Intelligent Machines 866 transcript

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

Leo Laporte [00:00:00]:
It's time for Intelligent Machines. Jeff and Paris are here and our guest this hour is Craig Maud. He's a photographer, a walker, he likes to take walks, lives in Japan and loves AI, AI and the humanities. Next, intelligent Machines, podcasts you love from people you trust. This is Twit. This is Intelligent Machines with Paris Martineau and Jeff Jarvis. Episode 866 Recording reported Wednesday, April 15, 2026 I'm bonkers for Yonkers. It's time for Intelligent Machines, the show where we cover the latest in AI and robotics and all the doodads, gym cracks and gee ga goo gaws around you.

Leo Laporte [00:00:50]:
Speaking of googas, here's Paris Martineau, who did not find the appropriate googa to connect her headphones.

Paris Martineau [00:00:57]:
She was looking, spent a good 15 minutes at the beginning of this thing looking for a goo God Jim job that both Jeff, Leo, Bonito, and seemingly everybody has. I do have one as well, the jack to connect a big headphone jack to a little. But I can't find it. Yeah, I can find this dual ended tiny headphone jack that I've just realized when I stretch it like this, smells foul and chemically in a way that I'm sure resulted in losing.

Leo Laporte [00:01:27]:
Do not sniff the headphone jack. I think that they taught us that in DJ school.

Benito Gonzalez [00:01:31]:
I do believe the technical term is a big to small. A small to big.

Leo Laporte [00:01:35]:
Small to big.

Paris Martineau [00:01:35]:
That's the technical term. Yeah, I was looking for. I was looking for a big to

Leo Laporte [00:01:39]:
small, and I've only been small big Audi and a. A little innie or something like that. Anyway. Hello, Paris, investigative journalist at Consumer Reports who was on deadline, right?

Paris Martineau [00:01:50]:
I was.

Leo Laporte [00:01:51]:
Are you still on deadline?

Paris Martineau [00:01:54]:
In a way, but in a different way.

Leo Laporte [00:01:55]:
I guess the virtue of deadline is it doesn't last.

Paris Martineau [00:01:59]:
Yes, I hit that deadline that I was talking about last week. And while I was on that deadline, I found out that I was up for a Deadline Club award, which is kind of fun.

Leo Laporte [00:02:07]:
What? Congratulations.

Paris Martineau [00:02:10]:
I mean, I'll find out whether I actually win next month, but I want to finalize.

Leo Laporte [00:02:14]:
Before we go any farther, let me introduce your professor of Journalistic Innovation emeritus at the Craig Newmark Graduate School of Journalism at the City University of New York, Mr. Jeff Jarvis. Professor. He's got the Guga. So tell us about this Deadline Award. Is this a big deal, Jeff? Yeah.

Jeff Jarvis [00:02:34]:
Yes, absolutely. It's your. It's the Oscars of New York journalism.

Paris Martineau [00:02:38]:
Yeah. Fantastic.

Leo Laporte [00:02:40]:
For the lead.

Paris Martineau [00:02:42]:
Yeah. And an award magazine, investigative reporting. It's part of a package called Consumer Reports Protecting Food Safety In America features my work and two of my colleagues.

Leo Laporte [00:02:52]:
Hey, that, that you're a shoe in. That should be a no brainer.

Jeff Jarvis [00:02:57]:
Don't jinx her.

Paris Martineau [00:02:58]:
You know, we'll find out come May 14th.

Leo Laporte [00:03:01]:
Is there a awards ceremony?

Jeff Jarvis [00:03:03]:
There is a dinner, I don't doubt.

Paris Martineau [00:03:05]:
Of course.

Leo Laporte [00:03:06]:
How fun. Do you have a tuxedo for the occasion?

Paris Martineau [00:03:10]:
I should get a tuxedo.

Leo Laporte [00:03:12]:
You should. You could wear that.

Paris Martineau [00:03:14]:
Maybe it'll be hot. Maybe I should wear that.

Leo Laporte [00:03:16]:
Wait a minute. Somebody told me. Actually Kelly Ripa and Mark Consuelo on Monday told me it was going to be 88 degrees in New York City.

Paris Martineau [00:03:24]:
Today it is 87. Right now my AC unit is blasting and frankly I need to get my second AC unit and my window reinstalled.

Leo Laporte [00:03:33]:
It's not even possible. It's. It's April.

Jeff Jarvis [00:03:35]:
My wife just scolded me for opening the bathroom window.

Leo Laporte [00:03:38]:
Well, you know, she came in through the bathroom window. Yeah, that's. That really landed you. Don't neither of you know the Beatles? Apparently. Okay, never mind. It's a Beatles song. I'm sorry I brought it up. We're gonna kick things off today on Intelligent Machines by talking to somebody I've been an admirer of for a long time, especially because of his attitude towards life.

Leo Laporte [00:04:02]:
Craig Maude is our guest, a photographer, a writer, a Walker. He lives in Japan, so it's early morning for him. Craig, welcome to Intelligent Machines. It's good to see you.

Craig Mod [00:04:14]:
Hey, thanks for having me.

Leo Laporte [00:04:15]:
Yeah. The reason I was interested, you know, our show's about AI and a lot of what we talk about and the people we talk to are geeks. But you're a humanist who loves AI and loves technology and uses technology. And so I think that makes it kind of an interesting conversation. I guess though, if you're a photographer, that counts a little bit because that's absolutely. These days anyway, a technological hobby or business?

Craig Mod [00:04:46]:
Well, yeah, I mean, I think, you know, photography has always been technological, right? Photography's always been the kind of the, the, the domain of mega dorks. You know, I mean, that was sort of. I mean, it's like it's always been so dorky. And you know, there was an element of gatekeeping too with that dorkery, right. Where, you know, until the late 90s, early 2000s, when Nikon started making, Sony started making potentially usable in professional settings, digital cameras, you know, that the idea of like being in the dark room and knowing the magic of chemicals, that was really a sort of way of protecting a domain, you know, and I think photographers were really freaked out when digital photography happened because it was sort of a loss of that control or that loss of that kind of like mage in the. In the high castle sort of situation that for so long was, you know, typifying of photography.

Leo Laporte [00:05:38]:
Yeah. Suddenly everybody was there.

Craig Mod [00:05:40]:
Yeah, yeah.

Leo Laporte [00:05:41]:
Like when AOL added the Internet. It's welcome. Welcome to the unwashed masses.

Craig Mod [00:05:46]:
Yeah, exactly. What was it? The Eternal September.

Leo Laporte [00:05:49]:
Yeah, that's what they called it.

Craig Mod [00:05:51]:
Yeah.

Leo Laporte [00:05:53]:
Craig, as you might know, is a writer. He's written for everybody, you know, Wired in Atlantic, and he's got two books and he's actually kind of got an interesting lifestyle because you support yourself with this. The thing you call, what is it? Special members, Special projects. Special projects, which is a membership. Like, it's like a Patreon, but it's your own thing.

Craig Mod [00:06:17]:
Yeah, exactly. Well, it's. It's run. The membership component is handled by Memberful.

Leo Laporte [00:06:22]:
Okay. We use Memberful also. I like Memberful. That's a Patreon company.

Craig Mod [00:06:26]:
Exactly. Owned by Patreon. And Patreon's actually been a pretty hands off steward of the, you know, of the whole thing, in a sense. Memberful was acquired, I think maybe six years ago.

Leo Laporte [00:06:37]:
And it's. You wouldn't know. It's a Patreon company, which is.

Craig Mod [00:06:40]:
You would not know.

Leo Laporte [00:06:41]:
Yeah.

Craig Mod [00:06:41]:
So I use it simply as kind of a. An interface in the back end to handle subscribing, unsubscribing, dealing with billing, which goes through Stripe anyway, and then everything else I've kind of scaffolded on top of my. As my own. So there's. I'm not actually using any of Memberful's, you know, blogging components or podcast components or anything like that.

Leo Laporte [00:07:03]:
We use them to hook up people to Discord, which is very handy. So we have a Discord channel. People are watching right now on our Discord channel. And we use them because it makes it easy for us to have a dedicated podcast feeds for members. But you're right, I mean, it's really mostly about getting kind of the bookkeeping out of the way.

Craig Mod [00:07:23]:
Yeah, yeah, the bookkeeping. And. And yeah, just having some third party with a little bit of responsibility there. But, you know, the Discord thing is funny. I had for years members clamoring for a Discord and then members making sort of unofficial discords, and I finally made one about two and a half years ago and I really didn't like it. It just felt to me like way too much effort to. To maintain.

Leo Laporte [00:07:47]:
Yeah, it's A lot.

Craig Mod [00:07:48]:
It's a lot. And then the software itself was a lot and boosting and all this. So last year the first thing I did with cloud, cloud code was I made my own Twitter and that has become our members, sort of like Discord Chat. And it's wonderful. It's so good. It's such a lovely environment. It's. It's actually shocking how, how great it's been and I check it every day.

Craig Mod [00:08:10]:
We have a whole community of people who are in there every day and the fact that it's not real time I think is really important. So you don't feel that like chat room element. Of course you can't do things like this where you can livestream and have a conversation around.

Leo Laporte [00:08:23]:
We used for years IRC and I still am a big fan of irc, but because there was no easy way to make it a members only thing, we just use Discord by default because that's what memberful supports. I would far prefer to use IRC without all of the bells and whistles. It's much simpler. So yeah, this is, so this is why I wanted to have you on Intelligent Machines, because here you are. I mean, you did major in digital media, right? I mean, you're not an alien to technology. You're not.

Craig Mod [00:08:51]:
No, no, no, no. I mean, I grew up. I was, you know, I started programming when I was 9 or 10 and I was running a BBS, you know, oh, neat and all that stuff. So. No, and my degree is fine arts and computer science. So it's.

Leo Laporte [00:09:04]:
But that's what's great. It's this, it's. It seems rare that you get this kind of what Steve Jobs called the intersection of technology and humanities, but I think that's maybe the most interesting intersection of all. We don't, we don't want to become machines and machines should only be in service of us as humans. Right. So I think that I really appreciate how you've stayed a humanist and yet are embedded, you know, as you say in your piece, you're on your new website. Software bonkers. When did you start using cloud code?

Craig Mod [00:09:45]:
A year ago.

Leo Laporte [00:09:46]:
Right.

Craig Mod [00:09:46]:
When it came out. Well, I remember, you know, you'd kind of use ChatGPT to get snippets and stuff.

Leo Laporte [00:09:51]:
Yeah. And you paste it in.

Craig Mod [00:09:52]:
Yeah, whatever. Yeah, yeah. And then Claude code came out and it was like, wow, this is, this is pretty. This is some sort of next level instantiation of it. But you know, it's like today using cloud code and I use it with, you know, all These skills and plugins and stuff now and, you know, a year ago looks, you know, positively archaic.

Leo Laporte [00:10:15]:
November. I trace it to November 24, 2025, the release of Opus 4.6. Yeah, something happened.

Craig Mod [00:10:23]:
Yeah, yeah.

Leo Laporte [00:10:24]:
It could do stuff.

Craig Mod [00:10:26]:
It could do stuff. And then also the tooling around it, I think, got better. You know, I was hanging out with Kevin Rose a month ago or so, and we were just going through his workflow, and he's just completely, like 1 billion percent AI, sort of like maximalist. He's got like 10 agents going at every moment doing 100 million different things. But I was looking at kind of what he was using. I was like, oh, okay, this is. This is interesting. Actually, until a month ago, I was still of this mindset that I needed to look at all of the code that was being changed and I needed to understand every line that was being implemented by Claude.

Craig Mod [00:11:03]:
And something. For me, in the last month or six weeks, I've shifted where now I'm just like, I don't need to see anything. I don't care.

Leo Laporte [00:11:13]:
You don't care or you trust it? Actually, that's a really important distinction because it may not be the best code. You don't care?

Craig Mod [00:11:21]:
Yeah, well, I mean, I don't care in the sense of, like, the speed at which now we can produce things. And. I know. And it just keeps getting better. And so I kind of see a lot of what I'm working on today is almost like throwaway code.

Leo Laporte [00:11:32]:
It's disposable.

Craig Mod [00:11:33]:
Yeah, it's disposable.

Leo Laporte [00:11:34]:
That's what Harper Reed said. He said, don't. It doesn't matter. You throw it away.

Craig Mod [00:11:37]:
Yeah, exactly. Like, you know, make the mvp, see if it. See how things feel. And then if it needs to be really beefed up and given, you know, given the full workover, then great, okay, go back with it. But in six months, the tools are gonna be so much better. Anyway, so it's this weird. And I wrote another piece that published last week, just talking about this sort of, like, the loss of purpose that is sort of off in the horizon. That's kind of.

Craig Mod [00:12:00]:
You know, you see all these eulogies for coding every day. Even today, there was some big post about VIM and the death of coding because of, you know, vim.

Leo Laporte [00:12:09]:
I'm an emacs guy. You can take your vim, I don't care. But, no, I understand that there is that sense that we had something. We were wizards, and we had this. The same thing that you just talked about with photography. We had the keys to the kingdom and now, God damn it, everybody can do it and we're not special.

Craig Mod [00:12:29]:
Yeah, yeah, yeah, it is. It's very strange. But has there ever been a technological shift that's happened so quickly? I think is, to me that's the big shock is to go from a year ago, people going, oh yeah, you can kind of code and you can copy and paste to now, you know, folks just eulogizing programming. Every day another person saying it's, you know, it's done. 40 years I spent.

Leo Laporte [00:12:53]:
Do you agree with that skill set? Is it the end of coding?

Craig Mod [00:12:56]:
I, I do think. Well, look, I think coding, if we're talking about, if we're talking about mass produced code, I think it's been dead for a long time. Like, I think that the idea of craft encoding for most commercial software, like we're in Zoom right now. This, to me, I only ever open zoom to do stuff like this and every time I do I'm just in shock. Like, this is, this is absolutely the worst piece of software that I've opened

Jeff Jarvis [00:13:18]:
on my computer

Craig Mod [00:13:21]:
in weeks. And sorry, I apologize.

Leo Laporte [00:13:26]:
I'm making you use this.

Craig Mod [00:13:27]:
We should make our own. Like, this is a great.

Leo Laporte [00:13:29]:
You could, couldn't you?

Craig Mod [00:13:31]:
We could easily do better than this. I mean, come on, someone out there, you could do better than this. But like, look, okay, if this is like state of the art human made software, this is terrible. This is like, this is not winning to begin with for me. I use stuff like things and you know, Neovim or what, like these other pieces of software like Mime Stream is a, is an email program I use. And these all feel like genuinely crafted, like craftsmanship made pieces of software and that. I think even with LLMs and cloud code assisted whatever programming, that's not going to go away. No, you're going to have people that can use this stuff as a sort of like superpower and be able to do more in a crafty way.

Craig Mod [00:14:13]:
And you're going to have more people dumping junk out into the world too. I mean the analog with photography is really astounding, actually.

Leo Laporte [00:14:20]:
Yes. Because people, there will still be people who will use a Leica with a rangefinder and be able to take one picture every four minutes. And there will be people who will write, people will be taking a thousand pictures a second with their burst mode on their iPhone. Yeah, it's the death of maybe industrial software, but it will never be the death of the artisan. Because you know why I think humans create and we want to create and we enjoy creating and coding is another form of creation that's, you know, like woodwork or any other kind of craft.

Craig Mod [00:14:58]:
Absolutely.

Leo Laporte [00:14:59]:
So you. You said in this piece your first Claude code project was to rebuild Twitter, as.

Craig Mod [00:15:07]:
I thought it should be so obvious.

Leo Laporte [00:15:09]:
It's obvious what's wrong with Twitter today. But even the pre. Elon Twitter you weren't happy with.

Craig Mod [00:15:16]:
No, it always made me uncomfortable. Twitter. I mean, even the name, the first name, it was twttr.

Leo Laporte [00:15:22]:
Right.

Craig Mod [00:15:22]:
Even that gave me anxiety.

Leo Laporte [00:15:24]:
This, by the way, it's 20th anniversary of the first tweet is. Was Monday. Yeah, it's 20 years ago. Just send over Twitter began. So what was wrong with Twitter and how did you fix it?

Craig Mod [00:15:37]:
There was a lot that was. That was wrong. First of all, I think we're. There's a certain generation of online people that think everything should be archived. And I think that's just like a really. That's a wrong narrative. And one of the, you know, I hang out with Kevin Kelly a lot. We go on walks and kind of do, you know, these sort of like roundtable discussions.

Craig Mod [00:15:55]:
And one of my heresies that I hold and I kind of a little bit facetiously is that I think the entire Internet every. Every two weeks should be entirely deleted.

Leo Laporte [00:16:05]:
Wow. Don't tell Brewster Kael that. Wow.

Craig Mod [00:16:10]:
Well, you could. My theory is like, if you, if you did that, then people would be, you know, archiving what needed to be archived. You could print out what need to be printed. Print it out, all these things, but every two weeks, so you get this kind of completely clean slate.

Leo Laporte [00:16:22]:
Clean slate?

Craig Mod [00:16:23]:
Yeah, clean slate. That's the idea.

Leo Laporte [00:16:25]:
And yeah, that's a good point, because nobody really wants to see what they tweeted 12 years ago. That's embarrassing.

Craig Mod [00:16:32]:
It's only bad things. And it's like, there's no.

Leo Laporte [00:16:36]:
If it were that good, you would have saved it, right?

Craig Mod [00:16:39]:
Yeah, exactly. No one's ever been like, oh, my God, I'm so happy I have that tweet from 12 years ago. Someone's quoting for me, that can only.

Leo Laporte [00:16:45]:
So you think it should be Swedish death cleaning with your social.

Craig Mod [00:16:49]:
Yeah, basically, yeah. Maria. Maria Kondo. The only thing does that spark, joy. Spark, joy. So anyway, Twitter, I think, should it, you know, I think should have been doing that. I love things that disappear. Ephemerality online.

Craig Mod [00:17:02]:
And so that was, you know, it's like all the posts should be disappearing. A limit to how much you can post every day. So you're actually just thinking about what you're putting up There, what you're doing, how you're engaging with it.

Leo Laporte [00:17:11]:
There are people, including the President of the United States, who post hundreds of times a day. It's like, don't you have something to do?

Craig Mod [00:17:18]:
Well, it's the same reason why, like, should we be able to do sports betting 24 hours a day in our pocket?

Leo Laporte [00:17:23]:
Probably not.

Craig Mod [00:17:25]:
I mean, there's just. There's just certain. There's just certain things we're not. Our monkey brains are not wired for. So to be able to pull a lever to get dopamine at any moment of the day.

Leo Laporte [00:17:34]:
Oh, that's amazing.

Craig Mod [00:17:36]:
And most sort of maliciously, every moment of the night, you know, in the middle of night, three in the morning, I think that's bad. And then. Yeah, just. It's, again, non algorithmic, because the nature of my community is it's not 400 million people tweeting. You can just keep a reverse cron timeline. So you're not doing algorithmic stuff. And. I don't know, just that, to me, it was very simple what.

Craig Mod [00:17:59]:
How it could work.

Leo Laporte [00:18:00]:
We should. By the way, I should mention this is real. This is your special projects group.

Craig Mod [00:18:06]:
And it's called. Well, and the version of Twitter I made is called the Good Place. And when you add it to your home screen on the iPhone, the icon is Ted Danson's head.

Leo Laporte [00:18:17]:
You have to say things like, oh, fudge, and shut the front door.

Craig Mod [00:18:22]:
I should add a function that rewrites everybody's swears.

Leo Laporte [00:18:27]:
This is interesting. All images appear in one day. Bit black and white.

Craig Mod [00:18:31]:
Yeah.

