Intelligent Machines 822 transcript

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

0:00:00 - Leo Laporte
It's time for Intelligent Machines. Jeff Jarvis has the week off, but we've got Jacob Ward joining us, along with Paris Martineau. Our guest Daniel Oberhaus. He's the author of the Silicon Shrink. Why you Shouldn't Trust AI With your Most Intimate Thoughts Coming up. Next, podcasts you love.

From people you trust. This is TWIT. This is Intelligent Machines, episode 822, recorded Wednesday, june 4th 2025. The one-man unicorn. It's time for Intelligent Machines, the show where we cover the latest in AI, robotics and smart machines, the ones all around us these days. Very pleased to see Paris Martineau on the other side of the camera. Hi, paris.

0:00:54 - Paris Martineau
Hello.

0:00:55 - Leo Laporte
Love your little kitty cat hairdo. That's cute Thanks.

0:00:58 - Paris Martineau
I believe that's the technical term for it Kitty cat, you got a kitty cat. Yeah, I know I'm with the kids. I'm hip you, you got a kitty cat.

0:01:02 - Leo Laporte
Yeah, I know I'm with the kids, I'm hip, you're risen. Yeah, you're hip, you're risen right now. Yeah, everything I know about this lingo comparison has taught me.

0:01:11 - Paris Martineau
Jeff has the day off, I don't take any responsibility.

0:01:17 - Leo Laporte
It's all your fault. Jeff has the day off so you don't have to hear the Craig Newmark jingle, unless our good friend Jacob Ward has started working in some way for greg newmark. Jacob is uh at the rip currentcom and uh has been on our shows before. We love having him on. I think you were a guest on this show before anything else, you know.

0:01:35 - Jacob Ward
You know what's funny, leo, is that when, uh, you guys reached out, I misread the email and thought it said that leo was not going to be able to be on there and that they needed me to fill in I need you to take over for me, Jacob, and I was like oh my God, I was like I can't believe it and I started thinking, oh, I'm going to do this, and how do I do this?

0:01:52 - Paris Martineau
And then they said, no, no, it's just Jeff and I was like oh, thank God, Just relax.

0:01:59 - Leo Laporte
Sit back. Just let's say sit back, relax.

0:02:02 - Jacob Ward
I don't have to drive the plane?

0:02:03 - Leo Laporte
No, you do not. Jacob is the author of a really interesting book about AI called the Loop how AI is creating a world without choices and how to fight back, which I really, really enjoyed, and I'm always glad to get you back on the show.

But this time you are on the other side because we have a guest today. Very happy to have Daniel Oberhaus. He is the author of a new book. Hello, daniel, welcome, hi, thanks for having me. His new book is called the Silicon Shrink. It came out in February. It talks about the use of AI in psychiatry. But Daniel knows Paris because you work together at wired magazine. Uh, he's also worked at motherboard vice and written for many, many, many magazines. His previous book we won't mention because he wrote it as a child a child at the wee age of 23 to mere 23.

He wrote about extraterrestrial languages, which we will, I think, have to get into, but first let's talk about the silicon shrink. Welcome, it's good to have you here. The premise, as I understand it, is that people should not be using AI for psychotherapy.

0:03:23 - Daniel Oberhaus
Yeah, that's a really great succinct explanation. I mean, the overall premise is that we shouldn't be using AI in psychiatry, writ large, so any sort of mental health application is kind of putting the cart before the horse. Is really the main thesis that we don't understand mental disorders well enough to automate them at scale. But that's exactly what's happening right now, so it's a bit of a warning about why it's happening at scale now.

0:03:49 - Leo Laporte
Oh yeah, man Correct.

0:03:51 - Paris Martineau
Oh yeah.

0:03:52 - Daniel Oberhaus
Oh yeah, like how Well you know the therapeutic applications in terms of using them as a therapist, I would argue is probably one of the more widely adopted ones. So there's tools that have been designed specifically to do that, such as Wobot. So are they apps?

0:04:08 - Leo Laporte
Is that mostly?

0:04:10 - Paris Martineau
or is it actually being used in a clinical situation, or is it kind of like a character AI situation or something else?

0:04:15 - Daniel Oberhaus
Both, yeah. So Wobot would be an example of an app that was designed for therapeutic purposes. What's interesting about Wobot is it won't call itself a therapist anywhere on this website. That's very intentional.

0:04:26 - Leo Laporte
Um, it's a, it's a wellness app, but it's it's functionally, it's a cognitive, behavioral, uh therapist what they're saying, though, is oh, as a lack of therapists, or it takes a long time to get to see a mental health professional, so the implication is that's what you're gonna see, right, but it's not it, it's an AI.

They've explicitly said it's not not meant to be a substitute for a therapist, and so when I, when I say this to people, is most common reaction is well, not everyone can afford to go to psychotherapy, let alone psychiatry, which is even more expensive. Isn't it better to have an AI shrink than nothing at all? Even more, expensive.

0:05:06 - Daniel Oberhaus
Uh, isn't it better to have an AI shrink than nothing at all? Yeah, I mean, that's the um, that's the kind of the profound ethical question, right? Um, so this is the primary argument I think made in favor of it, and there I'll start by steel manning it, which is there is an acute shortage of mental health professionals. I think something on the order of about two thirds of us counties don't have a practicing psychiatrist within them, which is wild, um. But even if you do have access to them, there's, uh, you know, a lot of stigma that is wrapped up in seeking, um, mental health uh services. Uh, there's just it's challenging to get access. I have lots of friends in new york city where who they can afford them. There's plenty of therapists in new york city and they still can't get access yes, and it is expensive.

0:05:42 - Leo Laporte
My therapist is $235 an hour.

0:05:45 - Daniel Oberhaus
You went for the bargain one.

0:05:47 - Paris Martineau
Yeah, I was going to say that's cheaper than my therapist.

0:05:49 - Jacob Ward
You got grandfathered in.

0:05:51 - Paris Martineau
You're kidding. Mine on a payment reduction is $250.

0:05:57 - Leo Laporte
You know, oh my God. Well, it is Petaluma, it's not New York City. So maybe I mean. Yeah, you got those small town prices do you, do you know offhand, daniel, what the what the you know mean cost for psychotherapy is?

0:06:13 - Daniel Oberhaus
I? I don't offhand, um, although I've never met anyone who has told me it's like, wow, that's, that was so affordable.

0:06:19 - Leo Laporte
No, it's not affordable, you know, and it's. And of course, a lot of health plans either don't offer it, or I'm on a Kaiser and they offer psychotherapy or even psychiatry, but they only give you 10 sessions. After that you're on your own Wait really.

0:06:36 - Benito Gonzalez
That's crazy.

0:06:36 - Jacob Ward
There's a limit to it.

0:06:37 - Leo Laporte
First 10 are free.

0:06:46 - Jacob Ward
Can I jump in, daniel? I just want to say first of all, just compliments on this concept. The concept of this book is so smart because I feel like I wrote this early worried book about AI in general and I touched a little bit on therapy. But that you went deep on this, I think, is such a smart thing.

And this notion you talk about, the idea that, well, first, this thing of should we be okay with trying to simulate a service we no longer have is sort of the industry argument, right, that like, well, nobody is going to do this job, so we need to come in and bring these folks in. And yet I'm curious, what do you think like, does it? There's the ethical question, but then there's also this question of like if we start to normalize simulated therapists and using large language models for therapists and so forth, will anyone ever be able to be a human therapist again? Or is that sort of like? Do you worry about a downward spiral in which we no longer have human therapists, sort of at all, at the end of this spiral?

0:07:45 - Daniel Oberhaus
in which we no longer have human therapists sort of at all. At the end of this, I'm not as worried about that. I appreciate you saying this was a good idea. The kind of an interesting backstory about this book is I started pitching this around or my agent did. Rather, I should give Robert credit there in 2019. So prior to ChatGPT becoming a thing, and we got in front of all the big four publishers and they're like this is a really interesting subject, but you know kind of niche. It's maybe good for academics, like save it for the university campuses. No, it's like it doesn't have broad appeal. No one cares about this. Then chat to PT happened and now actually one of the I had the same timing.

I had the same timing.

0:08:16 - Jacob Ward
We didn't even put AI on the cover of the book, cause we thought it would be too nerdy.

0:08:19 - Daniel Oberhaus
Yeah, and now it's like I literally just this, like a week and a half ago, I was at my barber and shout out Mark, but he was. He was like hey, have you heard of this thing called ChatGPT? And I thought he was joking at first. I was like Mark, you're like a 29 year old guy and you're telling me you just found out about ChatGPT in May of 2025. And he said yes, and so I was like okay, well, now I'm dying to know like, what are you using it for? And he's like oh, you know, at first just like asking a question, it's kind of like messing around with it, having fun. But he's like. Now he's like I use it almost daily for therapy and I was like there's no way really Like. And he's like yeah, he's like. So I was like tell me what you like about it and you know the things. You yeah, uh, you know it's, it's almost in a way, uh, which is both its advantage and disadvantage.

0:09:08 - Leo Laporte
Um, in impersonal, yeah right, I mean my therapist I have a relationship with, and I think it's important to have a relationship with your therapist, which obviously you can't have with an ai. But especially for people just starting out, it might be easier to talk to an AI.

0:09:30 - Daniel Oberhaus
Yeah, I mean I, um, I was just at a book festival this last weekend and I had a woman come up to me afterwards who, um was like you know, I think what you're saying is important, but she's like I. I just had my husband, who has rejected um therapy for a long time, and she's like he needs it.

That's like okay, that's really that's what every wife says, so that's okay yeah, but yeah, she, she's been trying to get him to uh, to kind of at least consider it, and so she used chat gpt as kind of a an easy way in and he, he really adopted to it to the point where now he's actually open to going to a human therapist got him started talking exactly safer, safer way yes, and there's actually a point out.

