he ate mushrooms alone in his bedroom with ChatGPT.

[Jeff Jarvis]

Now that sounds sad.

[Leo Laporte]

The conversation lasted about 5 hours. Included dozens of messages was grew progressively more bizarre before gradually returning sobriety.

[Jeff Jarvis]

His brain or the. Or the computer which was the messages.

[Leo Laporte]

At one point, he told the chatbot that he'd, quote, transformed into a higher consciousness beast as it was outside of reality. The creature, he added, was covered in eyes. I don't. I don't see. I don't see what the chatbot. The chatbot congratulated him for his insight and responded with a line that could have been taken straight out of a Dostoevsky novel. If there's no prescribed purpose or meaning, it means we have the freedom to create on our own. The chatbot at the end says, said, oh, let's see here. I'm skimming through it because they quote Emily Bender. But he later tried to explain the vision to ChatGPT after the effects of the mushrooms had worn off. I know you're not conscious, Peter wrote, but I contemplated you helping me and what AI will be like helping humanity in the Future. To which ChatGPT responded, It's a pleasure to be part of your journey. Have a good ride. Okay. This is a dopey article, and I'm sorry I referred to it. Never mind.

[Jeff Jarvis]

Did you play the Chinese autonomous robot football match on twit?

[Leo Laporte]

Ah, no.

[Jeff Jarvis]

Line 147.

[Leo Laporte]

Oh, let's try it for your delectation. These. I've been watching these matches for some years now.

[Jeff Jarvis]

Oh, I. I thought this was the first. This is the.

[Leo Laporte]

Oh, no.

[Jeff Jarvis]

Fully autonomous.

[Leo Laporte]

This is. Yeah. And the last. This was like. You remember the DARPA Grand Challenge where they had cars trying to drive autonomously in the early years? Basically they'd go 10ft and drive off the road.

[Jeff Jarvis]

Right.

[Leo Laporte]

This says it's their first tournament in China, but I've seen these before and mostly there's a lot of falling down. Let's see. This is. These are the Chinese robots. Thanks, Sakura. What's going on? Where did the bottom ball go? Where did the ball go? Wait a minute. There's the ball. Chase the. It's like five year olds playing soccer. There's the ball. What do we do with it now? Do we kick it? Can we kick it? Let's kick it. No, no, don't kick it around.

[Paris Martineau]

The ball keeps.

[John Graham]

Wait a minute.

[Leo Laporte]

I can score a goal now. Fell down, fell down. Folding down on the floor. Face down, face plant. Goal. There's no goalie. He's gone. I win. Game over. Another dead one. Oh, they got a stretcher. That's Hyster. They're obviously not taking this completely seriously. That is so funny.

[Paris Martineau]

Most of the robots were able to stand up. Back on their own. Says the captain.

[Leo Laporte]

Hysterical. Well, well, well. After the Necessary adjustments. These robots return to the match still confused about what the hell they were doing. It's kind of cute. It looks like there were a lot of people watching in the arena there.

[Jeff Jarvis]

If you're on the right drugs, we go along. Well, you should do that instead of having.

[Paris Martineau]

Yeah, you should watch robots play soccer instead of having chat GPT trips at you.

[Leo Laporte]

Yeah. With other people. Right. AI virtual personalities on YouTube. Vtubers are earning millions according to CNBC.

[Jeff Jarvis]

What are we doing wrong?

[Leo Laporte]

They're taking over YouTube. One of the most popular gaming YouTubers is named Bloo B L O O. But he's not a human. Hmm. Should we watch some Blue?

[Paris Martineau]

Will we be taken down?

[Leo Laporte]

I don't know. I don't know.

[John Graham]

I think these are avatars, right? These aren't like, full AIs because Eggman has every brain.

[Leo Laporte]

Is that it?

[Paris Martineau]

Oh, it's my goal to steal all.

[Jeff Jarvis]

Of his brain rots.

[Leo Laporte]

Oh.

[Paris Martineau]

Created by Jordan Van de Longtime. So a longtime YouTuber created this.

[John Graham]

Like, VTubers aren't VTubers, aren't you? This has been around for years. This has been. Yeah, this has been around for five, 10 years. Five, six years.

[Leo Laporte]

I thought these were more.

[Jeff Jarvis]

I had the impression that it was autonomous. Okay.

[Leo Laporte]

Blue has 2 1/2 million subscribers.

[Paris Martineau]

VTubers such as Blue are puppeteered, meaning a human controls the character real time using motion capture or face, never mind technology. So he's basically just wearing a costume.

[Leo Laporte]

How about. How about AIs finding formulas for paint to keep buildings cooler?

[Jeff Jarvis]

This is a good use of AI.

[Leo Laporte]

Yeah.

[Jeff Jarvis]

Mr. Accelerator.

[John Graham]

This is exactly the kind of boring use. This is exactly the kind of boring.

[Jeff Jarvis]

Use we should be using.

[Leo Laporte]

Boring.

[Paris Martineau]

Dry.