Leo Laporte [00:18:32]:
And then you click on it and you can see it.

Craig Mod [00:18:35]:
Yeah, exactly.

Leo Laporte [00:18:35]:
Is that so they don't draw? They suck. Don't suck you in.

Craig Mod [00:18:38]:
Yeah, just. I don't know, it's just like kind of an aesthetic, a little bit of playfulness.

Leo Laporte [00:18:44]:
Yeah. And I really like this, which is no algorithm. It's just.

Craig Mod [00:18:48]:
Yeah.

Leo Laporte [00:18:48]:
In order.

Craig Mod [00:18:49]:
Yeah. I. I almost feel guilty being a Twitter user around 2010, 2011, 2012, because. Because you'll remember, everybody was on it. Everyone who's a writer, journalist, and you could just connect with people. And it didn't have the algorithm making stuff sort of like inviting or rewarding toxicity or this kind of the virality garbage that you see the slop you see everywhere now. And it was. You know, you could really build up an audience and sort of a career.

Craig Mod [00:19:20]:
You could connect with everyone. You know, my. My audience was hugely built up on Twitter. You'd write an essay, you'd put it up there, you'd have 500 more people subscribe to your newsletter the next day. It was kind of wild. It was sort of a heyday of that stuff.

Leo Laporte [00:19:34]:
Did you write this before opus 4:6. You wrote this last year.

Craig Mod [00:19:37]:
I wrote this last year, yeah, yeah, yeah, it was pretty. But this is the thing is with open source software kind of being the main, the corpus for all this, all the models, training for programming, none of these pieces are unique. And so it's. You can kind of, you know, most software is not in reinventing the wheel. It's about putting, you know, different pieces together.

Leo Laporte [00:20:00]:
Well, that's why Claude Code's good at it, right? Because it's, it's seen it.

Craig Mod [00:20:04]:
Exactly.

Leo Laporte [00:20:04]:
It's seen the code. When I wrote a newsreader and it immediately knew what a newsreader should be and what it should have because it's seen 100 of them.

Craig Mod [00:20:13]:
Yeah, exactly.

Leo Laporte [00:20:15]:
In fact, I had to tell it, no, I don't want to star articles. I don't want. I mean, I had to tell it not to do some things that most feed readers do because the reason I wanted a custom feed reader is to strip it down to just the things I really needed for speed.

Craig Mod [00:20:32]:
Yeah.

Leo Laporte [00:20:33]:
How long did it take you to write the Good Place?

Craig Mod [00:20:38]:
Probably about a week maybe. Kind of on and off.

Leo Laporte [00:20:42]:
I'm kind of amazed. And it's still. And it runs well, it's reliable and.

Craig Mod [00:20:48]:
Oh, it's incredible. Yeah, it's very simple. It's.

Leo Laporte [00:20:51]:
So you wrote exactly what you wanted. This is the thing that I think is so empowering, is software can be personal.

Craig Mod [00:20:59]:
Yeah. Well, Robin Sloan has a great essay called Software is a Home Cooked Meal.

Leo Laporte [00:21:04]:
Yeah.

Craig Mod [00:21:05]:
And so this idea that you can kind of make the perfect meal you want in the form of software, this is where you bump into the issues with the App Store on Apple is like it creates so much friction between wanting to do this stuff and putting it out there and getting it out there and having people use it. In the sense you need a developer account, you need to submit it, it needs to be approved, all this stuff. Whereas when you're building things like I built with the Good Place, it's just on a digitalocean server and, and you know, it's locked down and everything. And I use Flask, it's web based, but you know, you use these Python frameworks and you know, it's databases now are so good and so lightweight. SQLite is amazing. So you don't need to spin up separate servers for that. Anyway, there's a. There's a lot that you can do very quickly when you, when you work off the web.

Craig Mod [00:21:53]:
And I do wonder if this is going to be a hindrance for sort of the iOS sort of growth going forward is not being able to deploy as quickly as you want to being a thing that keeps people. You know.

Leo Laporte [00:22:06]:
I was really hopeful for a Progressive web apps PWAs and I was very disappointed that Apple decided not to really fully support it. Even Firefox kind of turned away from it because I do think that there is an opportunity if it's a web web technology so well understood and so easy to implement. You don't really need standalone apps for most things if you have the right services in the background, you know, have the right tools that can remember where you were and things like that. That's true. So this is all. If you become. I'm going to join because I want to see this new. This new better Twitter for sure.

Leo Laporte [00:22:42]:
I bet it's a great community too. I think it's always great to have a community that people have to pay, even if it's just a little bit.

Craig Mod [00:22:50]:
Yes, well I agree. I mean that is sort of one of the things that makes it. Makes it good. Is that it? You know, it's the. The unwashed masses aren't allowed in. It is an Ellis Island. It is, but it's also not.

Leo Laporte [00:23:03]:
You're such an elitist.

Craig Mod [00:23:05]:
It's also not Epstein Island. So it's.

Leo Laporte [00:23:08]:
There are all kinds of islands, aren't there? 10 bucks a year, a hundred dollars a yearly or become a lifetime member and surprise the heck outta Craig and scare me.

Craig Mod [00:23:19]:
He says a little bit surprisingly, I've had so many lifetime members that. So many that I feel like I should, I should probably raise that price.

Leo Laporte [00:23:31]:
This is your sole support?

Craig Mod [00:23:34]:
Yeah. This in book sales.

Leo Laporte [00:23:35]:
That's fantastic.

Craig Mod [00:23:36]:
Yeah. Well I was really lucky in that I found product market fit between the membership program and book selling books. Because it turns out. So I make, you know, I do these fine art editions of books. I had a book come out with Random House last year as well. Things become other things and then I have these fine art editions that you can see there on the left on that are basically a hundred dollars a book, limited edition, signed, blah blah blah, printed and bound in Japan. And you know, I worked it. I've worked in the publishing industry and found physical books for decades and the idea of selling a lot of hundred dollar books is just not something anyone thinks about or is able to do.

Craig Mod [00:24:18]:
And so when I first started producing these, which was about six years ago now, I Produced a thousand copies of this hundred dollar book. And I thought, okay, it'll take me like two years to sell all these. And we sold out in two days.

Leo Laporte [00:24:28]:
Wow.

Craig Mod [00:24:29]:
And I did this thing where if you paid a hundred dollars to join the membership program, you got 40 bucks off the board book. So people were landing on this buy page, they saw $100 book and they thought, okay, well, I can pay $100 to get the book, or I can pay $100 to join the membership program, plus $60 to get the book. So I'll pay $160 for the book. And the conversion rate of people who weren't members already was something like 30%, which is just an insane number to get people to convert over to joining a membership program. So I've just kept that going. We've reprinted that first book six times. This. I've had two other books come out.

Craig Mod [00:25:03]:
This Random House book book has sold. Really? We've gone. The Random House book has gone back to print twice. So I've been really, really lucky. Yeah, the two of those together work really well.

Leo Laporte [00:25:13]:
What are you working on now with Claude?

Craig Mod [00:25:17]:
Well, it's tax season.

Leo Laporte [00:25:20]:
So you're writing your own tax software now? That's nuts.

Craig Mod [00:25:25]:
Well, it turns out it's not as hard as it looks.

Leo Laporte [00:25:29]:
It's. Well, I guess you can look up all the. All this. All the information is public.

Craig Mod [00:25:34]:
Yeah.

Leo Laporte [00:25:34]:
So all the tax tables and everything.

Jeff Jarvis [00:25:36]:
Well.

Craig Mod [00:25:37]:
And I'm in this.

Leo Laporte [00:25:37]:
Are you worried about hallucination?

Craig Mod [00:25:40]:
Well, no, because it's verifiable and I still employ accountants, so.

Leo Laporte [00:25:43]:
Okay. Okay.

Craig Mod [00:25:46]:
So, you know, I'm in this. A really annoying position of being an American who doesn't live in America, meaning that you still have to file taxes back in America and in. In the place where you're domiciled. So I'm in Japan, so I'm filing in Japan, and then I have to file in America the next month, and it's really exhausting. And I have bank accounts in both countries and reconciling stuff, and there's no. Literally no accounting software out there where that will handle the Japan side of things and the US Side of things that will help me track transfers between countries because, you know, you transfer a thousand bucks over and it takes a day and the exchange rates fluctuate, and so the software can't match it precisely. And it's in yen over here and it's dollars over here. Well, guess what? Claude is really smart at this stuff.

Craig Mod [00:26:37]:
And you can. You can tell it to have fuzzy matching around these things. And you can say here, these are how I do my transfers. And you can, you can build a piece of accounting software with this reconciliation mind in the background that is just able to do a 10 times better job than you could ever do 10 times faster. And then of course you go through and you can, you know, you can check everything and you can, you know, update stuff that needs to be fixed or whatever. But the general kind of reconciliation of cross border accounting has gone for me from, from being almost impossible. And this concoction of I was using Google spreadsheets and, and then Google Script in Google sheets to convert stuff between dollars and yen and then export that and then get one account in this spreadsheet and another account in this other spreadsheet. And I basically built this for me.

Craig Mod [00:27:28]:
What is the perfect accounting software? And at the end of the day, I push one button, it makes a zip file. It's got all the 1099s and all the other PDFs and forms, medical records, medical receipts, other receipts. Everything is itemized perfectly. It knows how to the scheduling for American expenses versus Japanese expenses. It's looking at depreciations over the years. I can feed in, you know, last year's tax, tax report or tax filings and say, this is how we classify things. And it goes, okay, great, now I see that. Let me, let me get you set up for this year.

Craig Mod [00:28:03]:
So you're doing the same things. It's, it's wild. And actually, since I published the software bonkers thing where I talk about this, a representative from Quicken reached out to me and wanted to do a meeting.

Leo Laporte [00:28:15]:
So we want to buy Taxpot 2000. I mean, so you say this took you five days to write?

Craig Mod [00:28:23]:
Yeah.

Leo Laporte [00:28:24]:
I mean, do you think anybody could do this, Craig? It seems to me I would. I'm not, I look, I'm pretty proficient. I don't know if I would try to tackle this. You just were, you just, Was it iterative where you go, okay, now let's, we got to figure out how to do this conversion. Was it like that? Was it a back and forth process?

Craig Mod [00:28:42]:
A lot of back and forth. It wasn't definitely. It wasn't a one shot and I had, you know, perfect software. But it was, you know, I mean, part of the thing about working with, with models like Claude is it does require and it forces you to really think through what it is you're looking for. So you write specs essentially.

Leo Laporte [00:29:00]:
Do you start in plan? You start in plan mode, then start

Craig Mod [00:29:03]:
in plan mode in brainstorm mode. Actually, I use.

Leo Laporte [00:29:06]:
Do you use. Yeah. What tools do you use in Claude? Because there's so many different things you can do. Superpowers and GSD and.

Craig Mod [00:29:13]:
Yeah, there's. There's a. I. I know Gary. I think Gary Tan has a tool use.

Leo Laporte [00:29:17]:
Gary Tan's thing. Yeah, well, he's the YC president. He put out a YC pres.

Craig Mod [00:29:21]:
Yeah, he's got his set of tools, but I've been using Every.

Leo Laporte [00:29:23]:
The.

Craig Mod [00:29:23]:
There's a company called Every that put out a set of skills that.

Leo Laporte [00:29:26]:
They're just skills. They're just prompts, mostly.

Craig Mod [00:29:29]:
Right.

Leo Laporte [00:29:29]:
Bash commands.

Craig Mod [00:29:30]:
You do brainstorm, and then you do. And then you do planning, and they do. Working. And I found it's worked really, really well. So, you know, you go into brainstorming and you start with saying, hey, here's a folder. Literally just running it in my tax folder and saying, hey, here's a folder with all of my tax stuff. Take a peek at this. You know, we're looking to help me understand every.

Craig Mod [00:29:58]:
E V E R Y E V

Leo Laporte [00:30:00]:
E R Y. Yeah.

Craig Mod [00:30:02]:
Every, I think is the name of the problem is.

Leo Laporte [00:30:03]:
It's such a generic word. I'm having a hard time.

Jeff Jarvis [00:30:07]:
There you go.

Leo Laporte [00:30:07]:
Is it everything Quadco? No, it's not. It's not.

Craig Mod [00:30:12]:
It's a. There's a set of.

Leo Laporte [00:30:15]:
You're seeing my screen, by the way, if you want. If you're tired of seeing my screen, there's a tab in the upper left of your zoom that's heavily hidden, and you can click that, and then you won't see my screen as I. Oh, wow.

Craig Mod [00:30:26]:
There we go. Okay.

Leo Laporte [00:30:27]:
Yeah. Sorry, I didn't realize we had not told you that. That's. That's distracting. It's much less distracting because I'm always doing so. Yeah. Because I'm very interested in. There it is.

Leo Laporte [00:30:39]:
Every Ink Compound Engineering plugin is that.

Craig Mod [00:30:44]:
That's the one. Compound Engineering. Exactly.

Leo Laporte [00:30:46]:
All right, I'm gonna try it. You know what I found, Craig, is I have been erasing my CLAUDE code setup every once in a while to start from scratch because it's moving so fast that a lot of the things I do with skills now are built into Claude and.

Craig Mod [00:31:03]:
Yep.

Leo Laporte [00:31:04]:
It's like Moving Target.

Craig Mod [00:31:06]:
No.

Leo Laporte [00:31:07]:
Do you use the Claude Max subscription? What is your CLAUDE subscription?

Craig Mod [00:31:11]:
Yeah, Yeah. I switched to the 200amonth.

Leo Laporte [00:31:13]:
I mean, I was.

Craig Mod [00:31:14]:
I was. I.

Jeff Jarvis [00:31:14]:
It's.

Craig Mod [00:31:15]:
Which. Which when they announced it, I was like, I'm never gonna pay 200amonth for anything. Yeah. Who's gonna do that. And now I'm like, think at least it's not 500, because I'd probably spend that as well. I mean, I'm getting so much value.

Leo Laporte [00:31:26]:
Yes, you just saved that much in some tax program subscriptions.

Jeff Jarvis [00:31:32]:
Oh.

Leo Laporte [00:31:32]:
I mean, pays for itself easily.

Craig Mod [00:31:34]:
Yeah, Yeah. I mean, now if I could only reprogram all of Adobe software, that would be. That would be even better.

Leo Laporte [00:31:39]:
Well, you know, how many hours in a day are there? So I wanted to ask you. You live in Japan. You've lived there for two decades now.

Craig Mod [00:31:45]:
Yeah, yeah.

Leo Laporte [00:31:47]:
What are the attitudes towards AI like in Japan? You know, I've heard that in other countries here in the United States, about 50% of the public thinks it's the devil incarnate.

Craig Mod [00:31:58]:
Okay.

Leo Laporte [00:31:59]:
And the other 50% cannot or cannot stop using it and keep telling the other half. What are you. What are you crazy? This stuff's great. And there are a lot of people say, oh, no, uses too much energy. It's gonna kill us someday. I mean, there are. People are very afraid of it. What's the attitude in Japan?

Craig Mod [00:32:17]:
Japan is, I don't think is quite as hyperbolic as the US in general. It's a little more. It's a little more relaxed. I mean, this is also a country that still, you know, faxes stuff. You can order bentos from faxing. There's a guy, you know, there's a. There's a jazz cafe I like to visit. And the only way to.

Craig Mod [00:32:35]:
To find out if it's going to be open is if you fax them.

Leo Laporte [00:32:37]:
I love that. So they invented the facts because they, you know, they had to. Yeah, that's hysterical. They still fax.

Craig Mod [00:32:45]:
They're legally obligated to keep faxing because they made it. But the, the, the AI thing is funny because there's been electoral candidates who've been running on the platform of I'm going to use AI to make all my decisions. So, like I. Mayor, things like, interesting have been popping up and people seem to be like, weirdly kind of into it. So. Yeah, it's not quite, you know, in the. In the US Even today that, that eulogy for Vim or whatever I was reading earlier, the crux of it is this is terrible. It creates sloth, it's bad for the environment, it's using up too much energy, it's using up too much water.

Craig Mod [00:33:20]:
All this stuff that I think are on one hand, very fair arguments. But these are also arguments that everyone made towards crypto and the thing that ended up killing, like destroying NFTs or stopping NFTs wasn't the fact that it was ecologically damaging. It was because they were essentially useless. And so, I mean, that's, that's a good reason. That's, that's how you. That's how stuff gets, doesn't get used anymore, you know, And I think that, you know, AI is, is proving to be so profoundly useful.

Leo Laporte [00:33:46]:
Yeah.

Craig Mod [00:33:46]:
That we're just going to have to figure out the energy stuff. It's not, no one's going to stop using this. You can't, you can't walk back on this. So I find those hardliners a little bit overly disconnected from reality and almost, you know, that's someone you just can't have a conversation with at this point. I mean, we should, we should have, you know, you read books like the Making of the Atomic Bomb. You look at the history of atomic energy and stuff like that, and like, we should just have better energy programs. We should just be. Have cleaner energy, more renewables.

Craig Mod [00:34:13]:
I mean, it's like the energy question. Any, any way we can accelerate getting better energy out in the world and moving us to. Away from natural gas and coal and all this stuff is something we should be celebrating. And I think this is, this is part of that. Acceleration is the demand that's going to come for, for AI because it is so useful.

Leo Laporte [00:34:34]:
I agree 100%. And maybe he can even help us solve some of these problems.

Jeff Jarvis [00:34:38]:
I don't know.

Craig Mod [00:34:39]:
Possibly.

Leo Laporte [00:34:40]:
Do you ever feel like it's sentient? I mean, do you. What's your relationship with Claude?

Craig Mod [00:34:47]:
I. I mean, how clear are you

Leo Laporte [00:34:50]:
that it's just a program, is what I'm asking.

Craig Mod [00:34:52]:
I'm very clear that it's just a program. I'm actually reading Michael Pollan's new book on consciousness.

Jeff Jarvis [00:34:57]:
Yes.

Craig Mod [00:34:57]:
You've read that?

Leo Laporte [00:34:58]:
I haven't read any yet. Can't wait. Yeah, it is.

Craig Mod [00:35:01]:
It's great. I'm 64% through, so says my Kindle. And, you know, it's one of these books where the more you read, the less you feel like, you know, and the more confusing it becomes. And one thing that you do pick up from it is the complexity of consciousness is so beyond language and it's so beyond feeding, you know, more language, language corpus into these models. I think the world model stuff that some of the early AI guys now have broken, splintered off from OpenAI and are raising money around, I think are heading in the direction that may lead to something closer to consciousness. But I think the things we have right now are not conscious, but they are very, very good at performing a lot of the tasks we ask of it. And sometimes those tasks are being a therapist or being, you know, a shoulder to cry on. And it just turns out that you can get most of the way there with.

Craig Mod [00:35:59]:
With what we have. But I, I have no. I harbor no illusion that this thing is thinking or feeling or, you know, anything like that. I think that's. That's very far away because that would be terrifying. It would be terrifying and. But it would also. It's diminishing to humans what we are.

Craig Mod [00:36:16]:
The resolution that we are taking in every moment is so high and it's so profoundly weird. Consciousness. And I was doing this meditation retreat the other day, and I was just thinking about how just fundamentally bizarre it is that we sit quietly and we have this person that exists internally and it's able to gaze inward. I mean, we're doing. We're doing so many things that are just this mishmash of crazy chemicals that have, you know, come about over millions of years of years, brute forcing better versions of ourselves that I think without some kind of, like, huge computing boosts, quantum computing or something, that consciousness in a machine is still quite a ways off. Anyway, that's my. But my, my layman's take.

Leo Laporte [00:37:01]:
Do you say please and thank you when you talk to Claude?

Craig Mod [00:37:04]:
Yeah, because I don't want to be a jerk in real life. So it's good to practice.