0:10:03 - Leo Laporte
I understand there could be downsides. You know, chat gpt could say you know, oh well, really, you, you, your life sucks, you shouldn't stick around. But I could see a human therapist and I know many cases of I mean. First of all, there's no requirements to put up a shingle saying I'm a counselor, uh, so they're, you know. You know, you might ask as a potential patient, what's your training, what's your degree? But you don't have to have a degree in it, you don't have to have training in it. And there are plenty of counselors who do have training, who might not be very good, who might even be dangerous, right?

0:10:41 - Daniel Oberhaus
you forget that there's a bell curve of doctors as well. Right, right.

0:10:44 - Leo Laporte
Same thing with physicians.

0:10:45 - Daniel Oberhaus
There is such a thing as the worst doctor in the world, and it might be yours.

0:10:53 - Jacob Ward
The thing is, there was a study that just came out that showed, basically, these makers of AI were trying to figure out a benchmark for the degree to which AI tries to suck up to you and make you-.

Sycophantic yeah sycophantic AI right, and they were trying to figure out the degree to which these models were doing it, because there was a lot of anecdotal evidence that these models will just try to flatter you all the time and reinforce your ideas. And da, da, da, and so they used like thousands and thousands of um am I the asshole subreddit posts, you know so the subreddit. Am I the asshole right which tells these incredibly embarrassing stories? Did I do something terrible and and?

0:11:35 - Leo Laporte
they were doing fake posts, though, right no, that was a different subreddit they pulled in. They pulled in like that was right, talk me out of it, or whatever.

0:11:47 - Jacob Ward
Right, and and what it? And what they found was that these models are incredibly sycophantic and that there is a way. And not only are they very sycophantic, they have no idea how you would fix this problem. And so, to me, one of the things is like no matter how bad the human therapist is and there's plenty of bad human therapists they're at least trained not to just tell you that your ideas are a good idea. They're trained in some way to try and challenge your thinking on this stuff.

0:12:11 - Paris Martineau
You know not all of them or at least they have, or at least the ones that are trained to have a perspective of some sort.

They're coming at the conversation with a perspective on the world that is going to shape how they respond to you and how they interpret your responses, versus a chatbot or a large language model powered through dynamic, is going to be inherently kind of reflexive and sycophantic. I mean, what did you find in your research on this, like about the nature of this kind of relationship? And I've assumed there are both, like there's a promise to it, but then also peril related to this.

0:12:55 - Daniel Oberhaus
Yes, so the research is actually pretty interesting on this and, to kind of go back to an earlier point, is DARPA. In 2014, they had partnered with the University of Southern California to do a study on exactly this. They brought a bunch of veterans in and ran a study on exactly this. They they brought a bunch of uh veterans in and ran a study where there was a female presenting a therapist avatar. One uh study group was told that there was a therapist on the other side and they were basically just interacting through this digital representation of them. And the other one was told uh, there's no person on the other side, you're just talking to a robot.

Um, and you know, they were doing a lot of like facial recognition stuff, which is a whole other can of worms in terms of that not actually being a thing.

But what they found actually was that the veterans who interacted and knew they were interacting only with an AI actually had better like, they would disclose more information. They actually got more value out of the therapeutic sessions than the veterans that thought they were interacting with a human through an AI avatar, which there's, like some you know, limitations to that study model, but outside of like and there's thought. This isn't the only study that has done this, but, going back to the very beginnings of the field, there was a researcher at MIT named Joseph Weisenbaum who, in the 1960s a lot of people are familiar with this has created this chatbot called Eliza yes, and was horrified that people would interact with it as if it were real and it had this shtick of being a therapist. But people would ask him to leave the room because they would actually want to talk with it and for him it was just a way to study human-computer interaction.

0:14:15 - Leo Laporte
Although he did want it to be kind of a Rogerian therapist right.

0:14:19 - Jacob Ward
But he was. I mean to Daniel's point. He quit the field. He was so horrified that he walked away.

0:14:25 - Paris Martineau
He point he quit the field. He was so horrified he was pissed away. He was pissed because the thing is, it wasn't even that like advanced, like, I believe, the main thing is terrible.

0:14:30 - Leo Laporte
It's like a total reflect whatever you said back to you yeah, that's basically what it did, and that's why he said it's like a rogerian therapist who does, in fact, reflect a lot, right, yeah, right, what do you think about that? What does that tell you? Yeah?

0:14:44 - Daniel Oberhaus
I. So this kind of comes back to what we were talking about at the very beginning, and so there's apps that and we should also mention that. You know, the AI therapist is what most people think about when they think about AI in this context. But, like, this is a very broad field. It's just kind of like the tip of that iceberg is this application, but within this application, there's apps like Wobot that were designed with the intention of being used for some sort of therapeutic purpose, even if they don't call them therapists. But then there's this other category of things, such as ChatGPT, that were never intended for therapy, that people naturally use them for, and that is truly as old as like it started basically with Joseph Weizenbaum, but people have been using Siri for this way. People have been using ChatGPT for this way. There is something so natural about doing this. Like my barber, first thing, what does he do? Like within an hour of having access to this tool? Therapist, I don't know what that is. So people have parasocial relationships with celebrities online. They have parasocial relationships with AI celebrities online, from what I gather now. So there is something about that that I find just very, very interesting.

The book doesn't go into as much of the kind of the why behind that relationship building, but it does seem to just come so incredibly naturally. But then there's challenges with that right. So, like a very recent news item, is the character AI thing with a young man taking his life, and I believe the American Psychological Association is now bringing this to the FTC to investigate. You know basically who's at fault for that, because you had mentioned earlier, it's like you don't necessarily need a license to practice as a counselor. You definitely do as a psychiatrist, but even if I go to my therapist, the data I give them is subject to HIPAA restrictions in terms of how they can use that. Most tools that you have access to in the context of AI and psychiatry they are not hipaa protected your data is, robot says it is um, well, do they say that?

uh, it's comparable to hipaa, because, like they're not regulated, they don't actually have to, they're just. You know right, we're going to abide by that because we're nice like they're not regulated hipaa compliant on their front page right but?

0:16:45 - Jacob Ward
but like he says, that's right, that's marketing more. I mean, they may very well be hippocampal, but they don't have to be right. Yeah, so they could not be checked?

0:16:54 - Leo Laporte
let's not be hippocampal and and they could really reveal everything to an advertiser or whatever.

0:16:59 - Paris Martineau
I mean it's interesting this kind of touches on you. I feel like discuss in the book this concept of like a psychiatric surveillance economy that these tools are enabling. Can you kind of elaborate on what that entails and kind of what impact that could have on like individuals' privacy or autonomy as they kind of use these tools?

0:17:17 - Daniel Oberhaus
Sure, yeah. And you know I'm a big fan of Shoshana Zuboff who, you know, kind of coined this idea of surveillance capitalism.

Yeah, who coined this idea of surveillance capitalism? The argument was that this is a new front in that. The core idea here is that when some people hear about AI in psychiatry, their first thought is, well, that sounds bad, but whatever, I'm not in therapy, it's not going to affect me. It sucks for the patients. Not, it's not something that's going to affect me in my day-to-day life. The. The problem with this is that the way these tools um are effective is, uh, essentially through dragnet um uh data ingestion, and the reason for that is because a we don't know what mental disorders are like. Psychiatry is the only medical profession that has not a single organic, like undisputed biomarker, uh, for any of the diseases there's the dsm correct.

0:18:09 - Leo Laporte
I mean they have a diagnostic undisputed is the kind of key.

0:18:12 - Paris Martineau
Well, it's accepted.

0:18:14 - Leo Laporte
I don't know if it's undisputed. Nothing in the world is undisputed.

0:18:16 - Paris Martineau
I mean, I feel like the dsm even among itself changes dramatically, I think there's some real problems with the dsm, I completely agree with you.

0:18:25 - Leo Laporte
But it is an accepted standard for diagnostic.

0:18:29 - Daniel Oberhaus
Well, I would, actually I would challenge that. So the the National Institutes of Mental Health under Thomas Insell in 2013, they said they weren't going to give grant money anymore to any study that was purely relying on DSM criteria because of how inaccurate um it was for uh, for diagnostics. So it is, it is the standard, but the the way that those diagnoses are a good standard? Well, it's not. It's it's generally not. We should explain.

0:18:54 - Leo Laporte
This is from the american psychiatric association. It's the diagnostic and statistical manual of mental disorders. If you go to a therapist or a psychiatrist on insurance, they're going to categorize you using the dsm. Uh, but for a long time dsm considered homosexuality, for instance, a mental disorder.

0:19:12 - Jacob Ward
There've been all sorts of problems with the dsm and it takes forever for things like gambling to get in as an official piece of you know, as an accepted addiction, that kind of thing, but in a way that underscores my initial point, which is humans are pretty flawed in this too.

0:19:29 - Leo Laporte
Right, this isn't created by ai, this is created by humans. Um, I, I feel like you could train an llm just to, for argument's sake, to not say things like, well, you're an idiot or you know you should take, try methamphetamine, but instead to say just encouraging listening things and that it would be of therapeutic, because part of the at least freud thought, part of the value of talk therapy is simply talking right, not necessarily the therapist's input, but the actual expression of what's going on with you is sufficient. Couldn't an LLM be? I think Google has an LLM that's supposedly trained to do that. Is that not acceptable?