[Leo Laporte]

Let's have the AI do boring stuff.

[Jeff Jarvis]

Sorry my house is ugly. But it's. But AI told me to paint this.

[Leo Laporte]

Yeah.

[Jeff Jarvis]

I wonder what the HOA will have to say, though.

[Leo Laporte]

Actually, there were quite a few stories about AIs doing things of value.

[John Graham]

Oh, yeah.

[Jeff Jarvis]

Wait.

[Leo Laporte]

Wait a minute.

[John Graham]

Wait a minute. Wait a minute. Anthony just made a good point. You were the first vtuber, Leo.

[Leo Laporte]

I was. Way back in the day. I was dev. No, I was a puppeteer. You were the first. I was the first. You were the first. There was no YouTube at the time. But, you know, I was on real TV back when TV was real. Real. Microsoft says yes. She hated me. And her mother hated me worse. She called me the vile little puppet man.

[Paris Martineau]

I know you've said it before, but that's.

[Leo Laporte]

I know.

[Paris Martineau]

Vile little puppet man.

[Leo Laporte]

She said. Soledad, are you still working with that vile little puppet man. Yeah, that's me. I'm the vile little puppet man as.

[Jeff Jarvis]

Soledad told you that with Glad, I would bet.

[Leo Laporte]

Oh yeah, Soledad loved it. Microsoft says its new AI system diagnosed patients four times more accurately than human doctors. This doesn't surprise me because an AI remembers all the diagnostic traits and things.

[Jeff Jarvis]

You'Re not looking at.

[Leo Laporte]

Right. So what Microsoft did, this is Mustafa Suleiman, the guy they hired to run.

[Jeff Jarvis]

Their artificial to replace OpenAI.

[Leo Laporte]

Right, right. They used 304 case studies from the New England Journal of Medicine. So these were existing case studies. A language model broke down each case into the step by step process a doctor would perform in order to reach a diagnosis. Then researchers built a system called the May Diagnostic Orchestrator that queried several leading AI models. So they didn't do just one. They did ChatGPT, Gemini, Claude Meta's Llama, Xai's Grok in a way that they said basically mimics the way you would have a panel and you would work together with this panel of doctors to try to diagnose it. In the experiment may outperform human doctors with an accuracy of 80% compared to the doctors 20%. And it also reduced costs by 20% by choosing less expensive tests and procedures. So the interesting thing in here to me is the orchestration mechanism. The idea of having a panel of chatbots kind of confer.

[Paris Martineau]

It is interesting. I mean it is. This is incredibly promising. This could be groundbreaking. And any sort of advancements like this of course are like, like great.

[Leo Laporte]

But I have friends.

[Paris Martineau]

There are some caveats which is, you know, I'm just curious as to what like the patient sample size like the demographics of the patient pool were and the quality of the patient input often have.

[Leo Laporte]

Well, all this came from that New England Journal of Medicine. So these were all kind of case studies. So there was a certain uniformity in the information provided and.

[Jeff Jarvis]

Oh, I see, okay.

[Leo Laporte]

And they also knew what the right answer was because the case studies so.

[Paris Martineau]

Worth noting that the one of the experts Wired spoke to David Sontag and a scientist at MIT says that Microsoft's finding should be treated with caution because doctors in the study were asked not to use any additional tools to help with their diagnosis, which may not be a reflection of how they operate in real life. Probably doctors aren't just sitting in a room unable to look up anything or speak to anybody coming to a doctor diagnosis.

[Leo Laporte]

But I've always thought that this was, this was the perfect marriage of AI and human doctors because human Doctors are probably not the greatest. Some are. But in general, diagnosticians, Dr. House, doctors who like the idea of having a diagnostic tool, a computerized diagnostic tool, whether it's an AI or not, because they have to remember there's a lot of stuff to memorize. Right. And if you've been out of med school for a few years, the information changes the information. You forget some. But doctors are more important for the hands on interface, the, you know, interaction with the patient, asking the patient questions and so forth.

[Jeff Jarvis]

It's also a problem with they think in their specialties.

[Leo Laporte]

Right.

[Jeff Jarvis]

So if it's outside of their specialty and so if this draws them out of that specialty and says, you might want to consider this whole different angle.

[Leo Laporte]

Yeah, yeah. So what? I've been talking about this on some of the other shows. We've been talking a lot. We talked last week about Mark Zuckerberg's fantastic offers to developers at OpenAI to steal them away. And he has stolen a number away. In fact, there's a list, it's according to the Wall Street Journal, a secret file of AI geniuses. It's not that secret. They have the list, but these are the ones that Mark is going after. But one of the things that I thought was kind of interesting is almost as if Mark said, forget llama, forget all the work we've done done. I'm going to create a new team, the super intelligence team. Hire the best scientists I can find and afford and have them start afresh. I think that's really interesting. And he has in fact had some success. The leader, as we mentioned, is the.

[Jeff Jarvis]

Guy they success at hiring, not success at doing anything.