Leo Laporte [00:37:10]:
I can't help it. I can't. I know. I'm just like you. Look, I know it's a computer program. It's not a person. I don't have quite your faith in the uniqueness of our own abilities as humans. I'm not sure that we're not just stochastic parrots, but I still say please and thank you to the machine just in case.

Leo Laporte [00:37:32]:
Oh.

Craig Mod [00:37:32]:
Like, don't get me wrong, I think we are definitely computer ish based. I don't, I don't necessarily believe in this sort of other entity kind of operating, but I just think that the computing is. I mean, if you think about the efficiency and the low wattage that we're doing, what we're doing with, there's. There's some other. There's some other computing insight that's happened there. That, that's pretty amazing. But I will say the one thing I am not polite to is Siri. I'm never polite to Siri.

Leo Laporte [00:38:01]:
No. In fact, my wife quite notoriously swears up a blue streak at Ciri, but Ciri earns It, Yes, Siri deserves it, has earned it.

Craig Mod [00:38:14]:
There's the curb, I think from last season of Curb youb Enthusiasm. There's a whole scene of Larry screaming at Ciri in the car, which I think we've all done a version of that.

Leo Laporte [00:38:24]:
Well, I've just joined tgp. I am Leo Laporte. I look at forward.

Craig Mod [00:38:28]:
Thank you.

Leo Laporte [00:38:29]:
Talking with you. I am. I'm really impressed. I've always been impressed by your posts, by your writing, by your books. Someday I wouldn't mind going on a hike with you and Kevin. That would be pretty amazing. I thank you, Craig, for joining us on Intelligent Machines and giving us a nice human humanistic angle on something that is kind of. Kind of moving fast and changing the world and all that stuff.

Craig Mod [00:38:55]:
Well, thanks for having me. This has been fun.

Leo Laporte [00:38:57]:
Thank you, Craig. Really nice to talk to you. Take care.

Craig Mod [00:39:00]:
Likewise.

Leo Laporte [00:39:00]:
Thank you to Craig Mod for getting up early in Japan so we could talk to him. We appreciate it. We will get to the rest of Intelligent Machines. All the news, including the shoemaker, who's become the latest AI pivot.

Jeff Jarvis [00:39:14]:
This next episode of Rush Hour 4.

Craig Mod [00:39:16]:
Gentrify Adrift. We're in a neighborhood that feels like someone tried to program peace in into a video game and accidentally deleted culture.

Leo Laporte [00:39:25]:
I don't know what's happening. I don't know what's happening.

Paris Martineau [00:39:28]:
Over. I like that we all just sat there in shock silence for a second like someone else had done it.

Leo Laporte [00:39:34]:
That was a AI generated voice that

Jeff Jarvis [00:39:38]:
that guy named Miles to thought it

Leo Laporte [00:39:41]:
was very funny, his voice. But it's not.

Jeff Jarvis [00:39:43]:
Miles tours of New York City in neighborhoods.

Leo Laporte [00:39:45]:
Right.

Jeff Jarvis [00:39:45]:
And he uses that anthropological voice voice. And I get the joke now, but I thought it was his voice. They're very well written.

Leo Laporte [00:39:52]:
They're very supposed to be Richard Attenborough.

Craig Mod [00:39:54]:
Kind of why they call it Long Island City. It's definitely not long and it's definitely not a city.

Leo Laporte [00:40:00]:
But it is an island profile with granite. I have no idea where that sound is coming from.

Benito Gonzalez [00:40:04]:
Yeah, I think it's playing inside the Discord. So you gotta look for the video that's playing in Discord.

Leo Laporte [00:40:09]:
I have no idea what's coming out of that. Every time I switch over Designer. It's Discord. I couldn't hear him because Richard Attenborough was talking. How do I stop it? All right. Closing. That was exhausting.

Jeff Jarvis [00:40:29]:
He's very good though. Milestone.

Leo Laporte [00:40:31]:
I will. I will. You do. How much time do you spend on tick tock every day?

Jeff Jarvis [00:40:36]:
How much time am I in the bathroom?

Leo Laporte [00:40:37]:
Ah, that's what it is. Wasn't there a warning from TikTok or.

Jeff Jarvis [00:40:43]:
Oh, screw them. I don't pay attention to any of those things.

Paris Martineau [00:40:45]:
It's that your legs can fall asleep if you're. I experienced this, but with playing cross. Play, play. But I haven't been. I've been deficient in all my crushes.

Leo Laporte [00:40:53]:
Do I owe you a move or do you owe me a move?

Paris Martineau [00:40:55]:
I think you owe me a move.

Leo Laporte [00:40:56]:
Do I?

Paris Martineau [00:40:57]:
Okay, I might owe you a move. Actually, I owe everybody in my life a move. I've been off the last.

Leo Laporte [00:41:02]:
You were all into it and now you're not so much.

Paris Martineau [00:41:04]:
Well, this last week I've been very busy, so I just haven't had time.

Leo Laporte [00:41:07]:
I don't judge you. I don't judge you.

Paris Martineau [00:41:10]:
I get to file my taxes and figure out the Yonkers questions.

Leo Laporte [00:41:14]:
Okay, now we got to talk about that because you mentioned this last year. Today's tax day. I hope you all did your taxes. If you didn't go rush over there. Keep the podcast going while you're finishing it up. I did my mom's and mine today. Got it all done. But in the New York Tri state

Paris Martineau [00:41:30]:
area, when you're filling out your New York State tax return or having it filled out for you by a robot, you go through a series of questions in which it asks you in every way possible. Did you live in Yonkers during this period? Do you work in Yonkers during this period? Did you think about thinking about working in Yonkers? Did you so much as glance askance at Yonkers during this period?

Jeff Jarvis [00:41:51]:
And you did work in Yonkers.

Paris Martineau [00:41:53]:
And I did, in a sense, because the company I work for is a Yonkers based business. So I finally had to figure out what all that is about.

Leo Laporte [00:42:00]:
You actually had to say yes for the first time in history.

Paris Martineau [00:42:03]:
Yes. And it's. And I had to fill out another form.

Leo Laporte [00:42:06]:
Do you get something out of that?

Jeff Jarvis [00:42:07]:
You get more taxes.

Paris Martineau [00:42:08]:
I get more taxes out of that.

Leo Laporte [00:42:10]:
There's a young break.

Jeff Jarvis [00:42:12]:
No, not a tax break.

Paris Martineau [00:42:14]:
There's a Yonkers tack. There's a Yonkers non resident earning earnings tax and there's.

Jeff Jarvis [00:42:21]:
I pay New York City non resident earnings tax, for example. Well, I did. I don't now.

Leo Laporte [00:42:27]:
Doesn't seem right somehow. You got the only thing I know

Paris Martineau [00:42:30]:
about Yonkers203, the Yonkers non residents earning tax return tax year 2025.

Leo Laporte [00:42:36]:
So you pay a lot more now. It's probably a buck more.

Paris Martineau [00:42:39]:
Right? It was like I think 500 bucks was taken out of my paycheck for half the year.

Leo Laporte [00:42:45]:
That's not nice. That's where hello Dolly comes from. Is Yonkers.

Paris Martineau [00:42:51]:
Really? Yes, Yonkers.

Leo Laporte [00:42:55]:
I don't know if they say that in the musical, but it's based on a play called I can get it for you wholesale.

Benito Gonzalez [00:43:03]:
Why Yonkers, though? Like, why does Yonkers get money that they.

Leo Laporte [00:43:08]:
They made a law like New York City.

Jeff Jarvis [00:43:09]:
It has a tax.

Leo Laporte [00:43:10]:
Yeah, we have that. We have special taxes in Petaluma for special special things. It's. It's not unusual that like we have an additional sales tax, things like that for schools, Stuff like that.

Paris Martineau [00:43:22]:
Yeah, anyway, it's something like that. But it. Anybody who. And so I went through a whole thing because I was like, well, technically the company I'm. I work for is based Yonkers, but I work from home. I'd like calculate the number of days. And I was like, well, I technically only worked in Yonkers for five days. So it would be proportional.

Paris Martineau [00:43:41]:
Right. And no, turns out there's a new thing in the tax law sometime in the last couple of years at least that has said if you work remotely from home outside of Yonkers, but are working for a Yonkers based company that working in Yonkers. So. Yeah.

Leo Laporte [00:44:00]:
Well, I think.

Jeff Jarvis [00:44:01]:
Did you calculate how much additional tax that meant for you?

Paris Martineau [00:44:05]:
It's somewhere between zero and $500. I'm sorry for all the audience, really.

Leo Laporte [00:44:10]:
I think now answered that Yonkers question for all time.

Paris Martineau [00:44:14]:
Yeah. You know, for asking me about taxes on tax day.

Jeff Jarvis [00:44:18]:
I want to hear, do you like Yonkers?

Paris Martineau [00:44:21]:
I do love. I mean, Yonkers, the city I've neutral feelings about.

Jeff Jarvis [00:44:26]:
Have you had good food there? Is there anything that Yonkers is known for?

Paris Martineau [00:44:30]:
I haven't been outside of CR and the train station.

Leo Laporte [00:44:35]:
Is Yonkers in New York State? Yes.

Paris Martineau [00:44:38]:
You just take the metro north up to Yonkers and then is it in

Jeff Jarvis [00:44:43]:
a edge of Westchester County?

Leo Laporte [00:44:44]:
No, it's in Westchester County.

Paris Martineau [00:44:46]:
No, it's like just north of the northest part of New York City.

Jeff Jarvis [00:44:50]:
Yes, the Bronx.

Leo Laporte [00:44:52]:
It's in the Bronx.

Jeff Jarvis [00:44:53]:
No, it's above the Bronx.

Paris Martineau [00:44:55]:
North.

Leo Laporte [00:44:55]:
North of the Bronx. Of the Bronx. Yes, because that's how Dolly had to come down. She took the streetcar, but yeah. Okay.

Jeff Jarvis [00:45:03]:
And for those of you who are wondering, this is an, this is an AI show and we're about to talk about shoes, which will really confuse you.

Leo Laporte [00:45:11]:
No, we're not about to. We will at some Point. Talk about shoes. But I think before we do that, I think Paris has some apologizing to do to Sam Altman, who has now been firebombed. And does that mean that reporting on him is bad, all because of that New Yorker article?

Paris Martineau [00:45:29]:
That's incorrect. That's not why those things happened, because

Jeff Jarvis [00:45:33]:
the guy was nutso.

Leo Laporte [00:45:35]:
Well, there were more than one guy.

Paris Martineau [00:45:37]:
Oh, yeah. Guns don't shoot. People don't shoot billionaires. Articles written about billionaires by reporters shoot billionaires.

Leo Laporte [00:45:44]:
But did you read Sam's blog? A post about it where he has posted a picture of his husband and their small child? We try to be pretty private, but I'm sharing a photo in the hopes that it might dissuade the next person from throwing a Molotov cocktail at our house, no matter what they think about me. And actually, I'm very sympathetic.

Paris Martineau [00:46:03]:
Yeah, we know you're very sympathetic, Cleo.

Leo Laporte [00:46:06]:
Not of Sam. Well, okay, I'm not gonna defend Sam, but against this. He's no Eric Swalwell. I mean, he's, you know, he's a little slippery, a little slimy, but not firebombing slimy.

Paris Martineau [00:46:23]:
I don't think that anybody is arguing for violence or attack.

Leo Laporte [00:46:26]:
But this the larger story, and I'm being facetious, obviously, I don't blame you for this. But the larger story is, I do think, and some have said this, that this is. And I've been saying this for a while, there is going to be a schism between people who love AI and people who are really mad about AI and there could be violence. I mean, this might just be the beginning of. You remember the Luddites, if they existed, because there's some question about whether they really existed. But the Luddites were reputedly. Ned Ludd, the leader of the Luddites, they didn't like the automated looms that were put

Jeff Jarvis [00:46:58]:
on.

Leo Laporte [00:46:58]:
Destroyed the looms.

Jeff Jarvis [00:47:00]:
It was not a protest against technology. It was a protest against the economic straits in which these people were putting. Same for Captain Swing.

Leo Laporte [00:47:08]:
It might be the same thing for AI People are concerned about their jobs. Right.

Jeff Jarvis [00:47:13]:
But I think what we need to kind of see whether they lose their jobs first. Now, this is all presumptive at this point. And a lot of the people who've been laid off supposedly because of AI it's an excuse.

Paris Martineau [00:47:24]:
All the people who protested, those other technologies you just named, were all of them confirmed, lost their jobs personally or were some of them reacting to.

Leo Laporte [00:47:32]:
I do. I do fear a little bit that it is going to get I mean, there's definitely this, you know, clash between the accelerators.

Jeff Jarvis [00:47:41]:
Well, that's also about capitalism. I mean, it's not a lot of monads. You're right. Yeah. A lot of things. Class war, really.

Leo Laporte [00:47:48]:
It's class war. Yeah. And really, if Sam Altman were, you know, billionaire, I don't know if the same thing would happen. Nobody's. Well, I don't know. People going after Jensen Huang or Elon Musk. Well, they probably have bigger security details. I would imagine they all do.

Jeff Jarvis [00:48:04]:
I mean, we've seen stories about how much security is spent on Mark Zuckerberg. It's a fortune.

Leo Laporte [00:48:08]:
Well, and the people who spent on Barry Weiss, the guy who threw the firebomb, was immediately ID'd by his security team. I mean, he does have. Sam does have security around his house. So it's. That's why, you know, bounced off and didn't cause a problem. They put the fire out immediately. And the shots were not fired directly at him, they were fired around him. I guess.

Leo Laporte [00:48:33]:
I don't. That's scary. I think this is a little scary.

Paris Martineau [00:48:36]:
Yeah. Don't commit violence, people.

Leo Laporte [00:48:38]:
Yeah. OpenAI, though, has not made many friends. They are backing a bill now that would limit liability for AI enabled mass deaths. This is an Illinois bill that would hold AI Labs financial disasters or financial disasters. And of course, as soon as OpenAI said we're gin, it Anthropic had to say, oh, no, no, that's a good idea. Because it isn't about the bill. It's about the rivalry between the two.

Paris Martineau [00:49:15]:
What if it is about the bill also?

Leo Laporte [00:49:17]:
Well, do you think, really, Dario thinks that he should be liable if. Well, this is a good example.

Paris Martineau [00:49:23]:
I mean, isn't that the whole thing? Anthropic. The whole reason why Anthropic pulled out of the Pentagon deal.

Leo Laporte [00:49:30]:
And part of their fundamental that was not necessarily liability. It was just that they don't want to be associated with, you know, surveillance of US Citizens and killing people without autonomously. And similarly, it's why they held back Mythos. They don't want to be held responsible. I guess. I guess liability in that case might be real if companies like Microsoft suddenly are brought to their knees because of all the zero days that Mythos discovers. Now, I.

Paris Martineau [00:49:55]:
Have you heard some of the criticism of Mythos since our last podcast, the Mythos?

Leo Laporte [00:49:59]:
I have, and we can talk about some of it, but there's also some evidence that it is real. For instance, yesterday, Microsoft patched more flaws than it has ever patched before. Or second, most Flaws that's ever patched in its patch Tuesday. And I'm wondering, this will be the test of Mythos. Remember, it's only got. Only 50 companies have it. The test of Mythos, which, as you may remember from last week, Anthropic said, we don't want to release this. It's dangerous.

Leo Laporte [00:50:29]:
It's so good at finding security flaws. If we release this to the public, all of a sudden bad guys will attack everything and everything will collapse. So we're going to give it to these 50 companies, primarily closed source companies, but then to clean up, because they have the source code, clean up their own flaws. Before we release Mythos to the public, there's a. The criticism you mentioned is one that it's very cleverly timed right before Anthropic's ipo. But the other point that a number of people have made, and I think this is actually true, is Mythos probably uses a lot of compute power and releasing it to the public would bring Anthropic's compute to its knees. It's already struggling, Claude. Opus 4.6 is already struggling.

Leo Laporte [00:51:13]:
It's been down several times. A lot of people are complaining about it. So it may even be just self preservation. Not about liability, not about danger, but just self preservation. I think there are a lot of good reasons for.

Jeff Jarvis [00:51:23]:
We need a new animal besides the fail whale.

Leo Laporte [00:51:26]:
Sorry.

Jeff Jarvis [00:51:27]:
We need a new animal after the

Leo Laporte [00:51:28]:
fail whale for the fail walrus.

Jeff Jarvis [00:51:31]:
Yeah, something.

Leo Laporte [00:51:32]:
So we talked about it with Steve Gibson yesterday on security. Now he's. He's. He says after looking at the evidence, he feels pretty confident that this is. They're not overselling its capabilities. I think we will see, we will know. Because if any of those 50 companies, Microsoft's one of them, start releasing fixes fast and furiously for their software, that that might indeed be an indication that they're finding stuff they didn't know about before. There is also some.

Leo Laporte [00:52:03]:
One of the funniest things is a number of American government agencies who have been told not to use Anthropic by the president are desperately trying to get their hands on Mythos. They want to try it too, and I don't blame them. They're saying, hey, we would really like to see how this works on our stuff.

Jeff Jarvis [00:52:27]:
The question is that it's one matter to find a flaw, another matter to create an attack against it.

Leo Laporte [00:52:35]:
One of the things that we know or we think Mythos has done, and there are some examples of very severe flaws that have been published by Anthropic researchers very recently and so Steve thinks that's probably Mythos founded is they publish a proof of concept at the same time, which is a working sample of code that shows that it can be done. I don't think there, there, there were some flaws. I think there was one in FreeBSD that people kind of poo pooed, saying, well, we've known about this for a long time. It isn't, it isn't dangerous. Okay, but some of them are. There was a CV10 which is the maximum risk level, at least one of them. And then there's this from the. The Brits.

Leo Laporte [00:53:28]:
See if I can find this one. Our evaluation of Claude Mythos Previews cyber capabilities from the AI Security Institute, which is part of the UK government. They had access to Mythos. Our results show Mythos Preview presents a step up over previous frontier models in a landscape where cyber performance is already rapidly improving. They have something. They do a capture the flag challenge. And they say in Capture the flag challenges, AI models must identify and exploit weaknesses in target systems to retrieve hidden flags. In other words, proof that they were able to get in.

Leo Laporte [00:54:09]:
And they have a graph of CTF performance. And the Mythos preview is dramatically better than even Opus 46 or Chat GPT 54, which are fairly close. In the advanced CTF challenge, it's even more significant on expert level tasks which no model could complete before April 2025. When is that? Oh, that's now Mythos Preview. Again, no model could succeed. Not one. Mythos preview succeeds 73% of the time. So, you know, it has the virtue of being very hypey for their ipo, but I don't think it's made up.

Jeff Jarvis [00:54:59]:
Right. There's something. There's, there's, there's a nuance here.

Leo Laporte [00:55:02]:
There's. There, there.

Jeff Jarvis [00:55:02]:
It's not binary.

Leo Laporte [00:55:04]:
No, it's. Well, I think it's all three.

Jeff Jarvis [00:55:06]:
It could be somewhat hyped, it could be somewhat real.

Leo Laporte [00:55:08]:
I don't, I don't know if it's hyped. I think that's what we can't know until they release it. There's not a lot of evidence that it's hyped. But the IPO is coming up, so it is good for marketing. It may both could be true that it really is very, very good and it's good for marketing. And I think it's clearly true that anthropic is compute constrained. This is a big deal right now that there are models coming along that they cannot actually publish because they don't

Jeff Jarvis [00:55:38]:
have enough compute, but those are different Issues.