0:20:13 - Jacob Ward
I think that might be a very valuable tool it's funny, you know, you guys, I was just literally just having a conversation for my podcast moments ago with a psychiatrist who's now a british, uh, politician. He's a really interesting guy and he's he was the architect of the good friday accords that ended the troubles in northern ireland fascinating guy named lord john alderdice and he he was saying that the key for in both individual psychiatry and in trying to get, you know, protestants and and Catholics to sit at a table and agree to not killing each other has to do with being able to express yourself absolutely but also being very effectively heard. That once you are that, you have to be heard in a way that feels real to you in order for it to have the true sort of, you know, catalytic quality. And for me that's a, that's a the thing that that it feels to me like ai doesn't do right. Is this feel? I mean, maybe, maybe it does, I don't know we, you know, we know it.

0:21:15 - Leo Laporte
We know it doesn't right because we know how it works.

0:21:19 - Jacob Ward
But if it feels like it does, it might maybe it feels that you know, until I said it I hadn't even really thought of that maybe, maybe for a certain, especially to a certain generation of people, like maybe that's a better, maybe you feel heard in a way that you don't with humans, if you're, if you've sort of grown up in that generation or the rest of it treatment uh, she's, she's doing very well.

0:21:36 - Leo Laporte
She talks to her chat bot. She has a therapist and a psychiatrist but she talks to her chat bot and really gets a lot out of it. If I thought it was dangerous for her and it could be because you know she's bipolar um, I would say you can't do that, but I think she gets a lot out of it you.

0:22:02 - Jacob Ward
You, daniel, just ask you a question here.

You have this really cool concept in the book that I really appreciate, which is this idea that somehow we're going to wind up in a kind of digital asylum situation.

And you use that term, asylum, as this sort of symbol of where things could go wrong. And I just think it's so interesting, this notion because, like I bump into this all the time as a tech writer and I know I know you think about this too that this thing of you know, if we, if we make a better, you know a more efficient, more cost-effective digital solution to a thing, and you wind up kind of giving up on the harder human version of it because you've achieved marginal utility by making something digital, are you giving up in a way that's bad? Is sort of my question. And in this case, I just want to hear you talk a little bit about this notion of creating a kind of asylum, which for me, as I understand it, is sort of this idea that we're somehow going to warehouse people with these digital tools rather than actually making sure we address them in effective ways. Am I understanding that concept right?

0:23:15 - Daniel Oberhaus
Yeah, and it comes back to what Paris was speaking to earlier about the psychiatric surveillance economy, and so kind of close the loop there. The challenge with a lot of these tools right now is we don't have a good understanding of what data is going to, or actually I'll unpack this a little bit more. So the reason why a lot of psychiatrists and people in the mental health profession find the idea of using AI so attractive is because if you go to a mental health provider assuming that you're not in an inpatient facility you're likely going to see them, maybe for one hour a week, maybe you can talk to them on the phone or like text them, whatever, but it's relatively little contact. So the therapist or the mental health professional we'll just use psychiatrists as shorthand has like a very limited data set. And when you come to the you know, an outpatient meeting, for instance, you're relying on the patient's recall how was your, how was your last week been, how have you been feeling, et cetera. And so there's, you know, surveys you can send out for momentary, like ecological assessments for uh, for instance, um, to kind of get more in situ uh data, but the vast majority of the time that you are awake, uh, you have no data on that patient how they're feeling, their behaviors, their thought patterns, et cetera.

What's nice about AI is like we all walk around with a pretty sophisticated computer in our pocket right now and we're just bleeding off all of this uh digital exhaust all the time, and so the thinking around a lot of the kind of push for uh AI and psychiatry is that we can use this data that we're generating, because most of like, most people are spending a significant amount of time, uh, every single day in front of a digital network device and this data, even if it's not, what am I writing into the computer? It might just be like my typing speed or my scrolling speed, and these things that are almost happening at a subconscious level might hold clues to my mental state and my mental health. Um, it's a really appealing idea. There's very little data to back it up, that it works and, in fact, the uh NIMH uh former chief who was there for about 10 years his name was Thomas Insel. He left to go work at Google to pursue this idea and then he launched a company called Mindstrong to pursue this idea. It's the most well-capitalized startup in history, raised a hundred million dollars and 2023, it completely shut down and they never said why. Presumably it's not because of how well it worked, and so there was very little data this entire time about this, and so the challenge right now is like that is a very attractive idea. If that works, I'm all for it, but we don't know what data is going to have like the best read in terms of actually measuring my mental health so I can correlate it with, perhaps, a diagnosis. How do I know that I'm about to enter into a crisis? So you basically have to be monitoring everything I'm doing all the time around the clock to make sure that, a I don't have a crisis while you're not watching, and then B well, maybe it turns out that my typing speed was like the X factor that you can map to the like it correlates very highly with major depression, but we don't know that. So right now we're kind of just hoovering up everything and seeing, like, what matches best and trying to like fit the box around this person. So that's kind of the surveillance economy aspect of it.

And then to the asylum piece. That's kind of an interesting historical analogy in the sense of asylums like prior to roughly the 19th century, the way that the mentally ill, what they would have called people with madness, were basically just warehouse. They were put in prisons, they were cared for, they were rich by kind of like they just hire a benefactor to like put them in an outhouse somewhere where they were kept away from polite society. Um then in the 19th century, around the enlightenment, um, uh, a bunch of uh, I think, people with very good intentions that, well, hey, maybe we can use the asylum for therapeutic purposes and we can actually use it as a place for rehabilitation. Um, people in prisons were thinking about the same thing. And it's just like you know it's it shouldn't have a custodial function, it should have a, a healing function. That went really, really well for a few decades, and then people just kept coming and the asylums essentially became overwhelmed, and so the number of patients in the asylum in the U? S hit its peak in the 1930s I believe you'll have to check me on that, but like pretty late into the game in the 20th century, before the population started declining. And then there was, of course, the turnout of all the patients around the Kennedy administration, with disastrous results that we're still dealing with today. But the asylum basically became a victim of its own good intentions, in the sense of it just became overwhelmed and it was no longer able to fulfill its therapeutic purpose and essentially reverted back to a custodial function. And so I think something similar is happening now.

It's just much harder to see because it's not physical, but these algorithms are beginning to be run on any sort of institutional computer, right. So, like you're seeing suicide detection algorithms in K through 12, you're seeing them in colleges, you're seeing them in government institutions, you're seeing them in offices, because happy workers, it turns out, are actually more efficient. So you can say it's for their mental health, but really it's because you'll get better products. But, like, judgments aside, these things are being implemented on our computers and you can say well, maybe I don't work in an institution Like this doesn't affect me. If you use Facebook, it does, because they have a suicide algorithm monitoring you too. Not if you're in the EU, because they banned it due to to privacy reasons, but if you're in the us, um, they do. So it, it is there, it's already here, um, and it's only they're doing that to try to mitigate correct teenage suicide, though, I mean, is it ineffective?

we don't know, they don't, they don't publish the data.

0:28:16 - Jacob Ward
Yeah, well, I mean, and the thing that we all worry about, right is this, is that they, they might like. My worry is that, like, because the market incentive is there to adopt it, because it's just so much easier to deploy a million digital therapists than it is to train and deploy human therapists, that we're going to sort of do that and, by the logic of Silicon Valley, just kind of like, you know, iterate as we go before we even know whether that's effective.

0:28:44 - Leo Laporte
Yeah, moving fast and breaking things isn't so hot if you're breaking people. Yeah, that's the thing I worry about. But what is the cause? Do we know what the cause of the therapist shortage is? Is it something we can fix?

0:28:58 - Daniel Oberhaus
That's a great question. There's several contributing factors. One is just it's a pretty brutal job. When you think about what a therapist is doing all day, you're basically a receptacle for people's problems, which could vary from very, you know, day-to-day kind of just the the, the living or the worried Well, I think it's what Freud called them to people who you know. Maybe they're harboring suicidal ideation, maybe they're the victims of abuse, maybe they're abusing others. So you're dealing with just heavy shit professionally, so not very attractive from that. So there's a huge rate of clinician burnout. Of course you could say the same of police officers.

0:29:31 - Leo Laporte
We don't have a lack of police officers.

0:29:34 - Daniel Oberhaus
I mean, I actually don't know.

0:29:35 - Paris Martineau
Payment. I think is also a big problem, I would assume, is police officers get salary and overtime. Therapists are typically hourly rate. They have to hustle, try and figure out how to make it worth it with insurance. They have to deal with reimbursement. They have to deal with the fact that clients can't really pay.

0:29:54 - Leo Laporte
I know a lot of therapists and a lot of very happy therapists who are working very well. My ex-wife is a therapist. It takes a certain kind of person. My guess is that there just aren't that many people who are would be good at it or would want to do it. Uh, even if you funded it, even if you promoted it, even if you had college programs and everybody said this is a great job, there just aren't that many people who have the skills to do it or the desire to do it.

0:30:23 - Jacob Ward
Mike would be my guess guess. I mean here in California they had to create a whole other category called a marriage and family therapist. Yeah, my wife was ex-wife was an MFT yeah. And an MFT right. That's a specific California thing. You can't practice in other countries, in other States. They created it because of this huge shortage. I mean, I'm looking at the American counseling association site. They're saying lack of funding, poor reimbursement rates, low retention, huge need.

0:30:48 - Leo Laporte
He had to get a master's degree. She had to pass an exam. She had have many, many hours in training, with an accomplished psychotherapist working with her, to get the MFT. Actually she had an MFCC, or initially so it is. It is not the kind of thing you hang on a shingle and the next day you're you're taking patients yeah, I certainly don't mean to suggest that she's not legit.

0:31:11 - Jacob Ward
I'm just saying that it's a. It's a reflection of how desperate california was yeah, but it's still a slight, a slightly easier way of getting as many people wasn't that easy.

0:31:22 - Leo Laporte
Let me just I just want to say it was a good quality qualifying uh process, I would say, and a lot of people probably wouldn't get through it. It's also expensive, by the way, uh, because you have to get a master's yeah, it's funny, I'm actually in the.