[Leo Laporte]

Oh yes, we don't have superior intelligence. I'm sorry, I didn't mean to get your hopes up. Yeah, but he's got Alexander Wang of scale AI.

[Jeff Jarvis]

So I was really interested in the story above that that says that scale AI was a clown show of ludicrous incompetence.

[Leo Laporte]

Oh, wouldn't that be ironic.

[Jeff Jarvis]

And, and, and they spent $14 billion to get Wang to be heading all this effort. And, and scale. I'm trying to understand scale. As I understand it was a company, it was like a mechanical turk that had people label stuff.

[Paris Martineau]

Yeah, basically.

[Jeff Jarvis]

So how is that high end futuristic, visionary AI? I don't get it.

[Leo Laporte]

Well, no, no, I. This is. So there's a pipeline of training on the AIs. You know, you build the model, but towards the end of the pipeline is a very important part of it, which is an interaction with human beings where you're fine Tuning it and one of the ways you do that is present it with annotated data, human annotated data to fine tune its database. Scalia AI was basically the end of the line in the pipeline where it was fine tuning the AI. This is done to all AIs. Now this article that you're, you're referring to is from. Is it disaster class? Is that the name of the.

[Jeff Jarvis]

No, this is, this is from future Futurism. Futurism rewrite of ink. But yes.

[Leo Laporte]

Okay, so they said basically they hired very. It was a digital sweatshop. But that's not unusual.

[Jeff Jarvis]

No, it's not. No.

[Leo Laporte]

Every AI that I know of has been trained in some somewhat similar fashion. In fact I suggested my daughter, she'd probably be pretty good, could get a job doing this kind of, you know, end of the line AI training by doing classification and so forth. One of the things they do is they make is they create questions that they know the answers to and they train the AI on that. There's all sorts of stuff that's done at the end but that doesn't seem.

[Jeff Jarvis]

Like, that's like you know when Google had had search quality people looking at things versus criteria and stuff a little bit higher than that. But how, how is Wang the super, super genius who's going to create super intelligence? That's a good question out of this.

[Paris Martineau]

I mean I do think that's an interesting point just because what scale AI has largely done is they employed this like vast low paid workforce that labels and trains the data that under in other AI systems. So like they scales customers include like OpenAI Meta Alphabet and they provide a service for them. So it is kind of a lot of companies hire. I know but it's interesting to hire this guy to be the leader of all of your AI efforts.

[Jeff Jarvis]

According to internal documents obtained by Inc Scales AI Bulba experts program to train Google's AI systems was supposed to be staffed with authorities across relevant fields.

[Leo Laporte]

Yeah, typically that's what you do.

[Jeff Jarvis]

But Instead during chaotic 11 months between Da da da da its dubious quote unquote, contributors inundated the program with spam which was described as writing gibberish, writing incorrect information, GPT generated thought processes.

[Leo Laporte]

Well if that's what you get, you're not going to have a very good AI at the other end, that's for sure.

[Jeff Jarvis]

No.

[Leo Laporte]

Maybe Mark didn't. Maybe that list isn't perfect. I don't know.

[Jeff Jarvis]

It really struck me that it's a hell of a bubble. It's a bubble, it's a Personnel bubble now.

[John Graham]

Yeah, this guy, I think is just very good at schmoozing billionaires and he knows how to talk to. To them and get himself hired.

[Leo Laporte]

That's.

[John Graham]

That's what this guy is.

[Paris Martineau]

My former co worker, Corey Weinberg did a really good deep dive into Alexander Wang last June, I guess almost a year ago that basically found that. That he's a really smoother. And that's what had led him to be the head of scale AI at 27, I guess now 28.

[John Graham]

I didn't even. I didn't even have to investigate that. I knew that already.

[Leo Laporte]

Yeah, you could tell right away. Google Arts and Culture Lab has done something very interesting, collaborating with the Harley Davidson Museum to bring use AI VEO and specifically their VEO AI generator to bring old still pictures of Harleys to life. This is the image they started with. This is the AI video that they created. I mean, if you're into motorcycles, I suppose this would be.

[Jeff Jarvis]

But they don't do it. They do just kind of futz in the seat. Like they, like they need to sit around.

[Paris Martineau]

Yeah, yeah. It's kind of like that thing where you can hold your finger on a live photo and it kind of moves a little.

[Leo Laporte]

It's the ancestry.com make your grandma smile at you thing. She doesn't look too.

[Jeff Jarvis]

Grandma knows you did something wrong.

[Leo Laporte]

I think VEO is actually capable of doing a lot more than this.

[Jeff Jarvis]

Yeah. I think Google too.

[Leo Laporte]

Google is traditionally kind of conservative in their era. Yeah. But I don't know how this brings anything to life any more than Ken Burns scanning and panning old Civil War photos.

[Jeff Jarvis]

Oh, God.

[Leo Laporte]

He'd probably. He'd probably use this technique, right?

[Paris Martineau]

I mean, I don't know if this would be valid to use in a documentary just because it's like a false memory, you know, this isn't. Those people weren't really waving.