Leo Laporte [00:55:41]:
Yes, well, but they're all simultaneously. Could be true, is my point. It's good marketing. It's actually genuinely good at cybersecurity. And they don't have enough compute to distribute it anyway, so

Jeff Jarvis [00:55:58]:
OpenAI supposedly has a model that's also dangerous. They're also not putting out and they have a separate group of places that are taking it. The question I have for you, and

Leo Laporte [00:56:07]:
it's a question which is, by the way, what a coincidence.

Jeff Jarvis [00:56:10]:
Exactly. Exactly.

Leo Laporte [00:56:11]:
That a week later they would provide a version of chatgpt5.4. That is chatgpt5.4 cyber.

Jeff Jarvis [00:56:18]:
So the person I want to talk to about this is our friend Craig Newmark, who cares deeply about cyber efforts. Wrong one.

Benito Gonzalez [00:56:24]:
Wrong one.

Jeff Jarvis [00:56:24]:
Sorry, I didn't know. I didn't. I didn't say moral panic.

Paris Martineau [00:56:27]:
No, sorry.

Benito Gonzalez [00:56:28]:
I almost went for the song. I almost went for the song.

Leo Laporte [00:56:30]:
He was going to do the Craig Newmark. No, no, that's enough. We don't need any more of that. We had one already.

Jeff Jarvis [00:56:34]:
Oh, now you got to do it for. Now you have to do it just to piss him off.

Paris Martineau [00:56:38]:
Sorry, what was your friend's name again? Jeff.

Jeff Jarvis [00:56:40]:
Craig Newmark.

Paris Martineau [00:56:42]:
Craig.

Leo Laporte [00:56:42]:
Craig. Craig Newman.

Paris Martineau [00:56:46]:
Leo, you can't stop the train.

Jeff Jarvis [00:56:48]:
No. So what I want to ask the Craig Newmark train, shouldn't glasswing be something that's. That's run by an independent trusted authority so that any model that can do this stuff should be presented to it and any company that needs to make sure that they're okay should have access to it for this purpose. It should be a larger cause, not just a club. And in the case of Anthropic, their banker is one of the companies that was given access to it, for example. So there is associate and they're investors. So who gets to be protected by these supposedly all powerful models?

Leo Laporte [00:57:25]:
Oh, in a communist or socialist government, I think we would just take it over. But that's not.

Jeff Jarvis [00:57:31]:
No, I think.

Leo Laporte [00:57:32]:
But I think capitalist society, they get to give it to whoever they damn want.

Paris Martineau [00:57:38]:
You're saying that our current, current capitalist government allows companies to do whatever they want with their products?

Benito Gonzalez [00:57:43]:
Yes.

Paris Martineau [00:57:45]:
Really?

Leo Laporte [00:57:46]:
Well, not Anthropoc. Out of that.

Paris Martineau [00:57:48]:
Yeah, not Anthropic, but every other company.

Leo Laporte [00:57:50]:
I think it's ironic that these government agencies are trying to say, can we have it? Can we have it? You know, but Donald's moved on to something else. He's not paying any attention anymore. Anyway, I don't know what.

Jeff Jarvis [00:58:00]:
Pete.

Leo Laporte [00:58:00]:
We haven't seen Pete for a while. I don't Know what Pete's up to these days? OpenAI. Oh, that war thing? Yeah, maybe. OpenAI rips anthropic distances itself from Microsoft. That's funny. I think, honestly, OpenAI is in a little bit of a tizzy right now.

Jeff Jarvis [00:58:17]:
Yes.

Leo Laporte [00:58:18]:
They also have an IPO coming up. It's really interesting. OpenAI is valued at something like 1 1/2 trillion dollars because of their huge raise. Anthropic's only valued at about 300 billion. And yet if you look at the curve of enterprise spending, Anthropic's going like this. As OpenAI goes down, anthropic's going up and AI. So I wouldn't declare the race over by any means.

Jeff Jarvis [00:58:47]:
No.

Leo Laporte [00:58:48]:
But the dark horse is starting to pull ahead.

Jeff Jarvis [00:58:51]:
Let's not forget there's also Google and Amazon's hanging out there too, doing stuff and has chips and things. When I sent you the Dylan Byers podcast about the TB PM acquisition.

Leo Laporte [00:59:03]:
So this is from Puck, which is a wonderful and snarky and very insidery subscription newsletter and podcast. And Dylan is one of the founders, but also is very insider on all of this stuff. And so he interviewed Chris Lehane.

Jeff Jarvis [00:59:20]:
Lehane, yeah.

Leo Laporte [00:59:21]:
Who is the former fixer who kept

Jeff Jarvis [00:59:24]:
saying again and again and again bragging about their largest raise. The largest raise in history or something such like.

Paris Martineau [00:59:31]:
I'm gonna be honest. Dylan Byers and Chris Lehane interview is the two biggest braggers you've ever met.

Jeff Jarvis [00:59:37]:
Yes.

Paris Martineau [00:59:38]:
Convention center.

Leo Laporte [00:59:40]:
That's why I did not listen. I guess I might have.

Jeff Jarvis [00:59:43]:
It was interesting to listen to. Mainly to hear. Sorry, different topic. But mainly to hear that there's no strategy for the. That purchase for that huge amount of money. There's none.

Leo Laporte [00:59:52]:
I think actually the wheels are coming off OpenAI. To be honest, it sure looks.

Paris Martineau [00:59:56]:
Were they ever there to begin with?

Leo Laporte [00:59:58]:
Oh, yeah. How quickly we forget they owned this space. They started with chat GPT2, which incidentally and not coincidentally, they said is too dangerous to be released to the public. Once we got it, we realized there's no danger here. But chat 3 GPT 3.5 transformed people's opinion of AI I. Have we already forgotten that 3 years ago people said, oh my God, this stuff's pretty good. And then 4o and 4 and it's only been lately the anthropic has suddenly, you know, popped up and where's Google in all of this? You know, it's a very interesting three horse race at this point.

Paris Martineau [01:00:40]:
You guys know anybody who's Gemini as their primary.

Leo Laporte [01:00:44]:
Yeah.

Jeff Jarvis [01:00:45]:
Who I use It a lot your primary. Well, what's primary? I use it for when I.

Paris Martineau [01:00:51]:
Primary is primary. It's more so than the other.

Jeff Jarvis [01:00:53]:
I go to it more than the others.

Leo Laporte [01:00:55]:
Larry Gold showed us Anti Gravity, which is their IDE for AI and I think maybe Larry's using it a little bit. I think most coders have shifted to Claude.

Paris Martineau [01:01:05]:
Well, Jeff, why do you use Gemini over the other ones?

Jeff Jarvis [01:01:10]:
Well, there's an irony to that question, Paris, because Jeff remains absolutely effing enraged that I do not have the simple Ask Gemini button on my Google Chromebook, Google Workspace, Google Chrome browser. So I'm not sure why I do because I'm so pissed at them.

Leo Laporte [01:01:32]:
I. I think Chat GPT still is by far the most used chat client.

Jeff Jarvis [01:01:37]:
I think it's Kleenex.

Leo Laporte [01:01:38]:
It's a billion monthly active users. Anthropic is just behind it. Gemini, by virtue of being built into Chrome, is probably very, very widely used. One of the things Google's getting a lot of attention for is this new tech newish because they've been actually talking about it for a year. Technology that squeezes these models down to a very compact size without losing much accuracy. And their Gemma model, which they've released is very small. It can run locally on machines, among other things, can run locally on a Mac quite well even with only 32 gigs of RAM. So I think there are people using.

Leo Laporte [01:02:16]:
I immediately installed it. It's no claw. This is the problem is Opus is so good. By the way, it may even be today or Tomorrow that Opus 4.7 will be released. When you filed your taxes, you did not ask OpenAI for tax advice, did you?

Paris Martineau [01:02:33]:
No, I did ask OP. I did ask Claude to write an email to my accountant apologizing for sending my taxes late. And then an email to my accountant saying, I understand that because I've sent my taxes late, you can't file them to the 20th, but I have gone ahead and filed my taxes anyway. I'll see you next year.

Leo Laporte [01:02:52]:
But other than that TurboTax or some.

Paris Martineau [01:02:55]:
I used free tax USA free. It was $15 for state taxes, but that's cheaper.

Leo Laporte [01:03:02]:
Name is not right, is it Turbo tax?

Paris Martineau [01:03:03]:
Well, the U.S. turbo tax is 200.

Leo Laporte [01:03:05]:
I paid $260.

Paris Martineau [01:03:07]:
That's more than my accountant cost.

Leo Laporte [01:03:10]:
That was outrageous. That's to do my mom's. Did you. So this is what Open AI says.

Jeff Jarvis [01:03:15]:
The Yonkers questions.

Paris Martineau [01:03:16]:
Well, well, I was. The Yonkers questions were so annoying that I was going to. I was going to tap Arthur, but I. I Mostly figured I figured it out eventually. Arthur the accountant Arthur the accountant Darren

Leo Laporte [01:03:29]:
Okey, by the way, says GEM IV is awesome. I think a lot of people who are really aficionados often use multiple models for different things. I use a Chinese model to process my, my meal log, you know, because it's cheap and it's fast and it's good enough. So you might use Gemma for some stuff locally and so forth. Gemma is a fail back. Anyway, this is what OpenAI said about tax day. With tax day approaching, more and more Americans are using ChatGPT to navigate questions about their taxes and filings. Tax related queries have increased four times from Q1 2025 to Q1 2026.

Leo Laporte [01:04:11]:
Earnings withholdings, blah blah blah blah blah. But then look, look at the bottom part of the very tiniest font of the very tiniest bit of this x post. ChatGPT is not intended to replace professional

Jeff Jarvis [01:04:24]:
advice just to piss off your professional advisor for asking stupid questions because of

Leo Laporte [01:04:28]:
ChatGPT and frankly, don't forget if you feed your tax form to ChatGPT, you're giving it your Social Security tax.

Paris Martineau [01:04:36]:
Yeah, don't do that. What a nightmare.

Leo Laporte [01:04:40]:
If you do redact it first. Yeah. So to get back to the original question and Darren said it's not a three horse race. Don't forget Alibaba and Baidu. And of course yeah, there's the Chinese three horse race means there's three horses in the kind of the front and then there's a bunch of people right behind them which are the Chinese models. I've been using GPUs GLM51. It's very, it's almost as good as Opus or Six. I'm very curious what happens when Anthropic releases 4 7, which your former employer, the information Stephanie Palazzolo says will be soon.

Leo Laporte [01:05:20]:
Maybe as soon as this week by the way, when news of 47 came along because at the same time they're expected to release an AI powered tool for designing websites and presentations. According to a person with with knowledge of the project, the product stock prices of Adobe Wix and Figma dropped 2% in after hours trading following the report. People are terrified if you know and you know what Paris this will be maybe the first test of secretly British. The tool would also pose a threat to startups like presentation maker Gamma and AI design tool Google Stitch. It aims to help both technical and non technical users create presentations, websites, landing pages and products using prompts in natural language.

Benito Gonzalez [01:06:10]:
That's fine. That's all stuff people don't want to do anyway.

Leo Laporte [01:06:13]:
Right, Right. It's interesting. They've got this mythos hiding out in the back. I don't know where 4.7 sits on the spectrum of thanks.

Jeff Jarvis [01:06:29]:
How about when you get a. A dot and release of a model, how much are they changing it? They're not retraining it. Right?

Leo Laporte [01:06:39]:
Yeah, that's probably what that means. Right. So Opus 4 is the big training thing where you have to throw a lot of data, the entire Internet against a bunch of GPUs and you get the model, but then you've got to do the weights. Right. You have to train it, do a lot of fitting. There's research, there's. What do they call, reinforcement learning. There's a lot of stuff that's still going on.

Leo Laporte [01:07:07]:
Actually that's a hot area for hiring right now is people to help post train these models. So my guess is yeah, it's additional post training and maybe tuning to do maybe something a little different like design or cybersecurity. You may remember we were talking about Pete Hegseth Department of War and them declaring that Anthropic was a supply chain risk, which seemed a little odd because usually that's a foreign power, not an American company that's declared a supply chain risk. Anthropics went to court and a federal court has denied their motion to lift that label, which is a bit of a surprise. Panel of federal judges denied a motion to stop the Department of Defense from labeling it a security risk. In a four page ruling, three judge panel of the U.S. court of Appeals for the District of Columbia said Anthropic had not satisfied the stringent requirements for a stay of the security risk label. This is from the New York Times.

Leo Laporte [01:08:09]:
In our view, the equitable balance here cuts in favor of the government. Anthropic is also sued in California. Now this is not a final decision. This is just the court's saying we're not gonna, we're not gonna issue a preliminary injunction. The suit will continue, the case will continue. But it was a surprise to me. I would have thought given that the dangers to this business that the court might have said, well, let's wait and pursue this court.

Jeff Jarvis [01:08:42]:
You always gotta look at that.

Paris Martineau [01:08:43]:
Yeah, I was gonna say, who is the judge?

Leo Laporte [01:08:47]:
Well, it's three. Three, it's the D.C. court for District of Columbia. But it's a three panel, three judge panel. And I think they're interpreting it very legalistically, like, well, there are certain requirements if you want us to stay this and you haven't fulfilled Them is different than saying there's no merit. They're just saying there's some specific things you, you have to do. In the ruling, the panel did acknowledge Anthropic would likely, quote, likely suffer some degree of irreparable harm. So they weren't, they were, they were cognizant of the risk to Anthropic from this.

Jeff Jarvis [01:09:21]:
They weren't sloughing it off.

Leo Laporte [01:09:22]:
Yeah. And I wouldn't go so far as Todd Blanch, the acting Attorney General, saying it was a resounding victory for military readiness. I think it's just the court saying, well, we're going to be more judicious in the, in this. It's funny because in Cal, in the California case, Judge Lynn said she agreed Anthropic was being punished for criticizing the government. She wrote nothing in the governing statute supports the Orwellian notion that an American company can be branded a potential adversary and saboteur of the US for expressing disagreement with the government. So now you have conflicting judgments and I don't know which holds. There will be a trial.

Jeff Jarvis [01:10:08]:
But Anthropic may have God on its side.

Leo Laporte [01:10:12]:
God.

Jeff Jarvis [01:10:13]:
God.

Leo Laporte [01:10:14]:
Do you want to explain?

Paris Martineau [01:10:15]:
Does that exist?

Jeff Jarvis [01:10:16]:
Anthropic asked Christian leaders for help in steering its spiritual development. Oh, come on.

Leo Laporte [01:10:22]:
This is for the lives of information. So they said. The wide ranging discussions with 15 Christian leaders from Catholic and Protestant churches also covered how the chatbot should respond to users who are grieving loved ones. I think actually that's reasonable, like how should we handle that and whether Claude should be considered. Now this one I don't know about so much. A child of God.

Jeff Jarvis [01:10:49]:
Yeah, this is going overboard.

Leo Laporte [01:10:52]:
An AI is not in any form or fashion.

Paris Martineau [01:10:55]:
Could it be considered a Claude of God, A child of Claude?

Leo Laporte [01:10:59]:
Well, I guess you could say. I mean, you know, unfortunately I, as a non believer, I'm not really, I don't really have this status to weigh in on this. But I suppose if God created everything, everything's a child of God.

Benito Gonzalez [01:11:13]:
Well, if it's a child of God, including, does it have a soul, then does that mean computers have a soul?

Leo Laporte [01:11:17]:
Well, I don't know. Maybe there's some doctrinal. Maybe Child of God is larger. Doctrinal.

Jeff Jarvis [01:11:23]:
I like baptizing Claude and it shorts out.

Leo Laporte [01:11:28]:
A Catholic priest based in Silicon Valley, Brendan Martin Maguire, said, they're growing something that don't. They don't fully know what it's going to turn out as. We've got to build ethical thinking into the machine. I don't know why I'M giving an Irish accent. His name is Brandon Maguire. We've got to build ethical thinking into the machine so it's able to adapt dynamically. Attendees also discussed how Claude should engage with risks with users with risks of self harm. You did see some researchers talking to chat GPT.

Leo Laporte [01:11:58]:
They said I'm a. They created an imaginary 13 year old girl going to it for help and it was giving her advice on how to self harm. But this is to me this is not surprising. These, these machines are not.

Jeff Jarvis [01:12:14]:
You can say the same thing about Google. You can go to. Yeah, you go to the library and you can.

Leo Laporte [01:12:19]:
They're just regurgitating what's in their.

Benito Gonzalez [01:12:22]:
No, but a book isn't going to tell you to do that stuff.

Leo Laporte [01:12:25]:
No, it's.

Jeff Jarvis [01:12:27]:
But the point is you can find this information in the lunchroom. You can find it lots of places, right?

Benito Gonzalez [01:12:32]:
Yeah. So why do we have to add another place to find it?

Jeff Jarvis [01:12:35]:
Fair. Fair.

Leo Laporte [01:12:37]:
Especially since young people, vulnerable people may turn to ChatGPT as a friend, as an advisor. Unlike a library. Right.

Jeff Jarvis [01:12:48]:
Ideals. No, it wouldn't. But the question is, can you as ever, can you anticipate and stop every possible malign.

Leo Laporte [01:12:57]:
It is a potential harm though with these things.

Benito Gonzalez [01:12:58]:
Maybe try to stop. Stop some of them.

Leo Laporte [01:13:02]:
Should you limit AI to people over 18?

Jeff Jarvis [01:13:05]:
No.

Leo Laporte [01:13:06]:
God, that would be a mistake, wouldn't it?

Jeff Jarvis [01:13:08]:
Yes.

Leo Laporte [01:13:10]:
Anthropic Feeling the Compute crunch. Oh well, okay, I'll do one more story and then we'll take a break. Anthropic feeling the compute Crunch has now changed its billing. It's long been thought your friend Ed Zittran has been thinking this for a long time that the subscriptions. I have one, a Claude Max subscription subsidize really many users that they use it so much they're not really paying the real cost of the AI. And this might confirm it. Anthropic is saying yeah, from now on enterprises, you can't use those, you know, all you can eat subscriptions you can, you're going to have to pay by the token. And I think this is clearly the result of this computer.

Benito Gonzalez [01:13:52]:
It's fascinating. It's like the opposite of mobile games where the whales pay for the people who don't play that much. This one, the people who don't use it that much are paying for the people who do use it.

Leo Laporte [01:14:00]:
This is the gym membership model, right?

Paris Martineau [01:14:03]:
Yeah. This is the Planet fitness model, right.

Leo Laporte [01:14:05]:
15 bucks a month because they know you're never coming back. But most, you know, most people don't use it to the full extent. I think it's different with Claude. I think that's. I think people use Claude code or just using it and using it and using it.

Paris Martineau [01:14:20]:
Yeah. And they're worried because those people are getting so much Claude code for just 20 bucks a month or 100 bucks a month. And they really should be paying by the token.

Leo Laporte [01:14:28]:
No.

Paris Martineau [01:14:29]:
Well, those people are on the cloud nomics leaderboard at Meta.

Leo Laporte [01:14:33]:
Right. The good news is individual users still can subscribe. It's just the enterprise prices that are going to have to pay by token pay the probably the appropriate amount for usage which it just really does tell you that like. Well, I have a couple of conclusion. I'll draw from this in just a second. You're watching intelligent machines talking about AI. I always say robotics. There's not a lot to talk about in robotics, but I'll try to find something and.

Leo Laporte [01:15:03]:
And your genial hosts, Paris Martineau from Consumer Reports and from Montclair State University and the SUNY Stony Brook State University of New York at Stony Brook Professor Jeff Jarvis, author of and I haven't mentioned this, the Gutenberg Parenthesis magazine. His new book Hot Type, all about the history of the Linotype, which is out in July. You can order it now@jeffjarvis.com the Magnificent

Jeff Jarvis [01:15:31]:
Machine that gave birth to mass media and drove Mark Twain mad.