0:31:35 - Jacob Ward
I'm in a because I'm in this funny moment where I'm no longer a full-time journalist. I've been thinking like what do I do with myself? And I started looking at being a therapist.

0:31:46 - Leo Laporte
I saw you're a fellow at Stanford in behavioral sciences.

0:31:50 - Jacob Ward
I did a fellowship there as a journalist and that got me thinking about it. But yeah, thinking about like becoming a therapist turns out it's such a hard road. It didn't you know. My wife looks at the numbers. She's a much more practical person than I do and she was like these numbers don't make any sense for our.

0:32:06 - Leo Laporte
I think you'd have to have a calling I honestly to be a good one anyway, and the good therapists I know they have a calling. It's like you know, it's like being a priest. It's like you have to really have that nature and want to be of service and it is a sacrifice. And want to be of service and it is a sacrifice, I'm sure it's a big. I can't imagine doing it. It takes a lot of training to be a good one. So the reason I ask is I suspect we're never going to have enough. Yeah, so maybe there is some way to use AI to. Is there, daniel? Is there any safe way to do this?

0:32:50 - Daniel Oberhaus
Yeah, and so, um, there's like a few, a few points that I really want to make sure I clarify. One is that, uh, this is this, isn't a book, and it's not my opinion at all that people just shouldn't use them full stop, right. So like, for instance, uh, your family member, I know lots of uh people who do this, uh, by Barbara included. Um, and I'm not, I'm not here to tell them, don't? Um, in the same way that I wouldn't tell someone, hey, don't take the pills that your psychiatrist gave you, right. But what I will say is like, hey, we, we actually do have a massive uh uh over-diagnosis problem. That has led to a massive uptake in uh, uh psychiatric drug use. Maybe that was out of best interest, maybe that was out of industry interest, it doesn't really matter. The fact is there are a lot of people on psychiatric drugs now, but what's interesting about that is the results have gotten worse over the past 20 years. This is the only medical specialty, once again, where, as it progresses, the outcomes are getting worse for patients. That's shocking. We spend billions and billions of dollars doing this.

I don't think it has anything to do with the intentions or the capabilities of therapists and people who work in this. I have nothing but very, very deep respect for them. So I just wanted to say, like I don't like people ask me this all the time. Now it's like I don't think that it's, you just shouldn't use it. But I do think you should like read, you know what's on the tin. What am I going to get from this? What are the risks, what are its limitations? I think that's very important to know and it's often very lost, Like we were just looking at Wobot's website and it'd be very hard to tell that that wasn't subject, like it wasn't regulated as a like a HIPAA compliant organization. It is HIPAA compliant, that is true, but there's this kind of unspoken thing about but like we to, and so that's the kind of thing where this was more meant to be kind of like a shot over the bow. I'd be like this is happening. This is real. I'm the last person who's going to say that there aren't people getting benefits from this. There's definitely people getting benefits from this and I think that's that's fantastic. I also think you know to your point, leo, you know, can, can we use these things effectively? My answer is like a like an optimistic kind of yes.

I think what's so interesting about writing this book is this is an issue that has become very topical because of chat, tpt, but it's about 70 years old and it goes right back to the beginning of artificial intelligence as a field, and, in fact, carl Sagan, in the 1970s wrote an article for natural science where he predicted a future where there would be almost telephone booth therapists to deal with a therapist shortage in his day 50 years ago. So it's definitely not going away. The tools have gotten better and I think what's really interesting about everyone kind of indexes on AI therapy. The book covers a lot of different technologies and there are certainly a lot of other ones that I think are much more nefarious different technologies, and there are certainly a lot of other ones that I think are much more nefarious. But as far as AI AI therapy goes, um, most of the AI therapists are implemented as a cognitive behavioral therapists, and what's interesting about CBT is when you really zoom out about what it's trying to do. Um, a lot of people, I think, look at therapists as like almost gurus, but like that's not what they are. They're actually teaching you how to help yourself, right, like they are giving you the tools to improve your, the way you think, to improve your behaviors in ways that are like, less maladaptive and, um, increase your overall wellbeing. That's what a good CBT therapist should do, um, and, to the extent that that's what they're there for and what they're trying to achieve, um, the programming for CBT is actually. It lends itself very well to being implemented in an algorithm in the same way that we were talking about Rogerian's therapists. That was a little bit more hacky what Joseph Weisenbaum was doing, but as a way of delivering care, cbt is actually very primed for application and LLMs. We're pretty early days here, but I think that's phenomenal if that works out and we can get people care and it's proven to be effective.

My question right now is why can't anyone show me data saying that this is at least as good as a human therapist? Because the pushback I usually get is it's better than nothing. But then I say can you prove that? And the answer is no, no, you actually can't. There's no data showing that.

Because, like, for instance, I just talking to the new york times a few weeks ago about k through 12 suicide monitoring platforms, right, and you're like how could that possibly be something that anyone would ever be like, hey, wait a second, we may not want that.

It's like it almost sounds like you're advocating for children killing themselves and it's like that's not what I'm advocating for, uh, at all. Um, but there's this tendency to say we must do this because it might save one life, and that is a philosophical and a moral argument you can make, and lots of people do. But there's also lots of people that say, hey, maybe we shouldn't all just be forced to like participate in these systems just because one person thinks it might help a kid, because we actually don't know if that would have happened regardless. There's just so little data and the companies that are building these aren't forthcoming with them, at forthcoming with it at all. I don't know if that's because the data suggests that maybe it isn't as effective as they want you to believe, or if it's just not, if it just doesn't exist. Yet my experience with tech companies is, when they have data that shows their thing works, they want to tell you about that.

0:37:40 - Leo Laporte
They will let you know.

0:37:41 - Paris Martineau
Yes, there will be seven press releases and a web conference.

0:37:45 - Daniel Oberhaus
Yes, yeah, there will be seven press releases and a web conference. Yes, so it's very curious that we're missing that. And you know, if you go on these sites and look at them, a lot of the sites will have data and it'll say you know it's. They'll make some very large claims, but if you actually click on the, on the peer reviewed research that they are claiming backs it up, what you'll find is um, you know, it had these massive mental health benefits for college students with depression. I was like, oh, that's interesting, maybe we found one that worked. And then I went in to look at the study and what they were actually showing was that this chatbot was better than a government pamphlet. In a cohort of about 50 college students over two weeks, that's not nothing. But, like I don't know, it's not really what you're saying on the website either. Right, better than a pamphlet.

0:38:24 - Paris Martineau
I was like okay, well, tell that to investors.

0:38:27 - Daniel Oberhaus
Yeah, I'm willing to. I'm willing to believe that actually, that it is better than a pamphlet. I'm less willing to believe that you are just across the board, like lowering depression rates by 50%. It was like something very grandiose like that, but that that was the case, but only with that very important asterisk, and so there's a lot of that in this field. Um, and like.

0:38:46 - Jacob Ward
I mean, that's the thing that is so disturbing to me about this world, that you're covering this book right? Is that, like in the united states particularly, this kind of thing is a business, whereas in other countries it's a civic service, it's a national service. Your therapist is not a, is not you know you're. You go to your national health service in the UK or you know other places for your behavioral therapy or whatever else you're going to need. It's only here where we've, where everything's privatized, that we don't, that we would even consider that we would need a uh, you know, a digital replacement for this stuff. And then we're all just kind of like hostage to the marketing, to how truthful the marketing is. It's so cynical.

0:39:33 - Daniel Oberhaus
It's shocking to me. So like to take Wobot, for instance. I feel like I pick on them. I actually have no beef with Wobot, it's just like probably the most well-known one. But they are currently doing a FDA trial and what's surprising to me about that is most things that have to go through an FDA trial, you cannot access them prior to it.

0:39:51 - Leo Laporte
So is it like is this, is this a?

0:39:55 - Daniel Oberhaus
medical technology Cause, like, what's going through the trial is the exact same thing. I can go download on Google play, so like, like, what, like what, what changed, so like. I basically just have access to an unregulated medical technology. And the reason I do is because you're not saying it's medical, but you are like you aspire to that, and that's that's. That's shocking to me that you can, that you can do that. So I don't know. The UK is actually an interesting example, because the NHS is also increasingly doing this. They have a pretty robust tech sector that's building stuff here for the same reasons.

0:40:31 - Jacob Ward
Right, it's therapist shortages, um, but I think the number two and number three, uh chat therapy bots are actually uk-based.

0:40:34 - Leo Laporte
Interesting. That is interesting. It's. I understand the the difficulty here. You should, people, you should absolutely read uh daniel's book because, uh, it goes into a lot more depth uh than than we could possibly uh go into in half an hour. But uh, it's, by the way, called the silicon shrink, how artificial intelligence made the world an asylum. Uh, it came out last month, a couple months ago, and is available on amazon, as you can see, uh, and kindle or hardcover, and at your favorite bookstore, which you should absolutely patronize. But I also I feel like absolutely patronize. But I also I feel like, you know, you don't want to experiment with people's psyche. That's a very dangerous thing to experiment with. Of course, we also experiment with full self-driving in vehicles, also dangerous.

Um, ai is involved in a lot of things. We're going to talk about some very dangerous, but at the same time, there is a need. We know there's a need. Uh, we know that in some cases it's useful. Um, I don't think there's a clear answer on this one. You're, I know you, you don't say don't do it um, but I think it's important to think about the consequences of doing it. I just don't know if there's a clear answer, like, is this a bad way? I feel like we need to look at it and think about it and move, and move forward judiciously. I feel like that's an.

0:41:53 - Paris Martineau
A possible answer is like that. It needs to be um I wouldn't throw it out.

0:41:58 - Leo Laporte
Actually thought about would you throw it out?