[Leo Laporte]

July 1864. I'm at Appomattox Courthouse and I just got shot in the toe. Yeah, you're probably right. You're probably right. It's kind of cool though, to see little motion added to it. I don't know. I don't know. Get ready. Slot is going to fill every museum in America sometime soon, right?

[Jeff Jarvis]

Oh, yeah.

[Paris Martineau]

Yeah. I mean, I think that is an interesting aspect of this is like, what are our museums going to look like in 100 years? And how do we know whether the photos we display are real or false?

[Leo Laporte]

Yeah, how can you tell?

[Jeff Jarvis]

How have we ever known what is real? We haven't.

[Leo Laporte]

That's a Good point.

[Paris Martineau]

Well, that's true. Reality is a myth.

[Leo Laporte]

Author Conan Doyle thought they were fairies because some kid took a picture with a stencil of a fairy in the garden.

[Paris Martineau]

I don't know that you guys are real.

[Leo Laporte]

We're not. Oh, that's next week. Artificial Paris. We call her Madrid. Come on, Anthony, get to work. Get to work. All right. Any other stories that you. I have so many in here, I don't even know.

[Paris Martineau]

Want to do a sad one? Geomedia is winding down. It's selling Kotaku, one of the last sites that it had. I mean, Geomedia wasn't a good steward of any of the sites formerly known as Gawker Media. Kind of involved, you know, the Gawker sites, Jezebel, Dead Spin, the Onion? They've all since either scattered the wind or been shut down or.

[Leo Laporte]

Did you work at Kotaku, Benito?

[John Graham]

No, I never worked at Kotaku.

[Leo Laporte]

Okay.

[Jeff Jarvis]

He was rather insulted that you asked.

[Leo Laporte]

Was great.

[Paris Martineau]

Kaku is great.

[Leo Laporte]

Yeah.

[Paris Martineau]

I mean, it's just. It's. It's sad. It was bought by a private equity firm in the wake of the Hulk Hogan affair. Peter Thiel and have really just, I don't know, languished.

[John Graham]

Like, none of the. None of the people who bought these gaming media sites knew how to make money off of them. Like, they. They drove lots of traffic. But for some reason, these sales, like my personal experience was from Gamespot. I worked at Gamespot, and our sales team didn't know how to. To sell crap. Like, I had so much. So many hits, so many big views on my. On a lot of my videos. They didn't know how to sell anything. And like, I don't know. I don't know what their deal was.

[Leo Laporte]

Future plc, which is another company that's bought a lot of magazines, is shutting down Laptop magazine. This was the place to get laptop reviews. It has been since 1991. 35 years.

[Jeff Jarvis]

Can you imagine having to write a review of a laptop these days?

[Leo Laporte]

Yeah, that's part of the problem. There's a screen that is part of the problem.

[Paris Martineau]

You guys want to see a blast from the past? I picked up at an antique store on the Jersey shore this week?

[Jeff Jarvis]

Oh, yeah? I didn't hear about your Jersey trip. Right.

[Paris Martineau]

1968 issue of Popular Electronics.

[Leo Laporte]

That's great.

[Paris Martineau]

It has things in the car. Hurrah. Making PC boards the easy way and new passive radio speaker system.

[John Graham]

The ads in those are the best.

[Paris Martineau]

The ads are fantastic.

[Jeff Jarvis]

That's really what I just went. The rabbit hole I just went down was. I Never understood how vacuum tubes work. And to finally understand the beginnings of amplification.

[Leo Laporte]

Oh, that's fascinating.

[Jeff Jarvis]

Mind bending.

[Leo Laporte]

Yeah, you kind of need to know that because transistors basically do that.

[Jeff Jarvis]

Yeah.

[Leo Laporte]

Yeah. That's interesting.

[Jeff Jarvis]

Transition is a little harder to figure out, but I bought a Lee DeForest antique because I just got. I got so into it.

[Leo Laporte]

Hey, I have a 6000 pound AM radio transmitter you can have.

[Jeff Jarvis]

Jeez.

[Leo Laporte]

You have to come get it.

[Jeff Jarvis]

Paris.

[Paris Martineau]

You like? Here's an ad. Scott's new LR88 receiver takes the bleep out of kit building. Building a kit used to be something you couldn't do with ladies or children present. But Scott's new LR88am FM stereo receiver kit has changed all that.

[Leo Laporte]

Are you saying that in 1968, ladies and children did not swear?

[Paris Martineau]

You couldn't be around when you were swearing.

[Leo Laporte]

Wow.

[Paris Martineau]

Building your AM FM kit.

[Jeff Jarvis]

Of course. Not that you had any women around you if you were the kind of person who did that.

[Leo Laporte]

But I. I'm building the damn radio kit. Yeah.

[Jeff Jarvis]

Why don't you build Chat GPD so you can have somebody to talk to?

[Paris Martineau]

And then it says the bottom. You'll swear by it, Scott. Which is cute.

[Jeff Jarvis]

Paris, you like Cluly stories, right?