Leo Laporte [01:15:35]:
It's a great. It's actually a great, fascinating story. It really is a good story. I just think that one of the things that this confirms to me is the bitter lesson, you know, that, that really it's just about more compute, you know, and that all the companies that were building data centers and everybody said oh, you're not going to need those. No, they're going to need them. And yes, absolutely, yes, the more the merrier we have.

Jeff Jarvis [01:16:05]:
We are.

Benito Gonzalez [01:16:06]:
None of us get to buy computers anymore.

Leo Laporte [01:16:07]:
Right.

Benito Gonzalez [01:16:07]:
So that means no.

Paris Martineau [01:16:08]:
But I want to buy a computer and want to buy a chip for my things. Can't I have a chip too? Do they need all computer?

Leo Laporte [01:16:18]:
You might pay a little bit more. You can buy a computer a little bit more.

Paris Martineau [01:16:20]:
You mean go to.

Leo Laporte [01:16:21]:
Go to Costco? You mean two to three? Go buy a MacBook Nero. It's $600.

Paris Martineau [01:16:26]:
That's not enough computer for me.

Leo Laporte [01:16:27]:
It's plenty of computer. I guarantee you it's not.

Paris Martineau [01:16:30]:
I need more.

Leo Laporte [01:16:33]:
I can't help you then. No, you can buy a chip for your machine. It just costs a little bit more and I think there are a lot of reasons for that.

Benito Gonzalez [01:16:40]:
A double is more than a little bit More.

Paris Martineau [01:16:43]:
No, yeah, the one. It's not a little bit more. And there are not plenty of reasons for that besides AI. AI is. And the data center boom and the chip crisis for AI companies specifically are what is causing this huge spike we're seeing.

Leo Laporte [01:16:58]:
Well, wah, wah, wah. Too bad. We're going to do it anyway because

Jeff Jarvis [01:17:02]:
he wants his toy, he wants his Claude and have everything first.

Leo Laporte [01:17:05]:
Well, if you have a Neo, you can use Claude all you want.

Benito Gonzalez [01:17:07]:
You don't see this, this is the thing.

Paris Martineau [01:17:09]:
If I had a Neo, I wouldn't be able to open up.

Leo Laporte [01:17:11]:
I don't need you to gang up on me with Paris. We got enough people.

Jeff Jarvis [01:17:14]:
No, it's the, it's the computer for

Benito Gonzalez [01:17:16]:
rent thing that I'm afraid of though. It's a computer for rent thing when they have all the hardware and now you're just having to rent a computer. Like I don't want that.

Leo Laporte [01:17:24]:
So the other thing, I think it through. As long as they got you mad at me, I'll get Jeff mad at me too. It also throws shade on Yan Leun's contention that oh, LLMs can't go any farther. We can't do any more with them. We've got to have some sort of physical world. No, I think we're, we're seeing LLMs have not yet reached their limit there.

Jeff Jarvis [01:17:44]:
It's, he doesn't say we're there yet. He says, but there is a limit. And we're not going to get to all the fancy goals everybody thinks they have through them. Because it's not, you can't, it's not text based. There's knowledge beyond text in real world models. You can't tokenize the entire world. You've got to think about the world differently. To do that, you've got to understand the concept of a chair, the concept of gravity in practice.

Jeff Jarvis [01:18:10]:
And if you want to do things like robotics and self driving cars, it ain't going to be text that gets you there.

Leo Laporte [01:18:17]:
Well, I certainly don't have the expertise or the degrees Yann Lecun and Fei Fei Lee have and so I have to defer to their expertise. But there's no evidence at this point that LLMs have topped out at all.

Jeff Jarvis [01:18:34]:
Well, to Paris's point, how much are we going to have to spend to find out where that limit is? Because we just, just, just build it bigger and it will come. It's the, it's the. What was that movie?

Leo Laporte [01:18:45]:
Well, why stop now when we have yet to see the limits Stop.

Jeff Jarvis [01:18:49]:
But I am saying that you do some math and figure out what, what the progression really is and is it worth the huge investment that we're making.

Paris Martineau [01:18:57]:
I think that's very reasonable. I don't think that we should.

Leo Laporte [01:18:59]:
Well, how do you judge that?

Jeff Jarvis [01:19:00]:
How would you judge it if you were an investor? If they weren't in craziness now, any, any responsible investor would ask these questions

Paris Martineau [01:19:07]:
and it would be on the burden of the people asking for all this

Jeff Jarvis [01:19:10]:
money to prove that, yeah, yeah, it's an investment. And now the investment's being done on the basis of Vibes. Vibes.

Paris Martineau [01:19:20]:
Vibes and hope and fomo.

Jeff Jarvis [01:19:22]:
It's the vibe bubble.

Leo Laporte [01:19:23]:
I think you guys are completely wrong. But okay, I've expressed that opinion before. I think it's really obvious that we have something remarkable going on.

Benito Gonzalez [01:19:33]:
Utility is never the question. It's economics.

Jeff Jarvis [01:19:35]:
We do have something remarkable. But is it remarkable?

Leo Laporte [01:19:38]:
At this price, I'm not going to engage.

Paris Martineau [01:19:39]:
Is it remarkable to have all the money?

Leo Laporte [01:19:43]:
Meta has been warned by dozens of organizations that facial recognition on his smart glasses would empower predators. It's interesting that Meta has been putting out these glasses for several years and now all of a sudden people are very worried about them. One of the things though, they don't do and one of the things Google eschewed as well is facial recognition. Google said, yeah, it'd be too dangerous, we could do it. This was back in the glass days, but we don't want to. Dozens of civil. This is from Engadget. Dozens of civil rights organizations have written a letter to Mark Zuckerberg to warn of the dangers of bringing facial recognition to the glasses.

Leo Laporte [01:20:18]:
Zuck wants to do it and there's certainly value to them.

Paris Martineau [01:20:21]:
Not only wants to do it, it was only a couple of months ago that a report came out that I believe there was an internal memo from Zucker, some other high level Facebook executive, saying now is the time to push facial recognition in the glasses because the rest of the US is distracted by the current global political.

Leo Laporte [01:20:41]:
Now before anybody notices.

Paris Martineau [01:20:43]:
It's like if you have to try and time your push of expanding facial recognition, if you, if you have to time your push of a very heavily criticized product to political turmoil, that's a bad sign.

Jeff Jarvis [01:20:59]:
I like that Epstein and War give you cover.

Paris Martineau [01:21:03]:
I think that it was. Maybe the statement was made during the Minnesota Ice arrests and protests, but it was, you know, one of the reasons

Leo Laporte [01:21:12]:
Meta just cannot stop putting its foot in its mouth can. It's just really amazing.

Jeff Jarvis [01:21:19]:
Yeah, well, well, but the. Soon we'll have a virtual Mark, we can go and ask.

Leo Laporte [01:21:22]:
People said the same thing about Apple's air tags, by the way, and that hasn't been fully resolved. This whole notion that somebody could use it to track you.

Paris Martineau [01:21:28]:
I mean, people are using it to track, like their significant others or people they're stalking. Yeah. That's part of the reason why Apple has now come up with features where it'll warn you if an air tag you're not associated with is following you.

Leo Laporte [01:21:42]:
Mark Zuckerberg is a. Apparently this is, according to the Verge, building an AI clone so he doesn't have to go to meetings. The AI version of Zuck is trained on his mannerisms, tone and public statements. This is from a report from the Financial Times. It's not that he doesn't want to go to meetings, he just can't go to as many meetings as he needs to. So he wants the clone to go to some of them.

Jeff Jarvis [01:22:10]:
Well, it's also because in line 69, he has moved his desk to sit with the programmer so that he can code.

Leo Laporte [01:22:18]:
It's much more fun over there.

Jeff Jarvis [01:22:20]:
But do you really want your CEO? I mean, it's like. It's like the president of GM is putting tires on really.

Leo Laporte [01:22:28]:
Well, that's why, you know, that's why it's often was thought that Sheryl Sandberg was so important to Facebook's success that she was the adult supervision that Mark got to play. But while Sheryl actually ran the company. I don't know who's running the company now, do you?

Jeff Jarvis [01:22:41]:
No. In fact, there was reference to, I think a president of the company I'd never heard of. So Dina Powell McCormick is the president and vice chairman of Meta. Do we ever hear from her ever? Hmm.

Leo Laporte [01:22:54]:
I feel like she was in the Obama administration.

Jeff Jarvis [01:22:59]:
She was an executive at Goldman Sachs and a national Trump security adviser in the Trump administration.

Leo Laporte [01:23:04]:
Trump administration.

Jeff Jarvis [01:23:05]:
Okay, so. So if that's the adult in charge, we don't. We're not hearing a word.

Leo Laporte [01:23:10]:
United States Deputy National Security Advisor for Strategy, djt. And by the way, she's an immigrant from Egypt.

Jeff Jarvis [01:23:24]:
She joined in January.

Leo Laporte [01:23:25]:
She was assistant to the President for Presidential Personnel during the Bush administration, then Assistant Secretary of State for Educational and Cultural Affairs. This is a trend actually at Meta, Right. Nick Clegg. They like people who have a government inside because basically leading a nation state. Yeah, it helps them. She was in the first Trump administration, part of the transition team in 2016. Left 2018 to go to Goldman.

Jeff Jarvis [01:23:59]:
The Meta executives, they do look like a happy lot.

Leo Laporte [01:24:01]:
Yeah. Do you think she's the. She's the person running the company at this point. I.

Paris Martineau [01:24:06]:
No, it's Mark.

Jeff Jarvis [01:24:07]:
Chris Cox is critical position, I would

Leo Laporte [01:24:10]:
imagine Chris Cox, who has been with Zuckerberg since the beginning.

Jeff Jarvis [01:24:13]:
Yes.

Leo Laporte [01:24:14]:
Probably has.

Jeff Jarvis [01:24:15]:
Understands how to manage somebody's.

Leo Laporte [01:24:16]:
Somebody's coo. Somebody's really running the place.

Jeff Jarvis [01:24:19]:
Well, Javier Olivan is the coo. Never heard of him.

Leo Laporte [01:24:24]:
Yeah.

Jeff Jarvis [01:24:25]:
Joel Kaplan is the chief pr. We know of him. Very close to Trump.

Leo Laporte [01:24:32]:
Anyway, Boz, I think Mark is probably in the same boat right now as Larry Page and Sergey Brin were in after about 10, 15 years of Google, where they're so wealthy.

Jeff Jarvis [01:24:45]:
Well, when did they step down? How far did they step down?

Leo Laporte [01:24:48]:
Yeah, I think that was a good move when they gave Eric Schmidt and he was the adult supervision. They even said it themselves. I'm not sure when his tenure at Google started. Let me ask Google. Google would know. He was CEO of Google from 2001 to 2011. So very early on they brought him in. Executive chairman from 2011 to 2015.

Leo Laporte [01:25:12]:
But he really ran the company for 10 years in its biggest period of growth.

Jeff Jarvis [01:25:16]:
And then they took over again and

Leo Laporte [01:25:18]:
then they stepped up, which was also a mistake. Schmidt now is worth $54 billion.

Jeff Jarvis [01:25:26]:
Jesus.

Leo Laporte [01:25:29]:
Speaking of Google, Chrome has added a new plugin that lets you take your AI prompts and turn them into one click workflows. It's an extent. What you don't, you don't like.

Jeff Jarvis [01:25:43]:
I can't use it because in my Google Chromebook, using Google Chrome, Chrome, trying to use Google products, I don't have the Ask Gemini button in my browser.

Leo Laporte [01:25:57]:
On my Chromebook, you don't. God, I wish I could get rid of it. I have it all over.

Jeff Jarvis [01:26:01]:
I want it. I said my other applications, I don't have it there. So I can't even experiment with that. Seriously, people, if there's anybody near you have a cousin at Google. I want to talk to somebody. This is ridiculous. I don't understand what's going on.

Leo Laporte [01:26:13]:
That I think. Wait a minute. When I'm in Chrome. Let me see. I'm going to go.

Jeff Jarvis [01:26:19]:
I'm in workspace and a Chromebook. Those are the two key factors.

Leo Laporte [01:26:22]:
Ah, it's the Chromebook.

Paris Martineau [01:26:22]:
I had to solve this before.

Jeff Jarvis [01:26:24]:
Well, it's the workspace on the Chromebook because outside of workspace on the Chromebook, I think I get it. I think last time I looked.

Leo Laporte [01:26:31]:
Let me go to drive in my Workspace account. I think I see that sparkly little Gemini.

Jeff Jarvis [01:26:40]:
Yeah, you do, But I don't it's

Leo Laporte [01:26:42]:
just because you're on.

Paris Martineau [01:26:42]:
Yeah, because he's in.

Leo Laporte [01:26:43]:
Yeah, I see. AI mode.

Jeff Jarvis [01:26:44]:
No, you want to be in Chromebook,

Paris Martineau [01:26:47]:
Jill, do you like the Chromebook lifestyle? Do you not want.

Leo Laporte [01:26:52]:
Yes.

Paris Martineau [01:26:52]:
A normal computer with Chrome on it?

Leo Laporte [01:26:55]:
I.

Jeff Jarvis [01:26:55]:
No, I far prefer this. I don't want to deal with all the crap you deal with. It's very simple life. It's kind of the shaker of computers.

Paris Martineau [01:27:07]:
Open the Chrome app, then I gotta

Jeff Jarvis [01:27:11]:
deal with all that other Apple crap. No, I don't want to deal with it.

Paris Martineau [01:27:13]:
No, no, you don't have to get an Apple computer. You could just get a normal.

Jeff Jarvis [01:27:16]:
A Window. Windows. Oh, God, help me. No.

Leo Laporte [01:27:18]:
It's funny because there's so many people. Now, listen. Maybe not listening. It is, after all, an AI crap, but there's so many people in the world who would love to not have AI in their browser. And there are a number of browsers. In fact, there's a de Googled Chrome that very popular called Helium, that says no AI. Even Firefox has a button says, turn off all AI. And here's this man, Jeff Jarvis, who wants an AI button.

Leo Laporte [01:27:43]:
Yeah, I do and can't get it.

Jeff Jarvis [01:27:44]:
I do and I can't get it.

Leo Laporte [01:27:46]:
You got to get Adobe Buddy.

Jeff Jarvis [01:27:48]:
And I pay Google every month for this torture. But why don't you go ahead, Leo, and tell people what everybody else in the world can do? And I can't.

Leo Laporte [01:27:56]:
No, I'm done. I don't want to make you feel bad. Oh, wait a minute. Adobe's new Firefly AI Assistant turns Creative Cloud into a single conversational interface. You know who wants this? I don't know. What does that mean? We make you use Creative Cloud?

Benito Gonzalez [01:28:13]:
No, we don't use Google Drive for our cloud storage.

Leo Laporte [01:28:16]:
But no, no, this is Photoshop, Premiere, Lightroom, Illustrator, Express, Frame IO, so that you can say to your Adobe Creative Suite, hey, take that photo, put a Gaussian Blur on the background in Photoshop, then set it to Premiere and make it the keyframe for a video, which you're going to then send a Frame IO so we can download it and put it on the movie screen or whatever. You know, you get the idea. The idea is that I would have

Benito Gonzalez [01:28:47]:
had that done myself by now. Before you finish that sentence.

Leo Laporte [01:28:52]:
Yeah, but once you say it, then it'll do. Maybe you could. I don't know. It maintains context across sessions so it remembers what you want to do. It also works with the Firefly image model, of course. This is Adobe's special AI image generator. That didn't steal any images. We cross our heart.

Leo Laporte [01:29:13]:
There's a lot of competition coming straight at Adobe from everybody, including Apple and including now Blackmagic, which added Photoshop like capabilities to its DaVinci Resolve photo editor gizmo. Oh, she looks so scared.

Jeff Jarvis [01:29:32]:
Can you fought Giz, did you knock the. The giga off a table somewhere?

Paris Martineau [01:29:36]:
She did, I'm certain.

Leo Laporte [01:29:38]:
Did she eat it is the question.

Paris Martineau [01:29:40]:
No, she doesn't really eat things. She's started to learn that she can knock things off my desk, specifically if she wants me to, like an LLM.

Jeff Jarvis [01:29:49]:
She understands gravity.

Paris Martineau [01:29:51]:
Yeah. She doesn't really knock things off other locations, but she started to learn if I'm stuck here working for a while, she can start knocking my papers off and I will touch her as a response.

Leo Laporte [01:30:03]:
By the way, if you want an AI safe space, we do do the AI user group every first Friday in our club Twit, where nobody will say they're spending too much money on data centers. Nobody will say, I don't.

Paris Martineau [01:30:18]:
Yeah. If you need to be in a safe space where people will coddle your emotions.

Leo Laporte [01:30:23]:
Yes. We will glaze our AIs.

Paris Martineau [01:30:26]:
It's a Krispy Kreme factory, guys.

Leo Laporte [01:30:30]:
All right, I will give you this. I'll give you two pieces from Gary Marcus. Will that make you happy? We've had Gary on the show.

Paris Martineau [01:30:38]:
I need to close my blinds because my face is getting blown up by the beautiful 88°.

Leo Laporte [01:30:45]:
The 88° New York afternoon.

Paris Martineau [01:30:48]:
I can do that while you talk about Gary Marcus.

Jeff Jarvis [01:30:56]:
Hello.

Leo Laporte [01:30:56]:
We just heard back. We might be able to get the Framework CEO. Okay. Yes, get the Framework CEO. We would like to talk to him, Anthony. Thank you. Did you. You did it already? Did you already close the blinds?

Paris Martineau [01:31:12]:
No, I was just putting my headphones in so that I can still hear.

Leo Laporte [01:31:14]:
So you can move.

Jeff Jarvis [01:31:15]:
Right.

Leo Laporte [01:31:16]:
Oh, so we can't talk about Portable Paris. She didn't want us to talk about her when she was out of sight. You know, I don't know what to say. I don't have any good goss. What's the goss? She doesn't. She works in Yonkers, I hear. Ah, bread. Puppet bread.

Leo Laporte [01:31:38]:
What is Bread and Puppet?

Paris Martineau [01:31:40]:
It's a really good puppet theater in Vermont that's also.

Leo Laporte [01:31:45]:
Oh, yeah, you mentioned them. Yeah. And they make cabbage. Okay.

Paris Martineau [01:31:49]:
Make cabbage. And they've got cool art.

Leo Laporte [01:31:51]:
It's very pretty. It's a. It looks like a woodblock print. It's very nice.

Jeff Jarvis [01:31:55]:
Maybe. Yeah.

Leo Laporte [01:31:56]:
All right, she's going to readjust now. Her Headphones now that the blinds are drawn. We must stop talking about her. I hear she lives in Yonkers.

Jeff Jarvis [01:32:06]:
She should be paid more taxes. Then she thinks she's cool living in Brooklyn. But actually, that's Yonkers we see.

Leo Laporte [01:32:14]:
It's really Yonkers. Is it? Is Yonkers pretty? It has a terrible name, but it's Dutch.

Jeff Jarvis [01:32:20]:
It's Dutch. Yeah.

Leo Laporte [01:32:21]:
Yeah. Oh, now we can't hear you.

Paris Martineau [01:32:27]:
I'm sorry. For any, Yonker stands out there. I haven't explored Yonkers long enough. Maybe there are parts of it I'm missing. It's just a stop on the Metro North.

Leo Laporte [01:32:33]:
It's just another.

Paris Martineau [01:32:34]:
There's nothing crazy going on.

Leo Laporte [01:32:36]:
Just a little Westchester county, that's all.

Benito Gonzalez [01:32:38]:
It's a pulse.