0:42:00 - Paris Martineau
no, I mean, I think that most things exist somewhere in gray scale, rather than all black or all white.

0:42:08 - Jacob Ward
Well, one thing I get out of this book is that we have come to at least expect in the United States that when someone offers you a therapy it has been tested and validated in some form, that if you go to a licensed medical professional, they've been through some sort of training and that your data will be kept private. What I think, Daniel, you've made such a good case on in this book is like those things don't apply to this sector and we're, you know, as I'm sure we'll discuss later in the hour or later in the show. You know, the budget bill that Congress is looking to pass right now would forbid states from regulating AI in any way for a decade. So the regulation of this kind of thing wouldn't even really be possible, if you know, under state law down the line, if that bill goes through. So like we're living in this world where it's totally unregulated, and to me, if there's nothing else here, it suggests to me that some standards should be imposed on what is currently a totally wild west.

0:43:05 - Daniel Oberhaus
Yeah, and I think to maybe to close the loop on this, like I'm I'm very strongly in agreement with what you had said, leo, because, like I think, um, the book is largely focused on what is the risk of if we, if we do this and we like we don't put a check in, right? But I think the reason why this is a problem at all is because there's a lot of well-meaning people who say what is the risk if we don't Like, if we must do something? The problem is that's like the famous politician's fallacy, right, like we must do something, this is something. Therefore, we must do this. And it's like.

That's where I'm, like this is one tool in the toolkit, and all I'm saying is like let's withhold judgment right now. Like these exist, they're all around you already, let's take a beat. And all I'm asking is like, please just show me the data. Like, does this work as advertised? And the answer is like it either doesn't exist or they're there. No one is showing it to you, and so it's hard though this is hard data to produce.

0:43:58 - Leo Laporte
I mean, it's not, oh for sure. Yeah, you know, I think we have a little bit of a crisis going. We have two school nurses for 1500 students in, uh, my little town. That's not adequate, uh, and we celebrate it because we've got two. So we have a cry, we have a problem.

I would you know, it'd be nice to say everybody, go be a therapist to your friends and neighbors, go help, go talk to them, so they're not talking to robots. Um, there are lots of things we could do. We should probably be doing them, um, and I'm not sure ai is the solution, and and you aren't saying it, but I will say it the sad thing is, ai has also become a gold rush and there are a lot of people out there trying to make a lot of money on any possible use of AI, because it's easy to raise money for it, and which means there will be a lot of bad actors involved in this space and every other, and in this space, the hazards are great. So that is something else to be very aware of, daniel. I want to thank you so much for your time, daniel Oberhaus. The book is the Silicon Shrink and well worth reading. I thank you for your time and I look forward to talking to you again yeah, thanks.

0:45:07 - Daniel Oberhaus
So much appreciate the interest and the time. Yeah, have a great one take care.

0:45:11 - Leo Laporte
See you guys, we're gonna take a little break. We have more to come on intelligent machines our show today brought to you by big id, the next generation ai powered data security, compliance and privacy solution. I mean, look, we know AI is transforming business, but there's risk. We talk about it all the time Data risk, bias risks and, of course, compliance challenges. The question for every business today is are you adopting AI responsibly? Well, bigid delivers end-to-end AI and data governance to help enterprises manage risk, enforce policies and ensure responsible AI adoption.

For one thing, businesses want to be very careful about the data they use for their AI right. You want to make sure AI-only access is safe to use relevant data. That's, automatically tagging sensitive data by policy and type, staying away from the stuff you don't want AI to have access to. That's why you need BigID. It's the only leading solution to uncover dark data through AI classification. It's actually a great use of that tool to identify AI risk, to manage your data lifecycle and to scale your AI strategy. Now, the nice thing about BigID it integrates with your existing tech stack, with unmatched data source coverage, and it allows you to automate privacy and security workflows. Take action on data risks with automated remediation orchestrations, you can automate privacy management, regulatory compliance, data rights requests and more. Partners, of course, include everything you're using, including ServiceNow, palo Alto Networks, microsoft, google, aws. With BigID's advanced AI models, you gain visibility and control over all your data. It's the platform Intuit named number one for data classification and accuracy, speed and scalability.

If you think about it, who would be a good customer for BigID? Somebody who has a lot of data in a lot of different places in the cloud on-prem? How about the US Army? Can you imagine? They're in the process of moving into the cloud. They've got dark data everywhere. The us army uses big id to illuminate dark data and to automate data retention. I've got this great quote from us army training and doctrine command. Listen to this quote. The first wow moment with big id came with being able to have that single interface that inventories a variety of data holdings, including structured and unstructured data, across emails, zip files, sharepoint databases and more. Continue the quote. To see that mass and to be able to correlate across those is completely novel. I've never seen a capability to bring this together like Big ID does man. If this can solve the Army's problems, it can solve yours.

Cnbc recognized Big ID as one of the top 25 startups for the enterprise. They were named the Inc 5000 and Deloitte 500. For four years running, the publisher of Cyber Defense Magazine says quote Big ID embodies three major features we judges look for to become winners understanding tomorrow's threats today, providing a cost-effective solution and innovating in unexpected ways that can help mitigate cyber risk and get one step ahead of the next breach. Start protecting your sensitive data wherever your data lives. At bigidcom slash im, get a free demo to see how BigID can help your organization reduce data risk and accelerate the adoption of generative AI. Again, that's BigID. Bigidcom slash im. There's also a free guide to help you understand the risks of generative AI and data-driven strategies to ensure responsible and compliant AI adoption. Again, bigidcom slash I am. We thank them so much for their support of intelligent machines. As I mentioned at the top of the hour, jeff Jarvis is on. Where is he? He's on assignment. He's out doing something.

0:49:25 - Paris Martineau
He's, I truly don't know he's being a big shot somewhere, whatever he's driving, because he was texting me earlier and was like, oh sorry, that was a spelling mistake, because I'm driving, and I was like jeff, don't text me and drive we're talking about the hazards of ai therapy.

0:49:39 - Leo Laporte
what about texting and driving? Well, anyway, we're really glad j Ward could be here, the author of, as I mentioned, a really good book, the Loop. But Jacob has a checkered career in many different areas. He's a very smart guy. We've had him on many shows and we're always glad to get Jacob in. So thank you for taking some time to be with us.

0:49:59 - Jacob Ward
Oh, it's a pleasure Just trying to be part of the club here.

0:50:03 - Paris Martineau
Well, you're in the club, Jacob.

0:50:06 - Leo Laporte
Welcome. It's a Mickey Mouse club in a way, but it's a good club. Lots of stories. We'll just do a tossed salad of stuff. The president of the United States has tapped Palantir to compile data on Americans. That's Alex Karp talking to somebody behind the american flag. Palantir is an interesting company. What do you know about palantir, jacob?

0:50:34 - Jacob Ward
so palantir, founded back in 2004, was trying to solve a problem that the us military and intelligence communities had, which was we. They'd been recording tons and tons and tons and tons of surveillance data, audio and video in Iraq at the time and they had no ability to parse it. You know, they literally had dudes in rooms, you know, scrolling through footage, and Palantir's idea was to get their arms around that and create basically a search engine for video. And they very quickly became sort of a data analysis cutting edge for the US government, and they have a couple of core products One is called Gotham and now the new one's called Foundry and what they basically say is you give us a huge amount of your data and we can parse it and, of course, now that we're in the world of transformer models and AI, they are vastly more effective than they used to be. So their genesis is as a surveillance company.

Now, the other thing that's really worth noting is you mentioned Alex Karp Part of his pitch. So Palantir was created by Peter Thiel, was founded by Peter Thiel, who's you know ultra he's a very conservative libertarian guy was one of the angel investors in Facebook and has been. He was JD Vance's patron. He you know, alex Karp, when he came into CEO, has basically been sort of boasting that this is, you know, not just a really effective you know form of data parsing. It is also, you know, he sort of considers it kind of an ideological and literal weapon in the fight to sort of, you know, push Western democracy around the world.

And I even have a quote here from from something that he said recently. He was, you know, the. He was on an earnings call earlier this year and he said palantir quote this is the quote palantir is here to disrupt and make the institutions we partner with the very best in the world and, when it's necessary, to scare enemies and, on occasion, kill them. And that's part of palantir's spookiness is that they are not just a tech provider, they're very, they're a little bit ideological. So much so that a group of Palantir ex-employees have actually put out a letter that basically called for the company to stop washing its hands of the creepy ways in which the technology is used, to start taking some more responsibility for it. And they wrote in the letter big tech, including palantir, is increasingly complicit, normalizing authoritarianism under the guise of a revolution led by oligarchs. We must resist this trend.

0:53:13 - Paris Martineau
So, palantir this is perhaps decide the point. But how do you become palantir employees and not realize that that's the system you're taking apart, taking?

0:53:22 - Leo Laporte
I could tell you a little bit about that I. I read his book the technological republic, and carp says in that that silicon valley got misled, that in the you know last century engineers, scientists worked with government for the general public wheel. So he talks about the manhattan project to build the atom bomb. He talks about the Manhattan Project to build the atom bomb. He talks about the Internet and he says unfortunately, in this century Silicon Valley has been distracted by crap, by gadgets, by toys, by making money. The first few chapters I was starting to say yeah, yeah, you're right. He says what the best minds in our country should be doing is working with the government to protect the American way of life, our Western values. You use the word democracy, jacob. I'm not convinced Alex Karp cares so much about democracy, although he has in the past been a donor to democratic party stuff.

I mean, he's not. I don't think he's a mega republican, but and in the book, one of the one of the victories that they tout and I think it's a genuine victory is in iraq. There was a big problem with ieds, these improvised explosive devices. You couldn't uh, you didn't know where they were going to be and it was killing a lot of American soldiers.