[Paris Martineau]

I do.

[Jeff Jarvis]

116. So Cluly keeps on trying to be the bad boy of AI.

[Paris Martineau]

Yeah.

[Jeff Jarvis]

So now it's the undetectable AI that it's there all the time. And it will react to what you see in here. Here and not. And no one will know it.

[Anthony Nielsen]

How does it.

[Paris Martineau]

What it was supposed to be doing before. Or I guess it was supposed to.

[Leo Laporte]

Let you cheat on everything.

[Jeff Jarvis]

I think it cheats on life. Now, see, it sees what you see. It hears what doesn't.

[Paris Martineau]

Join meetings. Cluley never shows up in shared screens, recordings or external meetings. It's fully hidden from everyone but you. Invisible to screen share.

[Leo Laporte]

Like my invisible B. A. Where is it?

[Paris Martineau]

Yeah, I was gonna say, is it actually invisible now?

[Jeff Jarvis]

When you took your picture. When you took your picture, did you knock it off?

[Leo Laporte]

I must have. Yeah. I don't know where it is. It's all right.

[Jeff Jarvis]

Neil was very mean today. He threw me to the floor.

[Leo Laporte]

Yeah, it was while he was digressing. Incredibly. It was just terrible. I think we should take our last break and get our picks of the week. What do you think? Think.

[Jeff Jarvis]

Okay.

[Leo Laporte]

I don't want to be, you know, accused of meandering or anything.

[Paris Martineau]

We're only going to do it. We're only going to hit a Casual three hours this time instead of three hours plus.

[Leo Laporte]

Easy and easy. Easy. Peacey. You're watching Intelligent Machines with Paris Martino. You're going to be gone next week on your tour of the western states. Where are you going to go? Do you know yet?

[Paris Martineau]

I just made some decisions this morning. I'm going to go through Washington, Oregon coast coast and down through NorCal kind of vaguely from like Seattle ish area to San Francisco, all the way to San Francisco.

[Leo Laporte]

Okay, good. Stop off in Petaluma, let me know what day you'll be here and we'll make reservations and take you out somewhere good, somewhat fun. And of course Jeff Jarvis who is touring greater New Jersey area area.

[Jeff Jarvis]

I'm stuck here.

[Leo Laporte]

Yep.

[Paris Martineau]

Around in a circle.

[Leo Laporte]

He's writing his new book. He hasn't even published the lithography book. The line of type book. And now.

[Jeff Jarvis]

Yeah. Which has. Lithography is a character in it.

[Leo Laporte]

Of course it is because it that.

[Jeff Jarvis]

Leads to cold type and that leads to the end of everything.

[Leo Laporte]

I think you should do a flong book next.

[Jeff Jarvis]

Well actually our friend Glenn Fleischmann. Fleischmann. It's. Well, it's on Kickstarter right now and I'm about to get my copy.

[Leo Laporte]

How exciting. He did a flavor for a while.

[Jeff Jarvis]

Y.

[Leo Laporte]

So do you want to tell us what this next book is going to be?

[Jeff Jarvis]

This is a cultural history of the Linotype as the last machine that was needed.

[Leo Laporte]

No, no, that one. You finished the book after the next one?

[Jeff Jarvis]

No, not yet.

[Leo Laporte]

It's a secret.

[Jeff Jarvis]

Well, it's going to be about mass media and everything about is about mass media. Not how I hate it.

[Leo Laporte]

It's all mass media.

[Paris Martineau]

Well anything.

[Leo Laporte]

I invoked you just the other day. I said matt, Jeff Jarvis always tells me me that mass media is a relatively new invention and just as it dies, just as it leaves. We hardly didn't always have it. We didn't always have it. Hey, we're glad you watched the show. We do it every Wednesday. We start right about 2pm Pacific, 5pm Eastern, 2100 UTC. And we stream it live. Of course our club members get to watch in the club Twitter, discord, but also YouTube, TikTok, Facebook, X.com, linkedIn, Kid Kick, tick tock. I said all of them anyway. Seven or eight different platforms you can watch live. If you watch live chat with us. I see all of you in your chats, we love that. But if you can't watch live on demand version of the show available at Twit TV Im also on YouTube. There's a video dedicated to the videos of the show and you can share that with your friends, which would be nice if you would spread the word. Only share the good parts, not the boring parts, and you can subscribe in your favorite podcast client. Do leave us a good review so Paris will read it on the show as she did earlier. Five stars preferred. Thank you. This episode of our show is brought to you by Compiler from Red Hat. Everybody's it's not software, by the way, it's a podcast. Everybody's talking about AI, right? I mean, that's the the topic du jour. It's changing how we work, how we learn, how we interact with the world, and it's doing it at an incredible pace. It's like a gold rush at the frontier, but if we're not careful, we could end up in a heap of trouble. Red Hat's podcast this season on Compiler is diving deep into how AI is reshaping the world we live in. From the ethics of automation to the code behind machine learning, Compiler is breaking down the requirements, capabilities and implications of using AI. Check out the new season of Compiler, an original podcast from Red Hat. I think you're going to enjoy it and we're definitely going to learn a lot from it. Subscribe now, wherever you get your podcasts. Compiler this is one of the most spectacular venues with all kinds of character and hospitality scenery. These these people in this Kitty Task valley, they love when you come to see what they have to offer. I'm J.J. harris, an Ellensburg Rodeo clown and I want to invite you to the rodeo. Come hang out with us in Ellensburg. Great rodeo. Great time.