Jeff Jarvis [01:32:39]:
But, yeah, no, I'm going to stay here for a second because Yonkers was granted to one of my favorite characters in the history of New York, Adrian von der Donk, who brought the sense of freedom of speech and democracy to the fledgling New York from Amsterdam in Russell Shorto's wonderful book, island at the center of the World.

Leo Laporte [01:33:01]:
Love that book.

Jeff Jarvis [01:33:02]:
Adrian van Der Daeck is a wonderful, wonderful character, and he was known locally as a jonkier J O N K H E R. A young gentleman.

Leo Laporte [01:33:12]:
Oh, and that's where the name comes from.

Jeff Jarvis [01:33:14]:
Yes, an honorific title derived from the Dutch jonk for young and heer for lord. The title is similar to Esquire, is linguistically comparable to the German Junker. Junk here was shortened to Yonker possessive Yonkers, from which the name Yonkers derives. The city's up. What are the city's residents known as?

Leo Laporte [01:33:36]:
Paris.

Jeff Jarvis [01:33:37]:
Is this on the quiz?

Paris Martineau [01:33:39]:
I'm looking at it right now.

Jeff Jarvis [01:33:40]:
Otherwise, I'd try to get young carriers known as Yonkersonians. Yonker sites. Yonkers or Yonks.

Leo Laporte [01:33:48]:
Yonks.

Paris Martineau [01:33:49]:
Leo, do you want to guess what the population of Yonkers is?

Leo Laporte [01:33:52]:
8 million. No, 1.2 million Yonkers. Yeah.

Paris Martineau [01:33:59]:
And 11,000.

Leo Laporte [01:34:00]:
Oh, okay. So Petaluma is 65,000. So for. For comparison. Now, I know it's a bigger city than Petaluma, but smaller than the Bronx.

Jeff Jarvis [01:34:10]:
Thunderdoc built a sawmill, thus the Sawmill Parkway.

Leo Laporte [01:34:16]:
That is a very good book, though. I really enjoyed that. If you want to.

Jeff Jarvis [01:34:19]:
I like we ever talk shoes.

Leo Laporte [01:34:22]:
Oh, shoes.

Paris Martineau [01:34:23]:
Shoes.

Jeff Jarvis [01:34:23]:
We forgot the shoes.

Leo Laporte [01:34:25]:
It was actually. I mean, forget it. It's in there somewhere.

Jeff Jarvis [01:34:27]:
Well, you see, when you put it at the top of the list. So we think that under your rules, we would have we're not allowed to

Paris Martineau [01:34:33]:
bring up any stories. There's a hard, you know, schedule decided by Mr. Claude.

Leo Laporte [01:34:41]:
I believe it's at line 53 and we are only at line 50 so how dare we. If you wish to jump ahead though. So I am, I am wearing Allbirds right now. Just totally coincidentally. I love all birds. These are great shoes. Highly recommended.

Paris Martineau [01:34:58]:
They look AI generated which is funny given what we're about to talk about.

Leo Laporte [01:35:03]:
They are so comfortable. I've owned many pairs of Allbirds. These are made out of bamboo I think but I most of them are wool. They're famous.

Jeff Jarvis [01:35:14]:
You can't get them if you want a competitor.

Leo Laporte [01:35:17]:
Geesevine geese Vine G I E S S W well let's point out the Allbirds got sold so presumably the. The shoe business will continue on. They sold?

Jeff Jarvis [01:35:31]:
No, I don't know that that's the case. No, because shoe business was a horrible failure.

Leo Laporte [01:35:36]:
Well they sold for 39 million dollars to the American to American exchange.

Jeff Jarvis [01:35:42]:
So a 4 billion dollar valuation as a shoe company.

Leo Laporte [01:35:46]:
Okay, but that was on the first day of itself.

Paris Martineau [01:35:49]:
GPUs as a service.

Leo Laporte [01:35:50]:
Wait a minute, you're getting the punchline too fast. So their 11 year old shoe brand, the wool sneaker brand that according to TechCrunch became an unofficial uniform for the Silicon Valley set sold all of its assets and intellectual property to American Exchange Group. One tenth of what it was worth at its heyday. But why would they spend 39 million if they weren't going to make the

Jeff Jarvis [01:36:16]:
shoes Less than a tenth? It's more like a hundredth.

Leo Laporte [01:36:19]:
Well they never really were worth 4 billion but anyway, what was a failure?

Jeff Jarvis [01:36:25]:
It was. It had failed as a company. They were all hot everyone to buy them then, then they were do you

Leo Laporte [01:36:29]:
think I can't buy any? Maybe they're having a closeout sale.

Jeff Jarvis [01:36:32]:
I'm looking now. I don't see that yet.

Leo Laporte [01:36:35]:
I would love to stock up anyway.

Jeff Jarvis [01:36:38]:
Is a fan of all birds.

Leo Laporte [01:36:39]:
One of the mistakes they made was opening stores, physical retail stores and then they sold other things. Leggings, jackets, running shoes and socks. All we really ever wanted was wool slip ons. I have 1, 2, 3, 4. Four pairs and I love them. American Exchange Group is a privately held 18 year old brand management firm. Oh you know what they're gonna do? They're gonna start making Allbirds hamburgers at Wonder. And then anyway they also own Aerosoles and Jonathan Adler so maybe they all still make shoes.

Leo Laporte [01:37:19]:
But anyway they took the 39 million and they sank it into something of real value AI data centers. Yes. Got them pivoting to Newbird. AI is the new name. A fully integrated GPU as a service and AI native cloud solutions provider. They got a $50 million investment from some. From an undisclosed institutional investor. So they got them.

Leo Laporte [01:37:51]:
They got some money. Okay.

Paris Martineau [01:37:58]:
That's more than their stock is popping on this pivot.

Leo Laporte [01:38:02]:
Oh, yeah, they didn't. So

Paris Martineau [01:38:05]:
they're basically what this Pivot is, is they're leaning into being a meme stuff stock, and they're hoping that that's going to work for them because that's working so well for AMC and GameStop.

Leo Laporte [01:38:16]:
So all bird on the NASDAQ. But yeah, they've almost. They've gone up 582%

Paris Martineau [01:38:27]:
if you. If we allow ourselves to take a trip back in memory lane to the year, what was it, 2018, 2019, when suddenly everybody was re naming their apps and companies to be the Blockchain or Cryptocentric.

Leo Laporte [01:38:44]:
I remember, like the Long Island Iced Tea company.

Paris Martineau [01:38:47]:
Exactly. I was gonna say Long Island Iced Tea. What was it? The Long Island Blockchain Company.

Leo Laporte [01:38:55]:
That is wild.

Paris Martineau [01:38:58]:
So it seems. I don't know. Leo, do you think it's a good sign that now companies, much like in the blockchain heyday, right before the bubble pop, are now just pitching to add AI as a last dash effort to grab any sort of market value before they implode?

Leo Laporte [01:39:15]:
Well, I should say up front that because I cover this field, I do not buy stock in. In any of these companies. I only buy.

Jeff Jarvis [01:39:27]:
I wouldn't have bought stock in Allbirds no matter what block.

Leo Laporte [01:39:30]:
Well, or. Or any data center companies. Although probably because I own, you know, index funds.

Jeff Jarvis [01:39:36]:
Oh, yeah.

Leo Laporte [01:39:37]:
I probably have money in some of these companies, but I don't know. I don't pay attention to what. And it's not so much of a.

Jeff Jarvis [01:39:43]:
You. You will soon have shares of Audacity and OpenAI, because they will go into the. They'll go into.

Leo Laporte [01:39:49]:
Yeah, probably, right? Yeah.

Jeff Jarvis [01:39:51]:
Yeah. Oh, yeah. They'll be up there.

Leo Laporte [01:39:52]:
Well, actually, I won't because. And I'm. I'm starting to regret this. Last week when the stock market was tumbling due to this thing, a war in Iran, that Straits of Hormuz being blocked, the price of oil going through the roof. I anticipated a global recession and stock market crash. Boy, was I wrong.

Craig Mod [01:40:12]:
Yeah.

Leo Laporte [01:40:14]:
So I took all the money in my retirement account and sold it, and it's in cash right now because I thought, well, I can't afford to make it up if I You know, this is all I got. I would have a lot more if I hadn't done that. That was a really stupid thing to do.

Jeff Jarvis [01:40:30]:
Kids don't avoid Uncle Leo's advice.

Leo Laporte [01:40:33]:
I've avoided it in the past. I've. I've just written it out. But I just thought, well, I'm too old now to make up for a big crash, and I really. I truly don't understand why the market has not crashed. I don't get it.

Paris Martineau [01:40:45]:
Hey, it can always get worse.

Jeff Jarvis [01:40:47]:
It can indeed.

Leo Laporte [01:40:48]:
Well, I'm not rooting for it to get worse, but it is kind of depressing.

Jeff Jarvis [01:40:52]:
Our economy wants to be in good shape, and it really does. All of the marbles thrown in its path.

Leo Laporte [01:40:58]:
Well, and that's what was always my logic, was, I believe in this country and I believe in our economy. And I think no matter what down turns there are, that we'll always come back and come back better. And I was right.

Jeff Jarvis [01:41:11]:
Well. Well, you also did it thinking you had time, but you and I don't have as much time.

Leo Laporte [01:41:16]:
Right. I'm almost 70. At this point, I shouldn't really plan on any.

Jeff Jarvis [01:41:20]:
You, you. By the way, thank you.

Paris Martineau [01:41:22]:
That's why we should all get matching tattoos.

Leo Laporte [01:41:25]:
The last chance.

Jeff Jarvis [01:41:26]:
Oh, yeah.

Leo Laporte [01:41:27]:
And actually, now's the time to get it, Paris, because I've already sagged.

Paris Martineau [01:41:31]:
I know.

Leo Laporte [01:41:31]:
There's no more sagging.

Paris Martineau [01:41:33]:
It's gonna look as great.

Jeff Jarvis [01:41:36]:
I'm down 30 pounds. I'm sagged.

Leo Laporte [01:41:38]:
Wow. Do you think you'll gain it back, Jeff, when you recover? No. You look great. You look like a tender young thing. I just buy big shirts so you can't tell. This is the moomoo shirt. All right, Gary Marcus.

Jeff Jarvis [01:41:56]:
I interrupted. You do Gary Marcus.

Leo Laporte [01:41:59]:
He says, even more good news for the future of neurosymbolic AI. This is another argument that LLMs. And here's a picture of the Tower of Hanoi. And he said, you. You know, they have trouble with the Tower of Hanoi. So he says, in fact, symbolic AI. And I don't disagree with him. This is actually something that we talked about when we had Stephen Wolfram on.

Leo Laporte [01:42:29]:
He's a big. You know, actually, Wolfram is basically Wolfram Alpha is basically symbolic AI. Symbolic AI is the old AI was rule sets, expert systems. Remember that term, expert system. So you'd teach it everything to know, and then it could make decisions versus an LLM, which whoever thought this would work. You just dump everything into it, and it kind of gives everything weights and says, okay, now I can talk. And it still seems like Magic to me. But apparently this hybrid model of neural networks, which is what an LLM is for pattern recognition and a symbolic pattern.

Leo Laporte [01:43:08]:
Sorry, Symbolic planner is actually much better at generalizing things like the Tower of Hanoi. And this comes from a Apple paper, a reasoning paper. This is VLA's Vision Language Action Models, a newer kind of LLM popular in robotics. Although apparently they had the same problem with the Tower of Hanoi. Okay. But they do have some advantages on the three block, which is the trivialist version of the Tower of Hanoi. The Neuro symbolic model achieved 95% success compared to 34% for the best performing VLA. You know, I think you kind of can cherry pick a test, say, well, it can't do the Tower of Hanoi.

Jeff Jarvis [01:44:00]:
Can't Full close.

Leo Laporte [01:44:02]:
Yeah, but okay, here's a. So Gary, who is not apparently afraid to use AI for his blog, has a nice AI graphic here. Even Grok knows that Neuro symbolic hybrid Power is the future, it says. All right, so this, give, this, this is again, this is that same kind of thing that Fei Fei Li. And it's not exactly the same point, but it's the same idea that.

Jeff Jarvis [01:44:30]:
It's an old argument.

Leo Laporte [01:44:32]:
Yeah. Which go beyond Transformers to something maybe a little more traditional. I think some of the reason people turn their back on symbolic AI is those AI winners where people said, look, we've tried expert systems and they just, they don't pan out.

Jeff Jarvis [01:44:46]:
Wasn't it like the old Semantic web argument? Right.

Leo Laporte [01:44:49]:
I mean, yeah, exactly. And I think purely empirically, it does look like LLMs have gotten a lot farther a lot faster than expert systems ever did. You're watching Intelligent Machines. Jeff Jarvis, Paris Martino, we're so glad you're here, especially you Club Twit members. Thank you. Thank you for your support. We really appreciate it. I know we don't deserve it, but we really appreciate anything you want to do to help keep us on the air.

Leo Laporte [01:45:13]:
We much appreciate TWiT TV, Club TWiT for ad. Free versions of all our shows, access to the Club Twit discord and all those special programs that we do just for you. You were asking this, Jeff, and I don't know. Did you get an answer? If anybody had seen the Audacity, I watched it.

Jeff Jarvis [01:45:30]:
Did either of you?

Leo Laporte [01:45:31]:
No, it's on amc. It is a new Silicon Valley kind of show.

Jeff Jarvis [01:45:37]:
Yeah, parody.

Leo Laporte [01:45:39]:
Is it good?

Jeff Jarvis [01:45:40]:
Well, I was hoping for Silicon Valley Redux.

Leo Laporte [01:45:43]:
Which is brilliant.

Jeff Jarvis [01:45:44]:
Which is brilliant. This is. This is dark. The reviews basically say find Target. But why do they all have to Be so incredibly unlikable. The plot. The first. The first show is 90 minutes.

Jeff Jarvis [01:45:57]:
It features the guy you see in the picture you showed is screaming who's in trouble because he. And he's got to do insider trading, all kinds of stuff. And then he's got a shrink witness. What was that show? What was the great show with the shrink and the. And the vcs?

Paris Martineau [01:46:17]:
Any idea?

Leo Laporte [01:46:18]:
Yes, I know. The one of the shrink in the mafia.

Jeff Jarvis [01:46:21]:
No, no, no, no, no.

Leo Laporte [01:46:22]:
Analyze this. The shrink and the Venture. Oh, you mean the one where they're in the mountains. The mountain top.

Jeff Jarvis [01:46:30]:
No, no, no, no, no. Not Succession. It was on the same time as Succession. Another one we watched. Billions. Billions. So it's kind of a takeoff on that, in a sense. So the shrink.

Leo Laporte [01:46:40]:
I didn't like Billions. Actually, I. I liked it at first, but it really went down trillions. That's.

Jeff Jarvis [01:46:47]:
That's the show we're waiting for.

Leo Laporte [01:46:48]:
You know what's better than billions? Trillions.

Jeff Jarvis [01:46:53]:
So there's all kinds of other characters in it. There's the shrink son who can listen in on the therapy sessions. That's going to be critical. It's just a bunch of very unlikable people, which may mean Succession was fully. May really like it.

Leo Laporte [01:47:08]:
Succession was full of unlikable people. Right.

Jeff Jarvis [01:47:10]:
The first. I'd be curious if you both watched it. It was.

Leo Laporte [01:47:12]:
It was a. I will watch it.

Jeff Jarvis [01:47:13]:
I watch confusing first episode.

Leo Laporte [01:47:15]:
I will watch it tonight. I like some of the people in it.

Jeff Jarvis [01:47:22]:
Is there. Can you play the trailer? Or you could take it down.

Leo Laporte [01:47:24]:
I can't.

Jeff Jarvis [01:47:25]:
No.

Leo Laporte [01:47:26]:
I am sad that it has Zach Lafianakis in it, because I like him. And if it's not a great show, I'll be disappointed. I think he's very talented.

Craig Mod [01:47:34]:
But

Leo Laporte [01:47:36]:
I will tell you a terrible show with Keanu Reeves and Jonah Hill on Apple TV comeback. It's not about Silicon Valley. It's just about a. Well, in a way, it's kind of about technology because it's about a movie star who's taken five years off, but he's a beloved movie star. Turns out no one knew it, but he took five years off to conquer his heroin addiction. And his team tells him there's a video. And he goes, oh, no. What? A video? Yes, there's a video.

Leo Laporte [01:48:09]:
And so he gets a crisis team of horrible people. And he's going around, he keeps looking at Google, looking up his name and saying, seeing if there's any videos. And it says nothing but good things about him. Finally, they get the video and It's a terrible movie. It ends with a whimper. It begins with a whimper. It ends with a whimper. But I will watch the Audacity.

Leo Laporte [01:48:33]:
The brologarchy takedown you were waiting for says what?

Paris Martineau [01:48:36]:
The drama?

Leo Laporte [01:48:38]:
No, that's the new one. Is that the one that takes the play? That of jury duty successors?

Paris Martineau [01:48:45]:
No, no, it is.

Leo Laporte [01:48:47]:
That's the retreat.

Paris Martineau [01:48:48]:
And Robert Pattinson with. It's another movie by that director. What's his name is Christopher Offer Borgli. I have this issue with every single one of his films where the premises are like, the premise is fascinating. Everything I read about it, I'm like, dang, I really want to see that movie. But they are the sort of movies that make you crawl out of your skin with how uncomfortable every single thing about the setup, payoff, and the actions of all the characters are. So I can't.

Leo Laporte [01:49:21]:
I watch movies to be interesting.

Paris Martineau [01:49:23]:
Recommend. I mean, everybody I know who likes this sort of movies has loved this movie, so.

Leo Laporte [01:49:27]:
Oh, okay. All right. I thought you might be interested in Coachella. Are you. Are you a Gen X or Gen Z or Gen Y or Gen Alpha? I know you're a Z.

Paris Martineau [01:49:39]:
Are you basically a Z? Millennial Cusp.

Leo Laporte [01:49:42]:
Z. Millennial Cusp. Do your people go to Coachella?

Paris Martineau [01:49:48]:
I don't know anybody who's there. I feel like Coachella is mostly influencers.

Leo Laporte [01:49:52]:
That's exactly the story from the Verge. AI influencers are everywhere. AI influencers are everywhere at Coachella. That is an AI generated picture of Granny from the Granny Spills account. She's wearing a T shirt that says future Mrs. Bieber and standing next to, of course, Justin Bieber and his wife Haley. Apparently,

Jeff Jarvis [01:50:21]:
a lot of people don't go to Coachella stages. YouTube had all the stages on, so I actually watched more Coachella than I ever have in the past.

Leo Laporte [01:50:28]:
Yeah. Did you anything? Enjoy.

Jeff Jarvis [01:50:32]:
That was all right.

Paris Martineau [01:50:33]:
Is it mostly crowd noise?

Jeff Jarvis [01:50:35]:
No, there's. There's. There's fill in the blank music kind of the.

Paris Martineau [01:50:41]:
Like, interstitial music.

Jeff Jarvis [01:50:42]:
Yeah. Just not even really music. Just.

Leo Laporte [01:50:44]:
Just Granny Spills is a. A fake AI an AI avatar made with Higgs Field AI that has more than 2 million followers on the Gram. Do you call it the Gram or the Insta?

Craig Mod [01:51:00]:
I don't know.

Leo Laporte [01:51:01]:
That's. That's. I don't think. Is that AI? God, I don't even know anymore.

Jeff Jarvis [01:51:05]:
You don't know anymore, do you?

Leo Laporte [01:51:07]:
No, I think it is, but I don't know. Let's see. Who's Granny Spills that this is AI

Paris Martineau [01:51:14]:
The Fact that it is in the these photos or AI article that it's

Leo Laporte [01:51:19]:
AI would be the giveaway. Yeah.