And Palantir came in, took these disparate you talked about all these different videos and data sources combined them and is able to make a tool that kind of predicted where IEDs are likely to be. And he says I haven't seen the numbers, but he says it was a great success, that it saved lives and that was the beginning of their relationship with the uh, with the us government. He has a 795 million dollar contract with the department of defense, got that last week, just signed 113 million dollar contract. Uh, actually, since trump took office, they've signed more than 113 million million in government spending. Among the things and this is what scared me in this story, in this New York Times story that they are going to be doing is combining data from the Social Security Administration and the Internal Revenue Service. Historically in the United States the IRS has been very protective of its data. For instance, the Social Security Administration does not get IRS data Historically?

0:55:50 - Paris Martineau
Well, no, now it does. Now they're kind of commingling all of it, but thanks to Doge.

0:55:54 - Leo Laporte
and now this is the next step. They want Palantir and it seems on the face of it.

0:55:59 - Paris Martineau
oh great idea, if you could just take everything the government knows about you, Paris, and put it in one big thing and analyze it with AI.

0:56:15 - Leo Laporte
Just think of all the insights we'd get Insights galore. So I can understand why the administration might want to do that, but I think it's also a privacy nightmare. The IRS knows so much about you and combine all the government sources, there's nothing they don't know about you.

0:56:26 - Benito Gonzalez
Also, isn't there a data center somewhere that has, like that's, been storing all of our stuff for like three?

0:56:30 - Leo Laporte
days NSA has been doing that. Oh, it's got all of our phone calls and everything like that the.

0:56:34 - Paris Martineau
FBI agent inside your computer. We're going to commingle that with our IRS data, whatever stuff immigration has all of the little cameras that facial recognize us every time I walk outside the street, and then great. And then we'll just have that in a big data center that I'm sure will never get breached or used for any nefarious reasons.

0:56:54 - Leo Laporte
We kind of know the value, by the way, of this combining, commingling the data. Well, we know it from a lot of different things, but most recently, there's a company called flock that does uh cameras on many, many city streets, license plate recognition cameras and so forth. And I don't know if you just saw a story about the texas cops using flock data in states where abortion is legal to try to track down a woman who went out of state to get an abortion, and this is starting to become a little bit dystopian. Uh, the times where it's creating detailed portraits of americans based on government data is not just a pipe dream. Oh, I was hoping it was a pipe dream.

The trump administration has already sought access to hundreds of data points on citizens and others through government databases, including bank account numbers, the amount of their student debt. They've already promised that they're going to send people who owe money on their student loans into collections, their medical claims, their disability status. You remember RFK Jr saying, head of health and human services, we want to make a database of autistic people. Yeah, imagine Mr Trump could use such information, the Times writes, to advance his political agenda by policing immigrants and punishing critics. It's a. It's a first step, frankly, to eugenics, to Nazi style cleansing eugenics.

0:58:31 - Jacob Ward
To nazi style cleansing. Uh, I I'm really concerned about how this data could be misused. Yeah, let me. I have two thoughts I want to share on this. So one is there is a good way you could do this and and I and I. So not to keep plugging my own thing here, but on the rip currentcom there's an interview I have with the former president of Estonia, a guy named Tumas Ilvas, and Estonia is a country that was crushed under Soviet oppression, incredibly impoverished coming out of the Soviet era and very quickly figured out a digital future for itself and really, like has turned itself into one of the great Western democracies and one of the great sort of things. So in this piece all American surveillance system is coming on the on the rip groundcom we sort of talk about this interview that I did with him.

0:59:13 - Leo Laporte
I have an Estonian digital ID, by the way. I thought that was so cool, so this is.

0:59:16 - Jacob Ward
This is what we're talking about. Okay, great. So when I said for it anybody in the world, I didn't know. I thought you had to be estonian. Well, so, so you, you, when you're, when you're born in estonia, you get a digital id, like leo seems to have.

0:59:31 - Leo Laporte
That's cool it doesn't say I'm an estonian, but it's you know no, but.

0:59:34 - Jacob Ward
But if you have any interactions with the estonian government, it's your way of accessing it. So, basically, what it is is it's it's a. It's a digital id that follows you through every single thing you do in your life Taxes, divorce, marriage. You get arrested, you go to the hospital. Your digital ID is how you are identified in each one of those places, how you vote, how you sign documents. That's right. Everything, everything and that is it creates an incredibly convenient system. You can pay your taxes on your phone in two minutes each year. There's no accountant required, because they already have under your digital ID exactly how much money you've made and who you've worked for.

1:00:09 - Leo Laporte
It's an incredible piece of a little terrifying, though, right.

1:00:13 - Paris Martineau
Okay, that's the thing that terrifies you that you could pay your taxes easily. They know where all my money is Okay, the IRS knows where all your money is.

1:00:22 - Jacob Ward
They're just making you a puzzle Sorry.

Sorry, here's the difference between it being really scary and being really awesome. So the reason it's awesome is that Tumasilvas said it's like a Christmas tree Each record is hanging separately on the tree. So your health records and your bank records and your police records are separate ornaments. You have access to all of those, but none of them can talk to each other without your permission. So the police can't go to your health records. They can't go to your bank records. You are in charge of who gets to see what, and he said that is the specific problem they were trying to solve, because as soon as you merge them all, then the police can start using your bank records and then your health insurance can start using your arrest records to determine your, you know whether you should be insured, and, and and you can start hunting down dissidents and critics. Uh, in the way, that.

Trump's thing could conceivably make.

1:01:16 - Leo Laporte
That's the problem is okay, maybe you like the government today, but that data will be available to the next government and the next government and the next government. I should also point out that shortly after I got my Estonian digital ID, they had to recall it because they had a security flaw. Oh, no way.

1:01:35 - Jacob Ward
Oh yeah.

1:01:36 - Benito Gonzalez
Oh God.

1:01:37 - Jacob Ward
This was back in 2017.

1:01:38 - Leo Laporte
They've probably solved that, I don't know, but they and the other point.

1:01:42 - Jacob Ward
I want to make. I don't know, but they and the other point I want to make.

1:01:46 - Leo Laporte
And this is why this is the central problem is, once you centralize this data, the flaw can let attackers decrypt private data or impersonate citizens, so they had to have all the cards are called and updated with new security certificates. There's your problem. Once you centralize it, however many controls you have, it's still all in one place.

1:02:10 - Paris Martineau
And it just becomes a bigger and bigger target.

1:02:12 - Jacob Ward
Yeah, Well, and here's the other thing that I would say is that so, for all of the ideological talk around Alex Karp and what he's the CEO of Palantir says about, you know, we need to get back to a world in which we're sort of serving a common good. He, like all these tech people, fundamentally seems to believe. So I interviewed Palmer Lucky, who's oh God what a nut job.

1:02:34 - Leo Laporte
He's back at Meta.

1:02:36 - Jacob Ward
He's the CEO. He's the former CEO of Anduril, which is another Peter Thiel company, like Palantir, both named for Tolkien, you know, lord of the Rings mythology.

1:02:45 - Leo Laporte
That's the thing I hate the most about them. Yeah, so weird right.

1:02:48 - Jacob Ward
You know so. But they all say, on the one hand, we need to be creating these, you know something that helps the government and pushes Western values, and blah, blah, blah, blah blah. But at the same time, when you ask them, hey, but you have built a system that involves, that has created huge new moral complexity, and shouldn't you have to also invent the morals to go along with what you've built? They all say no, that's not our job. It's up to people to vote in people to do that. That's democracy's job.

So they wash their hands of what could happen with this stuff. And in Palantir's case, when they've been asked by reporters what the hell are you doing giving Trump this potential? They refer reporters to the blog where they say our software and services are used under direction from the organizations that license our product. These organizations define what can and cannot be done with their data. They control the Palantir counts in which analysis is conducted. That's not a good answer, you know they're just saying it's not our problem, and that's what all these tech companies tend to say it's not our problem.

1:03:57 - Leo Laporte
It's up to the users. It's what the creators of Pegasus, the technology used to crack Apple cell phones, said is well, we only sell our technology to democratic countries. It's not what they do with it. It's not up to us, Of course, that may not even be true.

1:04:15 - Paris Martineau
I only created the ransomware, I didn't put it on the hospital computer system.

1:04:19 - Leo Laporte
I didn't say what they should be doing.

1:04:21 - Jacob Ward
I only manufactured the firearm. I didn't point it at anybody.

1:04:32 - Leo Laporte
Leo as our resident accelerationist. What do you think? Do you think that companies bear some responsibility in thinking about how their products should be used? Oh yeah, and that's? I mean, that's what Robert Oppenheimer realized after he invented the atomic bomb. He said oh my God, what have we done? And in fact, the thing that made him think that was the teller came to him and said now, next, we're going to make an H-bomb. And Oppenheimer said no, no, the war's over, we don't need to do that. And the government said we'll take it from here.

1:04:57 - Paris Martineau
I do think we need a new, some sort of tracker for the amount of times Manhattan Project is referenced in relation to new technologies, because it feels somewhat meaningful.

1:05:10 - Leo Laporte
Well, I think it's a good analogy. I don't think it's a perfect analogy, but it is a good analogy.

1:05:14 - Paris Martineau
I think you hear this again and again with various different technologies, and I just think it's very interesting that that is where our mind immediately goes.

1:05:22 - Leo Laporte
Yeah, well, that was a technology that we, you know. This is what they were talking about in the technological Republic. This is a technology All the scientists, the best minds of America got together to win world war two. Right, and, by the way, the reason? You know that they gave for that as well. If we don't do it, the Nazis are going to do it, which is exactly the reason AI accelerations give.

1:05:45 - Paris Martineau
China's going to do it.