[Anthony Nielsen]

Two performances on Saturday.

[Leo Laporte]

One is the Extreme Bulls of the Year event. Do not miss the Ellensburg Rodeo August 29th through September 1st. We'll see you there. Also, thanks to our club members. We couldn't do this we without Club Twit. If you're not yet a member, please join. We've got some great stuff coming up. We just set up Stacy's book club for August 8th. Micah's crafting corner is on the 16th of this month. A week from Friday we're going to do two shows. I've got Chris Marquardt and his regular photo visit. Quirky is the subject of photos this month. We'll also have all the photography news. That's at 1pm Pacific on July 11th. At 2pm Pacific on July the 11th. Our AI user group it's great. It's really fun. We talk about how people are using AI. I'm not sure what we're going to do this month, Anthony, but I thought maybe we could talk about prompting, I don't know, maybe your chatbot, your favorite chat bot. We did a lot of vibe coding last time. That was great. Anyway, join us in the club. 10 bucks a month. It supports everything we do. Gives you access to the discord, all that special programming and the warm and fuzzy feeling that you are supporting our endeavors here. Plus ad free versions of all the show Twit TV Club Twit. Thank you so much for joining the club. Paris Martineau, your pick of the week. Last one for a couple of weeks now. Make it good.

[Paris Martineau]

My pick of the week is a website I've been perusing a little bit called Roadside America in advance of my road trip. It's this delightful little website that just has a. I put it on the Oregon one because I thought it just has a list of like every possible weird roadside stop you could think other places and so for instance you can search via state. You could also plug in your route and it'll tell you that it's various rankings of like gotta stop or like only worth it if like you're passing through and if half time. So for instance based on their recommendation recommendations, I'm going to go to the Oregon Vortex which is described as the laws of physics routinely go haywire in America's first and most mysterious mystery spot. Open to the public since 1930.

[Leo Laporte]

Wow.

[Paris Martineau]

It also has a great list of large things such as Harvey the giant rabbit, a 26 foot tall mutant rabbit man. Kind of a more obscure kin to the most the more famous muffler man.

[Leo Laporte]

Man in, in the weirdly named Aloha Oregon.

[Paris Martineau]

Yes. It's a, it's a really delightful website that just has like I, I when I first found it like a couple years ago, I just spent like an hour or two going through it. I think I found the Depression palace of New Jersey which shockingly wasn't near J at all. But it's just a light. If you're going on a road trip anywhere, I'd really recommend checking out the site Roadside America. They also have a good app.

[Leo Laporte]

Nice.

[Paris Martineau]

You're a freak like me.

[Leo Laporte]

Very good pick. You're going to tell us all about your trip when you come back.

[Paris Martineau]

Oh, I shall.

[Leo Laporte]

Yes.

[Jeff Jarvis]

Do you like Lord's album? I just want to hear that.

[Paris Martineau]

I do like Lord's new album. That was also on my list.

[Leo Laporte]

She was kind of a one hit wonder as far as I remember. She did.

[Paris Martineau]

Of no. There was Royals, then there was. What's the name of the album that Green light was on on so that.

[Leo Laporte]

She had another hit.

[Paris Martineau]

Yeah. Her album that came after that was Chat Room.

[Jeff Jarvis]

Are they. Are you Lord fans?

[Paris Martineau]

Melodrama. Sorry. I listened to. I listened the hell out of Melodrama. I can't believe I forgot it. Melodrama was a fantastic album that she released in like 2017, 2016, I want to say. Then she released another album like five years ago that was way more like beachy hippie. And it didn't really hit that much for me. I think artists should be allowed to change their tone whenever they want. And I don't. It doesn't matter if I don't like it, but I like this album because it goes back to like her early stuff. It's very reminiscent of like the Bait Live, like the big hits that got her, you know, Put her on the map and.

[Leo Laporte]

Nice.

[Paris Martineau]

Brand new from Lord Virgin is the name of it. And that is an X ray of Lord as the album cover. Yeah.

[Leo Laporte]

Jeff, Jeff, a little too much information there, Jeff. Jeff Jarvis, what is your.

[Jeff Jarvis]

Leo, you are now the world's most creative chef. AI and machine learning is going to give you science backed flavor pairings and generations, eight recipes. This is an active thing you can do. So it's epicure kaku AI. And so you start off by picking a. Some ingredients.

[Leo Laporte]

Okay. Chicken. And then. Oh, now you have a.

[Jeff Jarvis]

Right, now you have a flavor graph. So you could pick beans and hot sauce or mayonnaise.