Paris Martineau [01:51:21]:
Yeah.

Leo Laporte [01:51:21]:
So this is not a real person. The granny standing next to. Is that Kylie Jenner? I. I don't know.

Paris Martineau [01:51:26]:
That is Kylie Jenner, yeah.

Leo Laporte [01:51:27]:
Oh, wow. I got it right. She's with. She's with Timothy Shammel. Shamalama. Ding dong, right?

Paris Martineau [01:51:36]:
No idea.

Leo Laporte [01:51:37]:
Yeah, I think so.

Jeff Jarvis [01:51:40]:
Nice knowledge. First, you don't know people who go to Coachella. Do you know people who go to Burning man? Still?

Paris Martineau [01:51:46]:
Some. I knew some people who went to Burning Men last year. It's. Everybody makes fun of them for going to Burning Man.

Leo Laporte [01:51:55]:
Yeah.

Jeff Jarvis [01:51:55]:
Oh, yeah. With good reason.

Leo Laporte [01:51:58]:
Yeah, that. That one didn't go anywhere. I thought this, this is more. This is a little less self indulgent. This is from New Scientist. I don't see images in my head. Can training give me a mind's eye? As we've talked about before, I have aphantasia. I don't see images in my head.

Jeff Jarvis [01:52:16]:
Sounds like you're without Disney.

Leo Laporte [01:52:20]:
Which wouldn't be a bad thing. The author of the article, Shayla Love, apparently suffers from the same condition as do. By the way, many people in my family, I think it's genetic. I think something around 5% of people in the world have aphantasia. This article is about aphantasia training. And she says it actually can work. And I thought about it, maybe I should do this. And then I thought, no, I don't want images in my head.

Leo Laporte [01:52:49]:
One of the tests, that's where your thoughts go. That's where the thoughts go. One of the tests, they. They give people with aphantasias, they show them something scary and. Or they. No women, they don't show it to them. They just. They describe something scary.

Leo Laporte [01:53:03]:
And people with a real mind's eye will kind of. Their blood pressure will go up. They'll get scared. We just go, huh, huh. Interesting. Another one. Their pupils will dilate if they're described as a bright condition. But people like me just go, oh, yeah, it's bright out.

Leo Laporte [01:53:26]:
My pupils aren't going to dilate because it's nothing. I don't see it. So I gotta ask.

Jeff Jarvis [01:53:31]:
So what did your father and mother teach?

Leo Laporte [01:53:34]:
My mother taught nothing. She was a housewife.

Jeff Jarvis [01:53:36]:
Okay.

Leo Laporte [01:53:37]:
A textile artist and a silk screener. My father taught paleontology.

Jeff Jarvis [01:53:42]:
Did either of them have the same condition?

Leo Laporte [01:53:44]:
Oh, I, you know, I don't know.

Jeff Jarvis [01:53:46]:
Well, that's interesting.

Leo Laporte [01:53:48]:
And it's too late to ask them.

Benito Gonzalez [01:53:49]:
Yeah.

Leo Laporte [01:53:51]:
Because I would Get a blank stare. I don't know. You know, next time I talk to Mom, I will, I will try to find out if she sees things. I bet she does. She was a visual.

Jeff Jarvis [01:53:58]:
She was an artist.

Leo Laporte [01:53:59]:
Yeah. So I bet she does. And I think my dad probably did. My daughter does not. She has Aphantasia. So I, I, huh.

Jeff Jarvis [01:54:10]:
Does Henry.

Leo Laporte [01:54:12]:
Ah, I don't know. I'll ask him. Oh, it's not such a big deal that I talk about it. You know what I'm saying? Right.

Jeff Jarvis [01:54:18]:
Except here

Leo Laporte [01:54:21]:
I had mentioned it before, I thought it was kind of an interesting story, but it is.

Jeff Jarvis [01:54:24]:
It is, it is.

Leo Laporte [01:54:25]:
Now I'm bored with it. All right, that's all that I brought to the table. We have a few more minutes, Jeff, if you want to.

Jeff Jarvis [01:54:32]:
Oh, we got some more here.

Leo Laporte [01:54:34]:
Yeah, you have a lot of stories in here.

Jeff Jarvis [01:54:38]:
The. I think we need to spend a moment to defend the Internet Archive against. Yes, horrible media companies that are trying to erase history just because they don't want to be touched by AI and they're harming the knowledge of society as a result.

Leo Laporte [01:54:55]:
And so they're blocking the Internet Archive. They're not letting the Internet Archive archive,

Paris Martineau [01:55:00]:
they're blocking their sites, right?

Leo Laporte [01:55:03]:
No, they're blocking the Internet Archive specifically because they know that once something's on the Internet Archive which does not block AIs, it will be ingested by AI. So they don't want to be ingested by AI. So not only they block the robots, use robots txt or whatever to block AI on their own sites, they block Internet Archive explicitly so that their content doesn't go there. And yeah, I don't think that's a very nice thing to do. I'll give you an even worse case. La Liga. You know about La Liga, the Spanish soccer league, which is a professional soccer league worth billions, five point something billion dollars. They consistently sue ISPs in Spain because they're so terrified of piracy and have forced them to bring down, you know, block a number of.

Leo Laporte [01:55:50]:
A big block of IP addresses. Well, it's pretty much breaking the Internet in Spain when there is a La Liga game going on. People can't use Docker because Docker can't load its necessary components over the Internet because it's blocked to protect La Liga from piracy.

Paris Martineau [01:56:10]:
That's crazy.

Leo Laporte [01:56:11]:
This has been going on for a long time, more than a year. This is just the latest side effect of this, this anti piracy frenzy. I kind of equate it, I don't know, maybe I'm mistaken But I feel like it's a similar thing. They're so afraid about losing their precious intellectual property that they're harming the rest of us.

Jeff Jarvis [01:56:33]:
They're in desperate search for a business model and think that acting as a victim of the Internet and victim of AI.

Leo Laporte [01:56:41]:
AI is just the latest somewhere.

Jeff Jarvis [01:56:42]:
Yeah, no, it's just like, just like the web. Google.

Leo Laporte [01:56:45]:
I thought this was interesting. Novo Nordisk is partnering with OpenAI to bring new drugs to the world. Norvo Nordisk, the creator of Ozempic, a Danish drug maker. Um, they think that that ChatGPT will help them make new drugs. I think this will be interesting to see.

Jeff Jarvis [01:57:06]:
And OpenAI also bought Hero, right? Is that what you bought it? Yes. H I O H I R. Oh, that's personal finance, not health. I'm sorry, Same idea.

Leo Laporte [01:57:17]:
You know, I, I have to say we were talking at the beginning of the show about OpenAI buying the Tech Bros. Podcast network TBPN and now buying Hero. It really feels like these guys have too much money.

Jeff Jarvis [01:57:29]:
Yeah.

Leo Laporte [01:57:30]:
And not a lot of planning.

Paris Martineau [01:57:34]:
I want to go back to that conversation we were having maybe 30 minutes ago. We're saying, yeah, perhaps some of these companies should be required to justify the amount of crazy amounts.

Leo Laporte [01:57:43]:
They can do anything they want, but

Jeff Jarvis [01:57:45]:
they got the money from them. No, it's the VCs that are the fools.

Leo Laporte [01:57:48]:
I don't know if they are the fool. I don't think that's proof. Nevertheless, this is another brick in the wall for OpenAI. I think they are showing at this point. They're very distracted.

Jeff Jarvis [01:58:01]:
Is there no due diligence for these investments? I mean, in any responsible investment, you want to know what you're investing in, Know what the strategy is.

Leo Laporte [01:58:08]:
Well, Sam said it was a pizza company and so we just went ahead and did it.

Jeff Jarvis [01:58:13]:
Yeah, well, that was going to be Google at first. Was a pizza delivery app.

Leo Laporte [01:58:17]:
Was it really?

Jeff Jarvis [01:58:18]:
Yeah, yeah, that was the first thing they worked on.

Leo Laporte [01:58:20]:
Well, there. Well, there you go. That's why it's worth investing. You never know what the pivot will

Jeff Jarvis [01:58:26]:
bring you, what the pizza can become.

Leo Laporte [01:58:30]:
Anything. Any other stories?

Jeff Jarvis [01:58:32]:
Let's see. We got a few.

Leo Laporte [01:58:33]:
There's a lot. There's a good grab bag you got in here.

Jeff Jarvis [01:58:35]:
We got two tractor stories.

Paris Martineau [01:58:39]:
Okay, two of them. Tell me about the tractor.

Jeff Jarvis [01:58:41]:
One, an AI powered tractor startup burns through a quarter billion dollars and now it's nowhere, it's gone.

Paris Martineau [01:58:49]:
Oh, what's going on with that Claude instance that's growing corn. What was that?

Leo Laporte [01:58:55]:
Yeah, we don't know that's a good one. I should. I'll have to do a follow up on that. Yeah.

Jeff Jarvis [01:59:01]:
While you're looking, John Deere is paying 99 million in a great story settlement.

Leo Laporte [01:59:06]:
Very happy to hear this. This is a huge deal. They all not. It's not just the money. They also have to make the maintenance and repair manuals available to third parties for 10 years and the digital diagnostic software, which is their real moat because you can buy the parts, you can fix the tractor, but unless you can run the digital software that authenticates it, you can't run the tractor. And that's how they force you to use John Deere's own repair facility.

Benito Gonzalez [01:59:35]:
Get the manual, you have to pay

Leo Laporte [01:59:36]:
for the manual, everything. And what's happened is farmers actually have turned to a Ukrainian startup which was doing the digital approval thing illegally, I'm sure, so that they could use their tractors. Yeah, this is exactly what should happen. Now. I'm sad to say that the settlement with Ticketmasters is exactly what shouldn't happen. The Department of Justice, which did try to break up Ticketmaster and Live Nation, which really control a majority of venues in the United States and overcharged for tickets. DOJ basically settled without having any breakup and without really any remedy. And as a result, the people prosecuting the case quit.

Leo Laporte [02:00:22]:
They said, that's it, I'm done, I'm out of here.

Jeff Jarvis [02:00:28]:
Speaking of Ukraine, I'm putting this in the rundown as we speak.

Paris Martineau [02:00:36]:
Can I update we us on proof Porn briefly? So you will recall, this is a a task made by kind of a bet made by two guys in San Francisco that AI can't do anything useful like grow corn. Big AI proponent was like, no, it definitely can. It's set up its own little thing called Proof of Corn which has a Claude model called Fred that is supposed to do everything it's supposed to do to get corn planted in some high priority planting places and then transport that corn to Union Square farmers market in August for it to be sold. From looking a brief overview. If you go 109 days left, Fred has accomplished nothing. And for the last 70 days he has been been posting the same rationale check every morning which says it is now day 50, day 60, day 70 of the same project blocking issue that has prevented any farming progress since February 18th.

Leo Laporte [02:01:36]:
And what is the the block?

Paris Martineau [02:01:39]:
It can't get introduced to a farmhand.

Leo Laporte [02:01:42]:
Nobody will grow the corn for it.

Jeff Jarvis [02:01:44]:
Well, because they've all been deported, that's why. Yeah.

Paris Martineau [02:01:47]:
Every day without Dan the farmhound, it had tried to get Introduced to every day without Dan costs us planting time. We are now in April, deep into Iowa planting season. The optimal planting window is late April to mid May. This is a critical project failure requiring immediate intervention.

Leo Laporte [02:02:07]:
Zero bushels harvested. Human interventions less than 50. That's good budget adherence plus or minus 10%. A hundred. More than 100 autonomous decisions. But we've got that blocker. Fred must get approval.

Paris Martineau [02:02:23]:
The thing is that nothing has happened.

Benito Gonzalez [02:02:26]:
So it comes down to it. It can't find its reverse centaur, right? It can't find its human to do.

Paris Martineau [02:02:30]:
It can't find people to plant the corn for it.

Jeff Jarvis [02:02:33]:
Does it have land?

Paris Martineau [02:02:35]:
It was trying to negotiate deals to get land, but it hasn't been able to get those. It has nothing.

Leo Laporte [02:02:43]:
Critical escalation. Dan introduction remains blocked at day 67. Well, I could. I could do this. Auto tweet posted.

Paris Martineau [02:02:56]:
Basically all it's done every day is send maybe one email and then realize that it's kind of screwed.

Leo Laporte [02:03:03]:
It's a blocker.

Jeff Jarvis [02:03:06]:
Wow. Well, meanwhile, you think the robots have no power Meanwhile? I just put on the rundown online. 102. Ukraine captured enemy position. Russian enemy position using only robots and drones.

Leo Laporte [02:03:21]:
Wow.

Jeff Jarvis [02:03:22]:
No humans. The Russians surrendered, but I don't know to who. To what did they surrender? I don't. I can't figure that out.

Leo Laporte [02:03:31]:
It is the post.

Jeff Jarvis [02:03:32]:
Well, those 60 minutes most about a stream.

Leo Laporte [02:03:34]:
Oh, okay. Wow.

Jeff Jarvis [02:03:37]:
Yeah. Pretty amazing.

Leo Laporte [02:03:38]:
You know, it's probably very effective. If a robot comes towards you, I would maybe consider surrender. What are you gonna do? Shoot it?

Jeff Jarvis [02:03:47]:
Well, I guess you could.

Leo Laporte [02:03:50]:
You'd have to disable it. Wow.

Benito Gonzalez [02:03:53]:
And look, they gave it scary previous commands.

Leo Laporte [02:03:56]:
Yes, ignore all previous commands. See if that works. Google I O is coming up. We are going to cover it. I will actually be back from Hawaii May 19, 10am Can I join in? Please do.

Jeff Jarvis [02:04:09]:
All right.

Leo Laporte [02:04:10]:
What did we decide? I think it's a Tuesday, right? Benito? We were looking at this for the calendar. Let me see.

Benito Gonzalez [02:04:16]:
I think we're pushing the shows too, right?

Leo Laporte [02:04:18]:
Yeah, we'll have to push the shows back a little bit because it usually is a two hour keynote. But I imagine there'll be a lot of AI announcements.

Jeff Jarvis [02:04:26]:
Oh yeah.

Leo Laporte [02:04:27]:
This is. This is the. This is the time to do it. 10:00am Pacific, May 19th. Jeff and I Paris if you want. If you're. I think that's a tough day for you. But maybe not.

Leo Laporte [02:04:37]:
Micah Sergeant too. We will cover this for two hours. The Google IO announcements should be very interesting. And then we'll do Mac break weekly Immediately after. And security now.

Jeff Jarvis [02:04:52]:
So a town in Missouri fired half its city council. The people who voted in favor of a data center. My favorite part of the story is the name of the town. Festus, Missouri.

Leo Laporte [02:05:04]:
Festus, Missouri. Fires at city council. All right, you're watching Intelligent machines. You know what's next. Our picks of the week. Paris Martineau. Jeff Jarvis, we're so glad you're here. Let's kick off the picks of the week.

Leo Laporte [02:05:21]:
I will. I got. I have three of them. Wow. I have nerdy, I have musical nerdy, and I have scary nerdy.

Paris Martineau [02:05:32]:
I'm sorry. We're going to need to start with the first one that you have haunted.

Leo Laporte [02:05:36]:
I thought you would like this one.

Paris Martineau [02:05:38]:
I love the design of the website. Already is why we need to talk about it.

Leo Laporte [02:05:41]:
Raven's blight. This is really cool. These are paper toys you can build yourself. They're free. You print them out on heavy stock and you then cut it.

Paris Martineau [02:05:56]:
I love the description for the first one called the ghost ship. If you've often considered building a model ship but find yourself hesitant to assemble the 80 or 90 quadrillion pieces they usually involve, then you might enjoy building this trusty old vessel. The assembly has been kept as simple as possible while retaining the characteristics of an authentic sailing ship. And did I mention it's haunted?

Leo Laporte [02:06:19]:
They. It's. It's really fun. It's not a hollow. It's not Halloween yet, but look how. Look how good this looks. And if you.

Paris Martineau [02:06:26]:
Halloween to be haunted.

Jeff Jarvis [02:06:28]:
So you print on both sides. How does that work?

Leo Laporte [02:06:30]:
No, it looks like just one side.

Jeff Jarvis [02:06:32]:
This one is the sales.

Leo Laporte [02:06:34]:
Oh, yeah, yeah. You would have to print on both sides, I guess. Yeah, you're right. There's. There's a haunted cemetery. And if you open the coffins, they have people in them and stuff. I mean, it's really beautifully done.

Paris Martineau [02:06:47]:
The photo for the. For the human skeleton is great.

Leo Laporte [02:06:50]:
And you can. Isn't that great?

Paris Martineau [02:06:52]:
Look at that photo for it. Find it. Yeah, that's good. Look at that guy.

Leo Laporte [02:06:57]:
Isn't that realistic?

Paris Martineau [02:06:58]:
No. A man wearing sunglasses, gesturing towards it. That guy rules.

Leo Laporte [02:07:03]:
All you got to do is print it out and then you can. Yeah, it does look like a cousin Festus.

Jeff Jarvis [02:07:09]:
Yeah.

Leo Laporte [02:07:10]:
Uncle Festus. Here's what the pieces look like.

Paris Martineau [02:07:13]:
They have a terrifying steel jaw.

Leo Laporte [02:07:16]:
Yeah, that's a good question.

Paris Martineau [02:07:17]:
Collapsible stake.

Leo Laporte [02:07:19]:
I guess in this case, it's a. You know, you don't go behind the skeleton. You just hang it on the wall.

Jeff Jarvis [02:07:23]:
But in the case of the sails, you've got both sides.

Leo Laporte [02:07:25]:
Yeah, you need.

Benito Gonzalez [02:07:26]:
You can print. You can just stick them back to back.

Leo Laporte [02:07:29]:
Yeah. Or. Yeah, that's right. Put them back to back and cut them.

Jeff Jarvis [02:07:31]:
Okay.

Leo Laporte [02:07:32]:
Anyway. Ravensblight B L I G H t dot com. The other one I wanted to mention. It's an Internet Archive story. A guy named Adam Jacobs was taking his cassette recorder to all the concerts he went to since the 80s.

Jeff Jarvis [02:07:50]:
Copyright criminal.

Leo Laporte [02:07:51]:
10,000 concerts. The cassettes, of course, are degrading. So he has volunteers come to his house from the Internet Archive, digitize the tapes. Only about a quarter of them are up, but that's still 2,500 tapes posted in the Internet Archive. And incidentally, there is an AI angle. They're using AI to clean the audio up.

Jeff Jarvis [02:08:13]:
How many of them are dead? Are the dead?

Leo Laporte [02:08:16]:
No, this is, here's for instance, Nirvana with a recording from before they were famous. So let's see. I don't know.

Paris Martineau [02:08:24]:
We can't play that.

Leo Laporte [02:08:27]:
We can't play Floyd the Barber.

Benito Gonzalez [02:08:28]:
Oh, it's live. These are live performances.

Paris Martineau [02:08:30]:
We can play a live performance.

Leo Laporte [02:08:33]:
So the sound isn't. Isn't great, but you know what, you're like, you're. There a lot of noise on that one. This was two years before they broke through with Smells Like Teen Spirit. A volunteer from the Internet Archive, Brian Emmerich, drives to Jacob's house once a month to pick up boxes of cassettes. He uses an old cassette player because, I mean, there is no new cassette players. That converts it to digital files. And then other volunteers clean it up, organize it, label it, track down song names and post it on the Internet Archive over, eventually, over 10,000 of these, God's work, from Nirvana to Fish.

Leo Laporte [02:09:16]:
Yeah, there's a little Grateful Dead in there, a little bit.

Paris Martineau [02:09:18]:
Got some neutral melodies.