1:05:47 - Leo Laporte
China's going to do it. If we don't do it, china or? Well, now we've given Saudi Arabia the means. Saudi Arabia is going to do it, somebody's going to do it. Maybe that's not untrue, but the problem is these technologies don't. They aren't controllable. I think AI has probably already gone past the point of it, Because when Google wrote those seminal papers on transformers, they made them public. They gave them away. The five or six papers that define how to make LL are in the part of our public knowledge. No one has exclusive access to it. The only way we've been able to control proliferation is controlling these high-end nvidia chips, but that I don't think that's going to last long. So it's out there just as the you know. Actually it's more out there than the atom bomb is. We were somewhat able to control proliferation of nuclear armaments. Your uh, your uh, substack is the rip current and uh, it's a podcast a great podcast it's also I literally had paris.

1:06:57 - Jacob Ward
On the other day, paris was my oh you were on guest because you brought us together, leo and that's nice, it awesome.

1:07:04 - Leo Laporte
We had a lovely conversation.

1:07:06 - Paris Martineau
What did you talk about? We talked about why all the tech workers hate us journalists.

1:07:12 - Jacob Ward
Yeah, that was the theme. Why is it so hard to get inside these companies? And Paris's answer was because they hate us.

1:07:20 - Leo Laporte
Why tech people hate us journalists? Yeah, wow.

1:07:39 - Paris Martineau
That was the theme built into the sourcing base. That and and a lack of like reciprocity in the sourcing relationship, at least in the perception of the sources that you don't see in, for instance, like dc politics, like in politics, the people on the hill realize like, hey, we got to play ball with these journalists and it's also a good way to make sure my boss, you know, gets up higher and my enemies go down lower it's considered.

1:08:05 - Jacob Ward
it's considered like an asset to be, to have a relationship with a journalist when you're in DC, whereas in in Silicon Valley, being in touch with journalists is a not just a fireable offense, but you'll be banished from the industry forever if you're caught in some places talking to journalists.

1:08:19 - Paris Martineau
And so, yeah, we talk a lot about how Beyond just sending a poop emoji automated, which is why yeah, exactly, I mean so.

1:08:28 - Jacob Ward
But the theme of the podcast is basically just how hard it is for Paris to get inside these companies and what a good job she has done.

1:08:35 - Leo Laporte
It's funny. In some ways that's changed, because in my day, tech journalist was very much like beltway journalism, where in order to get access to these companies, you had to play ball with these companies. That's your problem, paris. You don't play ball, that's right, uh, and so?

you remember this too, jacob, you've been around a little bit. Uh, so there was this. It was same thing with, like, auto magazines, where you really couldn't trust the journalism, the tech journalism we have. We've had a guest on talking about this near it. Oh, I can't. I'll have to get her name. I'll look it up. She's been on a show, the tech press was somewhat. Pardon me Nikita Nikita Roy. No, no, no, no, no.

1:09:20 - Paris Martineau
I mean. No, this was entirely true in the Halcyon days of tech journalism when, you know, the biggest, flashiest pieces of tech journalism were like a Wired magazine cover, with a bunch of boys sitting in a dorm room being like gee, aren't these misunderstood geniuses? Swell, they've got some big ideas. Who's going to give them money? Was, I think, maybe useful at a time that the industry was nascent and uh, there was. Well, there's a more coverage or understanding. But now that the power dynamic has shifted in tech versus the rest of the world much less like tech versus even a smaller industry like media it uh begets more, uh critical and careful coverage.

1:10:02 - Leo Laporte
That's because, again, you have a conscience, but you know I have to it would be a lot easier if I didn't.

It wasn't that long ago that if Apple wanted a story, you know, to set the stage for something they were up to, they would go to the Wall Street Journal and plant it, and there were plenty. There are plenty of journalists I won't name names who were kind of in order to get access to the top. Apple was notorious for this, by the way, but I'm sure it was commonplace in Silicon Valley to get to, you know, access to the executives. You play ball with them. This is that beltway journalism and uh, and Apple was is notorious for almost. I mean, they blackmailed me in a way because they blackballed me, let's put it that way, because I play ball and I noticed John Gruber, who's been for a long time, is the creator of the Daring Fireball newsletter, an Apple fanboy went apostate a couple of months ago.

He used to have he has a thing called the talk show. He would do every year at WWDC, at the big Apple conference, where he had the big Apple executives on and all of them, you know the top executives would go on the talk show. This year they're not going on and I think, when it may not be the case, there are other reasons Apple might not want to talk about what's happening in WWDC this year, but, uh, I think it's because he went apostate about a month ago and said Cupertino's lost its way, apple's lost its way, they screwed up badly, they broke my heart, all of this stuff, I mean.

1:11:40 - Paris Martineau
Apple, I feel like, is such a famous example of this that even in like 2016, 2017, just as it was starting to shift a little bit the mood of tech journalism from more like rah-rah to a bit more critical, even at that time, they were infamous in the industry of being like if you did so much as not even criticize one of their products but maybe described it in a way that they didn't want, that was still neutral. They'd be like sorry, you get no review access ever.

We will not answer any of your pr questions, even I.

1:12:13 - Jacob Ward
I went on nightly news on nbc nightly news, probably two years ago something like that, and I listed apple among the companies that I consider to be surveillance companies. I said apple oh, they didn't like that got a phone call over the weekend from a top Apple executive saying how can you call us? That that's outrageous. And I said well, if you look in the dictionary like surveillance just means you're closely watching your customer, and they were like but no person is doing the surveillance. I was like that doesn't matter, your system is automated.

1:12:43 - Leo Laporte
What are you talking?

1:12:44 - Jacob Ward
about. That's the last conversation. That was the last time I ever. That's the last invitation.

1:12:48 - Leo Laporte
They have a list conversation. That was the last time I ever, uh, that's the last invitation. They have a list, they have a list, but lists have gone on also forever. I mean it isn't anything new. Uh, you're right, paris, because I'm john c. Devorex says he remembers being at microsoft 30 years ago and walking by an office and seeing his name on an enemy's list of people that nobody should talk to this guy. He's on our enemy's list and he says he's. He's asked many times since then and it was always denied, but he saw it. So it I believe.

1:13:13 - Benito Gonzalez
I believe him I'm sorry though, but like you guys talking about how tech hates the journalists, but nothing compares to how much game companies hate game journalists nothing but, and yet, at the same time, there are, you know this benito's worked for both of them. I work for cnet and game spot so I know both sides of that he's been.

1:13:29 - Leo Laporte
But there are companies that are in the pockets of game companies as well right, I mean it's true, yeah, definitely and you know when you read them, they.

1:13:41 - Jacob Ward
You know that dynamic of game reviewer, you know, like movie reviewer, or you know, at one time Paris and I would have been employed, you know, and Leo, you, you were probably part of this too, where we were sort of like reviewers of tech rather than talking about its ongoing implications. The rest of it, in the same way, like auto journalists are about like how's the handling as opposed to like what's the horsepower how many kids are killed, getting killed by cars every year?

right, that's not what auto journalism is about. And so that's interesting, benito, that I would think that they'd be okay. You know, sure you might get a bad review, but you're not talking about like it's ruining kids' brains, right, like gaming journalists aren't doing that. I would no.

1:14:22 - Benito Gonzalez
No, no, it's more about, like, access. It's like, especially now, it's totally about access. It's all about access. You don't get an early copy of the game now because you talk bad about this company.

1:14:31 - Leo Laporte
I'm seeing all these articles today.

1:14:32 - Benito Gonzalez
It's not even talking bad about the company.

1:14:33 - Paris Martineau
Yeah, it's not even talking bad, I'm forgetting what recent example there was of some maybe it was a Nintendo game, some large game in the last six to nine months A reviewer had described a level in just plain English terms. A level in just plain english terms and like their entire outlet got, like their access revoked from games from that company for a while because, like they had said in some page of the pr thing, we don't want you to describe levels past this or like something like that.

1:15:02 - Leo Laporte
it's just, I mean, a strange mercurialness and they really love apple might have made this because apple might have really brought this to a head because they're so famous for secrecy and uh, and they're really adversarial. I mean your experience, jacob, is not rare. Uh, they're. They were so, by the way, in my career. I stopped talking to pr, I stopped taking review units. I basically cut off any relationship. I understand that that's a sacrifice, because I don't have the inside. You know, look at what's going on in a company. But my experience has been that's overrated. Often the company's just playing you.

1:15:41 - Paris Martineau
Yeah. But I mean, I think part of the interesting thing is you're able to do that, because your podcast is largely talking about news, talking about your experience of products, covering this industry from kind of a bird's eye view. But when you are a working reporting journalist part of the nature of the job people like to, I feel, like, kvetch online and like, oh you know, journalists get to publish whatever they want and I mean I guess there's some people out there. But if you're an actual journalist employed at a place worth its salt, you have to reach out to a company if you're reporting something on them beforehand, Go over what you have reported on them on background or on the record, Get their take on all of it. And oftentimes that just means sitting there on the phone, sometimes while someone yells and screams and lies at you and tries to get you fired for reporting things that are just normal.

1:16:30 - Leo Laporte
But you can't opt out of that system. Oh yeah, well, I've opted out. I got the privilege of opting out because I own the company, and that's delightful.

1:16:37 - Benito Gonzalez
Yeah.

1:16:37 - Leo Laporte
Hey, we're gonna take a break. We come back. I will talk about a journalist who has gone apostate and has really upset the AI community, but that's coming up in just a bit. Jacob Ward is here, so good to have you. The author of the Loop, his podcast and newsletter, theripcurrentcom. Obviously he's got the best guests on of all. I mean just really good, just following in your way, jacobwardcom. Great to have you. Paris Martineau, tech journalist, who is not afraid to offend any company and whose lava lamp is finally heated up.