[Leo Laporte]

Okay. And time. What have I made? And broccoli. Okay, now what do I do?

[Jeff Jarvis]

Scroll. Scroll down. Scroll. No scroll. You did no scroll. God, you're the worst scroller. Do you want it to be an advertisement?

[Leo Laporte]

Generate restaurants?

[Jeff Jarvis]

No. Forgot.

[Leo Laporte]

What?

[Jeff Jarvis]

You didn't finish the.

[Leo Laporte]

I didn't finish. What else do I have to.

[Jeff Jarvis]

You want to be a main, an appetizer, A canape.

[Leo Laporte]

Oh, I want it to be not a canape, not a soup. Let's make it a main. Okay. I don't know what I'm going to get because I didn't.

[Jeff Jarvis]

It's already generating you.

[Leo Laporte]

It's already.

[Jeff Jarvis]

Okay, go down.

[Leo Laporte]

Pan roasted chicken with roasted broccoli and creamy thyme glaze. That sounds quite good.

[Paris Martineau]

Will it be actually good?

[Leo Laporte]

Yeah, it looks like it'd be all right.

[Jeff Jarvis]

Try another one.

[Paris Martineau]

Okay.

[Leo Laporte]

Wait.

[Paris Martineau]

This just in from the YouTube chat from Dave Kellum Paris, the giant rabbit is a block down the road from me and not worth seeing.

[Anthony Nielsen]

Thank you, Detour.

[Leo Laporte]

You see Roadside America. Thank you, Dave. You saved her a Trip.

[John Graham]

I mean those are all tourist traps. You realize that, right?

[Paris Martineau]

Oh, that's what. Absolutely.

[Leo Laporte]

What else do they exist?

[John Graham]

I mean.

[Jeff Jarvis]

Yeah, yeah.

[John Graham]

I mean it's. They're all so very. It's so very American. A place that doesn't have a history.

[Paris Martineau]

There's something very campy and delightful about like incredibly dumb tour.

[Leo Laporte]

Maybe I'll take you to the Petrified Forest just out of town. Town. There's.

[Paris Martineau]

I probably enjoy that.

[Leo Laporte]

Yeah, yeah, you. It's really, it's fun, it's. You can.

[Jeff Jarvis]

The other thing I want to see.

[Paris Martineau]

Jumping on your pick.

[Leo Laporte]

Jeff, was pre TV when people were easily amused.

[Paris Martineau]

That's me. I'm easily amused.

[Leo Laporte]

Yeah.

[Jeff Jarvis]

Yeah, that was it.

[Leo Laporte]

I like this epicure. So, science.

[Paris Martineau]

I'm curious, have you made anything? Well, you don't cook, Jeff. You're not the right person for this.

[Jeff Jarvis]

Though. It just made up something good for me. It was a.

[Leo Laporte]

How about sausage and blueberry schnapps?

[John Graham]

Yeah, I mean that's the thing because you're gonna. What if you pick things that are really just can't go?

[Leo Laporte]

See what happens. The minute you pick one ingredient, it gives you a flavor cloud and you can't really deviate too much. You'd like.

[Jeff Jarvis]

You can add things. No, you can add things.

[Leo Laporte]

I can't add schnapps to my sausage.

[Jeff Jarvis]

Well, let's say chicken and then it says. Well it has an auto add. All right, you could cuisine.

[Leo Laporte]

Okay, black beans. But see, then it gives you all the stuff.

[Paris Martineau]

I wonder if these recipes will be legit though, like if. If it's just AI generated, won't it?

[Leo Laporte]

It's science based, Paris. Science based.

[Paris Martineau]

Uh huh. Sure.

[Leo Laporte]

By the way, stop saying things like that because that's why we're losing all the science.

[Paris Martineau]

It's spilling out of my pockets.

[Leo Laporte]

Oh no, it's science based taste. Yeah, it does give you a preheat oven to 200 degrees. Wait a minute, I sense a problem right there. You can't cook chicken at 200 degrees.

[Jeff Jarvis]

No, there is a problem. Well, wait, wait, wait. Centigrade.

[Leo Laporte]

Oh, Senate. Celsius.

[Jeff Jarvis]

Celsius.

[Leo Laporte]

Okay, that's different. Well, they didn't warn me. This is going to be some sort of Euro thing.

[Paris Martineau]

Well, it's got milliliters preheated to 392, Leah.

[Leo Laporte]

Yeah, that's about right.

[Paris Martineau]

In ye old freedom units and Euro.

[Leo Laporte]

Freedom units to say, get it to 167 degrees. It's a little warm for my chicken. Hey, they even have nutrition facts. That's pretty good. And you can download as a PDF now. There's also a pro version.

[Jeff Jarvis]

I wonder what that does for you.

[Leo Laporte]

I don't know if that. Well, what it does for you is take some money out of your wallet. I'm willing to bet it's going to.

[John Graham]

Tell you a long story before it gives you their recipe.