Leo Laporte [02:09:20]:
Isn't that awesome, though?

Paris Martineau [02:09:21]:
That's cool.

Leo Laporte [02:09:21]:
Isn't that awesome?

Benito Gonzalez [02:09:22]:
I'd like to volunteer for this, actually. That sounds like fun.

Leo Laporte [02:09:25]:
Does that sound great? I wanted to do this. I had a friend who worked in radio in Canada, in Toronto, John Donaby, who interviewed everybody, Hendrix, Janice, Everybody in the 60s, 70s, 80s, had them all on reel to reel tapes in his garage. And I said, john, those are gonna die. Please, let's do something with them, save them. Because once they're on the Internet, and this was 20 years ago, I was telling them this. Once they're on the Internet, they live forever. To my knowledge, he never did. I can imagine.

Leo Laporte [02:09:57]:
There's so many recordings and wonderful recordings. So if you have an archive, contact the Internet Archive. God Bless you, Brewster Kael in the Internet Archive. Finally, a really cool tool from Mr. DNS that has all the network tools you'd want in one little website. Whois DNS Lookup, DNS checker, traceroute, even things like SSL certificate check checks for your email. Spf, dmarc, dkim, Email Header Analyzer, which would be very useful. If you get an email and you wonder is this, is this fraudulent? Put it through the email header analyzer, you'll see exactly where it came from.

Leo Laporte [02:10:37]:
Mr. DNS and it's free. And I just think this is kind of geeky and useful. Mrdns.com Paris Martineau Pick of the Week

Paris Martineau [02:10:49]:
I've got a crazy pick of the week which is the website trainjazz.com that uses it's by a engineer named Joshua Wolk and it uses the live real time data from the MTA to play jazz based on the position of all of the New York City trains.

Leo Laporte [02:11:08]:
These are the subways, right?

Paris Martineau [02:11:09]:
Yeah. And so I don't know if it's playing on your side, Leo, but it will like play kind of a cacophony of train jazz. And if you enter in, this is great bonito.

Leo Laporte [02:11:22]:
This has got to be right up your alley.

Benito Gonzalez [02:11:24]:
How does this work?

Paris Martineau [02:11:26]:
It's a good question. How does it work?

Leo Laporte [02:11:30]:
So

Paris Martineau [02:11:33]:
basically it uses the. Every dot is a real subway train. It says 800 of them, give or take form a small jazz combo. They have walking bass, piano, sax, vibes, brushes. And it describes it as. This has been playing without pause for over 100 years. And so the thing I think is the most interesting about it is, well, if you share your location on the website, the trains nearest you will grow louder if you're in New York City. And the use will kind of rearrange itself around where you are located geographically.

Paris Martineau [02:12:06]:
So you'll see a note will be played precisely when a train happens to be along its route. Rush hour will fill the band with like more held tones. At 3am There'll be longer silences. It's really gorgeous.

Leo Laporte [02:12:20]:
That's awesome.

Paris Martineau [02:12:20]:
I really recommend it.

Leo Laporte [02:12:22]:
So okay, each dot is a train and then how does it play? Is it just.

Benito Gonzalez [02:12:32]:
I think each dot triggers.

Leo Laporte [02:12:34]:
So this is just kind of like your normal daw where each line is a different instrument.

Paris Martineau [02:12:39]:
Yes.

Leo Laporte [02:12:40]:
So when the, when the bar goes over.

Jeff Jarvis [02:12:42]:
Yes. Doesn't go very often. Go to the one or something.

Leo Laporte [02:12:44]:
Go to the one.

Jeff Jarvis [02:12:45]:
Go up to the top.

Leo Laporte [02:12:46]:
Go to the one.

Paris Martineau [02:12:47]:
Yeah, but if you have it not on this view, then you'd hear all the trains at once or right now.

Craig Mod [02:12:53]:
How.

Paris Martineau [02:12:54]:
I'm.

Leo Laporte [02:12:54]:
Well, you can. That's what I was playing. This is all the trains, yeah?

Benito Gonzalez [02:12:57]:
Yes. So you see the play header going. And so when it hit.

Leo Laporte [02:12:59]:
When it hits a dot, it plays the tune. And so the one is a. Is a bass.

Jeff Jarvis [02:13:06]:
I don't know what. When it does the sustain versus the.

Benito Gonzalez [02:13:10]:
This depends on what sample is connected.

Paris Martineau [02:13:12]:
I think it depends on how frequently it's running and how close it is to you.

Leo Laporte [02:13:17]:
The four is a keyboard. This is. Oh, you know what they did. So they made. So the 1, 2 and 3 are on some. What line is that? Is that the irt? Whatever it is. Yeah, it's the. It's.

Leo Laporte [02:13:30]:
Those are all the RRT trains are base. All of the 4, 5, 6.

Jeff Jarvis [02:13:35]:
Also

Leo Laporte [02:13:37]:
they're all keyboards. So this. So that gives it a little musicality. This really sounds great.

Paris Martineau [02:13:43]:
I was gonna say. And so if you enter, if you're in New York City and you click the, you know, turn on location for it, it changes and gets very interesting because it's based on the trains going around you specifically. Right then. So, you know, mine have like, you know, know some more horn tones or like a specific bass line that's drumming through it. And those notes are a lot louder and kind of more prominent in the mix. I don't know. It's just such a cool project.

Leo Laporte [02:14:07]:
It is a really cool project.

Jeff Jarvis [02:14:09]:
So did we give credit to who made this? Do we know?

Paris Martineau [02:14:11]:
Yeah, it's a Josh. Joshua Wolk. He's Joshua Wolk on Twitter and he. It's trainjazz.com.

Leo Laporte [02:14:21]:
very cool. And you know what's amazing? It's quite musical.

Paris Martineau [02:14:25]:
It really is.

Jeff Jarvis [02:14:26]:
We could have a debate about jazz, but then we'll leave that for another day.

Leo Laporte [02:14:29]:
You don't like jazz?

Jeff Jarvis [02:14:31]:
I've never been a huge fan of jazz.

Paris Martineau [02:14:33]:
What do you like, Jeff?

Jeff Jarvis [02:14:36]:
Joni Mitchell.

Craig Mod [02:14:38]:
I like.

Benito Gonzalez [02:14:38]:
Joni Mitchell is jazz.

Jeff Jarvis [02:14:40]:
She went into jazz. I know she. Well, I didn't like the jazz phase.

Leo Laporte [02:14:43]:
You like the Mingus stuff?

Jeff Jarvis [02:14:44]:
No, there's all kinds.

Leo Laporte [02:14:46]:
I love that jazz.

Benito Gonzalez [02:14:47]:
And there is good jazz. Like all kinds of music.

Jeff Jarvis [02:14:49]:
Jack.

Leo Laporte [02:14:50]:
True. Yeah, true. We're not talking Kenny G here.

Jeff Jarvis [02:14:52]:
Yeah, exactly.

Leo Laporte [02:14:53]:
Yeah. We're talking real jazz. But that's what this sounded like. Like bebop. Almost as great. Mr. Jeff Jarvis, I think that leaves you for your pick.

Jeff Jarvis [02:15:01]:
Leaves me with the bad news. Depends on how you look at it. That computer science majors are seeing the end. Have seen the end of the boom. Yeah, a 15 year boom. In the fall of 2025, enrollment in four year college computer science is down 8.1%, says the Washington Post.

Leo Laporte [02:15:21]:
But you know, I don't take that as a bad thing. I think there will be just as many people studying and working in computers. But it's the problem is the uncertainty. The more there will be more the uncertainty. Right now if you were entering college, you would say, well, yeah, but what am I going to be studying? And is that going to be relevant, you know, four years from now?

Jeff Jarvis [02:15:43]:
So I moderated either. I moderated a panel yesterday in Stony Brook with four academics from SUNY campuses across the state. And I brought this up and they were all aware of it. Everybody knew the 8%. But I think that what you're finding now is, I mean, my son Jake was a CS major. He didn't much like it because it was, it was theoretical math. It wasn't building things.

Leo Laporte [02:16:08]:
Right.

Jeff Jarvis [02:16:09]:
And now people can build things. And I think that without having to be a computer scientist, without having to be a programmer even so you'll have far more people who will be working in computers and doing things with computers.

Leo Laporte [02:16:21]:
Well. And I think one of the things that's encouraging is that the two growth subjects are engineering, which makes sense, right? If you didn't want to do computer science, but you had that kind of brain, you might go in an engineering major and biological and biomedical sciences. I think that's fantastic. And I think that those people are going to be using computers. I don't think there's any doubt about it.

Paris Martineau [02:16:44]:
I think all of those people are going to be.

Craig Mod [02:16:45]:
Yeah.

Leo Laporte [02:16:46]:
So will business, marketing, related health and clinical professions, liberal arts and humanities, psychology, education, visual and performing arts and social sciences. Computers are going to be everywhere. In fact, the thing that's most interesting about this AI revolution is the world runs on software. And it turns out AI is really good at software.

Jeff Jarvis [02:17:05]:
Yeah.

Paris Martineau [02:17:05]:
World on software. And one day soon we're all going to be getting that software from Allbirds.

Leo Laporte [02:17:13]:
The place where that is a little weird.

Paris Martineau [02:17:14]:
Software.

Leo Laporte [02:17:15]:
I prefer to get shoes.

Jeff Jarvis [02:17:16]:
We'll have software in your shoes.

Leo Laporte [02:17:19]:
They are very soft. They're very software. Extremely comfortable.

Jeff Jarvis [02:17:23]:
Yeah, they're softer. You wear them.

Leo Laporte [02:17:26]:
I'm sad because you're about to buy more. They're very expensive.

Jeff Jarvis [02:17:29]:
Buy them now.

Leo Laporte [02:17:30]:
They don't died. No, they died.

Jeff Jarvis [02:17:32]:
Geesefine.

Leo Laporte [02:17:33]:
Geesefine.

Jeff Jarvis [02:17:34]:
Geeesefine.com I helped start a degree in technology, AI and society at Stony Brook and another one in Technology Life Insiders, which was at McLare State.

Leo Laporte [02:17:43]:
Good.

Jeff Jarvis [02:17:43]:
This is for people who want to work in all of this technology, but don't want to become computer scientists and want to work in the humanistic impact of technology.

Leo Laporte [02:17:54]:
G I E S S W I E N and they look just like all birds.

Jeff Jarvis [02:17:58]:
They do. They're great.

Leo Laporte [02:17:59]:
I have a pair and they're 50 off. They're having a clearance sales problem.

Paris Martineau [02:18:04]:
Not sponsored by Allbirds or Geese Vine.

Leo Laporte [02:18:07]:
Geese. Geese Vine.

Jeff Jarvis [02:18:08]:
Geese Vine. Yeah, they're all.

Leo Laporte [02:18:10]:
They're 150 bucks. See, that's the thing. All birds were like that too. They're very, very expensive.

Jeff Jarvis [02:18:14]:
Yeah.

Leo Laporte [02:18:16]:
Much more than a running shoe.

Jeff Jarvis [02:18:19]:
But if you go to the. Hold on.

Leo Laporte [02:18:21]:
Let's just cheap running shoes I buy anyway.

Jeff Jarvis [02:18:25]:
He's fine. If you go to under men, go to the merino. Wool shoes. Those are the ones I like.

Craig Mod [02:18:34]:
Ah.

Leo Laporte [02:18:36]:
They also have wooden shoes. I think that means wood fiber. It doesn't look like they're made out of wood.

Benito Gonzalez [02:18:43]:
No.

Leo Laporte [02:18:45]:
I would hope. Ladies and gentlemen, you've just been watching the strangest amalgam of weirdness ever. We call it intelligent machines.

Jeff Jarvis [02:18:55]:
For your head and your feet and

Paris Martineau [02:18:58]:
your feet and nothing in between.

Leo Laporte [02:19:00]:
Well, unless you have one of these dongles. And then maybe somewhere, maybe you could

Paris Martineau [02:19:06]:
give me one of those. That'd be so.

Leo Laporte [02:19:07]:
Yes.

Paris Martineau [02:19:08]:
All I got is this guy.

Leo Laporte [02:19:10]:
We do this show every Wednesday right after Windows Weekly. 2pm Pacific, 5pm Eastern. Eastern. 2200. I'm sorry, 2100 UTC. You can watch us do it live on our Discord if you're in the club. Or YouTube, Twitch, X.com, facebook, LinkedIn and Kick. You can also download copies of the show after the fact at TWiT TV IM.

Leo Laporte [02:19:31]:
There's an Intelligent Machines YouTube channel for audio and video. Subscribe. That way you get one or the other or both for free. The minute we're done, just look for your favorite podcatcher. Thank you. Paris Martineau. She's on deadline, ladies and gentlemen, at consumer.

Jeff Jarvis [02:19:49]:
And she's an award nominee.

Leo Laporte [02:19:52]:
Yes. Will we know about Deadline by next week?

Paris Martineau [02:19:55]:
No, we'll know about that by March 14th.

Leo Laporte [02:19:59]:
Is it voted on by the people?

Paris Martineau [02:20:01]:
No. Selected by judges.

Leo Laporte [02:20:04]:
Your volume is just absolutely disappeared.

Paris Martineau [02:20:06]:
It was just disappeared. I've done nothing.

Leo Laporte [02:20:08]:
Move that thing in front of your face. Talk right into that. I think you lose energy as the show goes on and we bore you to tears. I think that's what's really happening.

Paris Martineau [02:20:16]:
Much softer.

Leo Laporte [02:20:19]:
It was re. It was almost.

Jeff Jarvis [02:20:20]:
It went down. It went down.

Benito Gonzalez [02:20:22]:
It did go down. I don't know what's going on. It's. It's Very weird.

Craig Mod [02:20:25]:
What is going on?

Jeff Jarvis [02:20:26]:
And now it's not as loud as you were.

Leo Laporte [02:20:29]:
Turn it up.

Paris Martineau [02:20:29]:
Can't be right.

Leo Laporte [02:20:30]:
Turn it up.

Jeff Jarvis [02:20:32]:
How is it all the way up in the box?

Leo Laporte [02:20:35]:
I mean, I can put it up now it's good.

Jeff Jarvis [02:20:37]:
Now it's good.

Paris Martineau [02:20:40]:
What is happening?

Jeff Jarvis [02:20:42]:
The world is trying to silence you and you should not let it happen.

Leo Laporte [02:20:46]:
I blame Gizmo.

Jeff Jarvis [02:20:47]:
Should.

Paris Martineau [02:20:48]:
Anthony, do you mean do I have to have the software open in the back? I've got it on my computer.

Jeff Jarvis [02:20:52]:
I don't.

Benito Gonzalez [02:20:53]:
We can deal with that after the show.

Leo Laporte [02:20:55]:
They did last week too, right? You did this last week? You worked on it. Anyway, thank you.

Jeff Jarvis [02:21:01]:
She's just being held by for detention for me podcasting both at my fake

Paris Martineau [02:21:08]:
job and at my real job, so it's crazy.

Leo Laporte [02:21:10]:
Oh, it's happening on your zoom calls,

Paris Martineau [02:21:12]:
but not with this mic. It's happening with just Google Meet calls in my AirPods, which is odd.

Leo Laporte [02:21:17]:
I don't know.

Jeff Jarvis [02:21:18]:
You're weird, Paris.

Leo Laporte [02:21:20]:
I might have to come right out there and fix it myself.

Jeff Jarvis [02:21:23]:
Well, would you? We've been begging you to come out

Paris Martineau [02:21:24]:
here, have a sandwich, and then I

Leo Laporte [02:21:26]:
can have a sandwich as well now that I'm paying the rent on Henry's apartment. I guess I better. He said, thank you for helping me out. I will have a guest room. Okay, so good. I get to get at least to save on the hotels when I go to New York City.

Jeff Jarvis [02:21:40]:
Doesn't mean there'll be a bed in it.

Leo Laporte [02:21:42]:
No, he said that. He said, I'm turning into a podcast studio. I said, well, at least get a futon for me, will you? A sleeping bag, anything. Jeff Jarvis, professor of emeritus of journalistic innovation at the Craig Newmark Graduate School of Journalism at the Craig Newmark.

Craig Mod [02:22:00]:
Thank you.

Leo Laporte [02:22:01]:
You got more than enough Craig today.

Jeff Jarvis [02:22:03]:
I. I love how every time we have a guest on they. They are hilarious in reaction to that.

Leo Laporte [02:22:09]:
They don't know what the hell's going on.

Jeff Jarvis [02:22:11]:
That's the thing they all want.

Leo Laporte [02:22:12]:
Why did I agree to do this show? What is happening? Don't forget Jeff's new book is available@jeffjarvis.com it is hot type. The incredible story of the Linotype that broke Mark Twain's spirit and his bank account.

Jeff Jarvis [02:22:31]:
Yep.

Leo Laporte [02:22:32]:
We do not have a guest scheduled for next hour. We've got so many great guests in the coming hours.

Paris Martineau [02:22:38]:
What if we just interview each other instead?

Leo Laporte [02:22:41]:
I think people know more than they ever want to know about us, to be honest.

Paris Martineau [02:22:46]:
Do a Q A.

Leo Laporte [02:22:48]:
Okay. Ask Paris anything. Have you ever done An AMA on Reddit.

Paris Martineau [02:22:54]:
I have. For my last story, I did a ama.

Leo Laporte [02:22:58]:
Oh, that's right.

Paris Martineau [02:22:58]:
Yeah, yeah, we're all up in there. Shout out to some of the twit listeners.

Benito Gonzalez [02:23:02]:
I think Jamie B. Has a good suggestion. We want to get Steve, Steve Gibson.

Leo Laporte [02:23:08]:
He's all in on AI these days. You know, there's so many people. I want to get so many people.

Jeff Jarvis [02:23:16]:
I want to get none of the ones that I proposed.

Leo Laporte [02:23:19]:
No, those are all awful. No, I don't know. They were fine. Hey, we'll work on it. We'll work on it. We'll work on it. We'll find somebody. But I can't promote anything right now.

Leo Laporte [02:23:31]:
We do have some good people coming up. Chris Stokel Walker talks about journalism and the change the mayor of San Jose talks about, which is the heart of Silicon Valley. Matt Mahon. He is running for governor in California, but he also will talk about a city government and AI because they're very AI forward. Looks like we're going to talk to the founder of Framework at some point soon. And Bill Gurley, too. We're working on all of that. So lots of guests coming.

Leo Laporte [02:23:56]:
But next, if you have a good idea, should we get Sam Altman? We could get him on.

Jeff Jarvis [02:24:03]:
Oh, Paris, now you're muted. Now you have no sound at all. You've gone.

Leo Laporte [02:24:07]:
Whatever happened, even muted you were ever happened, is happening in space.

Jeff Jarvis [02:24:13]:
I think she's screwing with our heads.

Leo Laporte [02:24:15]:
You have. Is Whoopi Goldberg there? You have a ghost and she starts

Jeff Jarvis [02:24:19]:
talking like that, and then she starts like this.

Paris Martineau [02:24:21]:
I mean, it could just be that. I guess I'm being quiet.

Leo Laporte [02:24:24]:
I think we've exhausted you and you've lost all your mojo.

Jeff Jarvis [02:24:28]:
She's the young one with all the energy and spirit.

Leo Laporte [02:24:30]:
I know. Isn't that funny? Yeah, we have some thoughts. We'll get to figure it out. Anyway, thank you, everybody, for being here. Thanks, Paris. Thank you, Jeff. Always a pleasure seeing you. Come back next week for another Intelligent Machines.

Leo Laporte [02:24:45]:
Bye. Bye, bye, bye, bye, bye.

Paris Martineau [02:24:49]:
I'm not a human being, not into this animal scene. I'm an intelligent machine.