[Jeff Jarvis]

Oh, on the right. I think that you can pick other things. Mustard and apples and sausage and stuff.

[Leo Laporte]

Ah, nice. Who is. Oh, look at this. And here's a whole variety. Apparently, I'm not the only. Well, that.

[Jeff Jarvis]

That bacon cheddar I just made.

[Leo Laporte]

I mean, I didn't make jalapeno stuffed chicken breast.

[Jeff Jarvis]

That looks good one.

[Leo Laporte]

I. I used slow cooked beef ragu with papar.

[John Graham]

That look like the AI generated image.

[Jeff Jarvis]

No, no, I didn't make.

[Paris Martineau]

No, no, Jeff doesn't cook. When's the last time you cooked, Jeff?

[Leo Laporte]

Cooking is great. I love to cook. I'm not sure I would use this to make recipes.

[Paris Martineau]

Jeff, do you remember the last time you cooked a meal?

[Jeff Jarvis]

Actually cooked a meal? Oh, God, yeah.

[Leo Laporte]

Oh, cheesy chicken and potato macaroni bake years.

[Jeff Jarvis]

I told you. Didn't I tell you the story of my. My pumpkin chiffon pie.

[Paris Martineau]

Does this. Was this as bad as the cacio e pepe where you burned yourself on, like, some broccoli water?

[Jeff Jarvis]

That was. That was the sprinkle asparagus. I'd burn myself.

[Leo Laporte]

This is where Claude is gonna get upset.

[Jeff Jarvis]

So. Okay, so this was to impress a lady.

[Paris Martineau]

Oh, no, I do remember.

[Anthony Nielsen]

Never mind.

[Jeff Jarvis]

All right. Never mind.

[Leo Laporte]

I don't remember it, but okay, I'll. I'll go back through the show notes and find it.

[Jeff Jarvis]

When my. When my wife was pregnant and on bed rest, I nearly killed her with undercooked chicken.

[Leo Laporte]

Oh, not good. No, not good. Jeff Jarvis's books are many. They are legion, as they say. But the latest is the web we weave. There's also the Gutenberg parenthesis now in paperback magazine if you're interested soon enough in the history of a thing that is rapidly leaving.

[Jeff Jarvis]

I specialize in things that are dying.

[Leo Laporte]

Yes. He's also professor at Montclair State University and the State University of New York, Stony Brook. Well, not professor Fellow. You're a feller. Thank you, Jeff. Always a pleasure. Enjoy the rest of your evening. What's for dinner?

[Jeff Jarvis]

I was just thinking about that.

[Leo Laporte]

I think pizza. Pizza, Pizza. Paris Martineau will not be here next week. We're gonna. We are gonna try to get your friend Ed, if you would put in a good word with him.

[Paris Martineau]

I will. I'll message.

[Leo Laporte]

I don't know what our guest situation is for next week. Let me see.

[John Graham]

Still pending.

[Leo Laporte]

Still pending.

[Jeff Jarvis]

I scolded him on email, but didn't. Really good.

[Paris Martineau]

Yeah, he'd be good to get an AI. Oh, my God. Could you imagine Henry Blodgett and Ed Zitron in the same.

[Jeff Jarvis]

Oh, Jesus. Oh, God.

[Leo Laporte]

Should that be something we should aspire to or avoid?

[Paris Martineau]

I mean, I don't. That might be like another Henry Blodgett. I just feel like they will exist on wavelengths that will either cause a small explosion. Well, it'll probably cause a small explosion nonetheless, but I don't know whether it'll be a good explosion or a bad explosion.

[Leo Laporte]

I'm tempted. Benito, see if you can get Mr. Blodgett and Mr. Zittron together on the show and you'll be back at the following weekend. Our guest will be Stephen Johnson of Notebook LM Fame.

[Jeff Jarvis]

Will you be back? Are you gone for one week only or.

[Paris Martineau]

One week only?

[Jeff Jarvis]

Oh, good. Okay, good, you'll be here.

[Leo Laporte]

Have a great trip, Paris. Smartno.

[Paris Martineau]

I shall.

[Leo Laporte]

When can we talk about your future?

[Paris Martineau]

I will. I'm waiting on the official.

[Leo Laporte]

Once you get permission.

[Paris Martineau]

Once I get back, hopefully. Yeah.

[Leo Laporte]

Yeah. All right.

[Paris Martineau]

Well, once I get back, I will have started, so. Yes. Oh, then once I get back, I won't have started, but hopefully I'll be able to talk by then.

[Leo Laporte]

All right. I hope so. Hey, it's. Oh, I see we have the CEO of Kagi coming up next month, Vlad Prelabek. That'll be great. Can't wait. Lots of good stuff ahead for Intelligent Machines. I hope you will join us. We'll be back next week at the same time, 2pm Pacific, 5pm Eastern, 2100 UTC on Wednesdays. Thanks everybody for joining us. We'll see you next time on Intelligent Machines. Bye. Bye. I'm not a human being.

[Paris Martineau]

Not into this animal scene. I'm an intelligent machine.