It's heated up, it's really it's popping, it's having a great time it's a popping lava lamp Our show today, brought to you by our friends at OutSystems. This is the leading AI powered application and agent development platform. Now, this is a really interesting story. They've been around for 20 years. You might say, wait a minute. Outsystems has been around for 20 years. They weren't using AI back then. They were very early on the low-code, devsec automation, things like that. They've been doing this for a long time. The mission of OutSystems to give every company the power to innovate through software. But I have to tell you, these guys are smart and when it became possible to merge AI with DevSecOps automation and low-code, no-code environments, they jumped right on that and they have created something truly powerful.

Traditionally, you've got two choices when you need software for your enterprise. The IT teams can buy something off the shelf a SaaS product. Yeah, it's fast, you got it right away, but you don't have any flexibility and you've lost all differentiation because you're not the only ones using that tool. Oh, okay, so if it's not buy, maybe we'll build it, but I take it for me. I've done it. Building custom software costs a lot of money, a lot of time and, frankly, you often get a subpar product. So that build versus buy conundrum has been around a long time, but now AI has changed all that. They've forged the way for another path. It's, as I said, the fusion of AI, low-code, devsecops, automation into a single development platform, outsystems. Your team will build custom applications with AI agents just as easily as buying generic off-the-shelf sameware. And what's really great about OutSystems they know what you need, so they've built in flexibility. Security comes standard. Scalability comes standard. With AI-powered, low-code team can build custom future-proof applications at the speed of buying with fully automated architecture. It's all built right in security. The integrations you want are there, the data flows absolutely control with permissions. The truth is, outsystems is the last platform you need to buy because you can use it to build anything. You can even customize and extend your core systems. What I'm saying is build your future with OutSystems. Visit OutSystemscom slash twit to learn more. That's OutSystemscom slash twit. We thank them so much for supporting intelligent machines and you support us if you go to that address, so they know you saw it here, outsystemscom slash twit.

Cnn offended the AI community and, in particular, dario amode, who is the anthropic ceo, by saying there's a little bit of hype going on. Uh, he's amode. Uh, told anderson cooper on cnn. Ai starting to get better than humans at almost all intellectual tasks and we are going to collectively, as a society, grapple with it. Ai is going to get better at what everyone does, including what I do, including what other CEOs do. He says 50% of jobs will be wiped out in the next two years by AI, but not his. He even says he could be replaced. Cnn says it's all part of the ai hype machine. There will not be. What do you think? Will there be a white collar bloodbath?

yeah, hello there is a certain amount of hype, though, isn't there? I mean, sam altman is the king of this, over hyping what ai can do, but is ai going to replace that many jobs?

1:21:09 - Jacob Ward
I don't think I mean yeah, so you can't trust what, the, what, the what the ceos of the companies say, because they it serves their purposes to make this thing seem disruptive, but like or even scary we've.

1:21:19 - Leo Laporte
We've always said some of this is them saying, oh my god, they're so good.

1:21:24 - Jacob Ward
No, I know it's a weird form of marketing but it is a form of marketing it's true, but but at the the same time, you know, kevin Roos at the New York Times had a good piece that came out about the job apocalypse and he made the good point that, like labor, figures are showing that new college graduates, you know employment rate is down.

1:21:40 - Leo Laporte
It's a terrible time, terrible time, terrible time.

1:21:43 - Jacob Ward
They begin out of college, and it's also a time when I mean so there's the general tightening of the belt in advance of the recession that everybody seems to be convinced we're going to go into, but there's just this sense from all these companies that, like you know, they just don't think whole categories of jobs are going to be necessary. The Bank of Singapore just said that it's going to replace like thousands of people that would normally review loans and, you know, be a clerk and all these different ways. They're all going to go away, and they were the first big bank to say it's specifically because of AI that we're going to get rid of all these people, that stuff. I think, even if it proves that it's not able to replace people, I think these companies, and these shareholders especially, are so eager for that to be the case that we're going to experiment with that for a little while, and I think that's going to be a terrible thing for people in the job market.

1:22:35 - Leo Laporte
Amode said that within five years, ai could eliminate half of all this is important entry-level white collar jobs, which means unemployment could spike to 10 to 20% within the next five years. Maybe that's true. There's a long-term consequence, though, if you lose entry-level jobs, because there's no way for somebody to get to a senior management position without going through the entry-level job.

1:23:00 - Paris Martineau
This is what my friends that are senior software engineers bring up because obviously I mean, I think these people are typically the biggest AI proponents other than you, Leo, that I talk to on a day-to-day basis. They're like, yeah, love AI, love using it for my work. It's incredibly helpful. But the one qualm they bring up again and again is they're like I don't know what this is going to do to the pipeline of people that come up to work with us. Like the sort of tasks that typically you'd have an intern or fellow or like really entry level, like junior engineer be doing and working on to kind of cut their teeth, are suddenly being handed off to AI and obviously that makes things a bit faster. In some ways it's cheaper, but I do wonder what the consequences to the pipeline is going to be.

1:23:48 - Jacob Ward
Yeah, I mean I don't know that you hear from you know Sam Altman was was on. I have a piece on this. You know was talking about how there's only a text thread like a WhatsApp group or signal group or something of all the big tech CEOs, right and, and supposedly they he's got a a bet going with a few of them as to who when they're going to have the first billion dollar, one person company and that they are kind of cheering the idea that you're not even going to need a hierarchy to have to move up through the one man, unicorn.

1:24:22 - Leo Laporte
Yeah, that's right. And it'll be a vibe quarter to boot, because you can't write enough code by yourself.

1:24:28 - Jacob Ward
And I think that that is the dumbest possible thing to be jeering for, I mean, but I also think it's like so perfectly on the nose for how these guys think.

1:24:39 - Leo Laporte
Because if you're, that one guy, though that's good. No, no, no.

1:24:42 - Paris Martineau
You're not even thinking what's going to be the first company that, with a single person, creates a billion dollar in like value for people. It's what's going to be the first billion dollar one person company, which means what is going to be the first company probably vibe coded by one guy that ends up with a billion dollar valuation, because they've kind of talked investors into believing that one guy and his chat gpt rapper are worth uh, them giving hundreds of millions of of dollars of venture capital at a $1 billion valuation, which is not the same thing as creating a billion dollars of actual value in the world.

1:25:16 - Leo Laporte
That's a good point. Well, I'll give you an example. Here's an Indian AI startup Builderai. It was a unicorn. They promised to use AI to create software. They raised over half a billion dollars. Turned out that when you had their AI code something, it was sent back to the home office where 200 Indian engineers wrote the code. There was no AI. That just seems inefficient.

1:25:50 - Jacob Ward
At least they had a job.

1:25:54 - Paris Martineau
At least they had jobs, at least they were working.

1:25:57 - Leo Laporte
On may 21st, builder ai announced it was entering bankruptcy, saying historic challenges and past decisions have placed significant strain on our financial position.

1:26:11 - Jacob Ward
That's funny, I don't find myself outraged by that one.

1:26:14 - Benito Gonzalez
So this is. It was so. To them it was cheaper to actually hire people hire people.

1:26:18 - Leo Laporte
Yeah, that's a good point.

1:26:21 - Paris Martineau
I'm like that seems where the customers unhappy, they got the code this is the conspiracy theory I have, which which may be played out, may not which is that we have all of these different fields that, like you just said, executives are like well, why don't we give it a shot? We'll get rid of thousands of workers, replace it. Ai companies are largely in their infancy and charging rates that don't allow them to actually make money. But I wonder, if, let's say, even a world where it very easily becomes a tool that can do the same jobs that you're eliminating, then they're going to jack up the prices, much like Uber and Lyft, whenever they suddenly had to make money. And will that be viable for any of these companies? Then if suddenly you're paying a thousand times a thousand dollars a month to replace some of your employees, perhaps it wouldn't be um, I'm gonna take a break when we come back.

1:27:29 - Leo Laporte
I think this might be the best named website ever walter writes dot ai. Uh, walter white, of course, was the bad guy in breaking bad, but this is something, a different kind of bad guy. Uh, we will take a break and come back with jacob ward. Great to have you, jacob paris martineau, to have you, jacob Paris Martineau. Let's just tell Jeff that we moved.

1:27:53 - Benito Gonzalez
Nobody home, Jeff Tell.

1:27:54 - Paris Martineau
Jeff, that the podcast is located on the other side of a bridge, we can't come anymore.

1:27:58 - Leo Laporte
Sorry, jeff, I miss Jeff. We can't wait to get him back, but I love having you on, jacob. I really appreciate all the content Our show today brought to you by this little doohickey here. Let me show you. This is my, this. Uh, let me push the right button. This is my thinks canary. Look at that. It looks like what I don't know an external hard drive. It's about that size, about the size of my hand, got two connections, power and ethernet. Put this on the network and it can be almost anything I want it to be because, you see, this Thinkst Canary is not what it looks like. It can be.

Well, mine is set up right now to be a Windows 2019 server. For a long time it was a network attached storage device. It can be pretty much anything. Let me log on to my Thinkst Canary and I'll show you. This is we call this one attic because it's sitting in my attic and I can make this thing be. Let me show you. Well, yeah, as I said, it's a Windows 2019 Office file share, but let's just see what else it could be. It could be IIS, it could be SharePoint, it could be Windows 7, windows 10. It could be a Linux database, a Linux proxy, a Mac OS X file share and, on and on, even a scatter device. This thing is awesome.

Right now I'm going to turn it into a, let's say, a Siemens Cymatic 300, plc, 300 plc. Now why would I want one of those in my studio? Well, because if somebody breaks into my system, if they get through my perimeter defenses and they start snooping around, they're going to see that siemens sematic 300 plc and they're going to want to know what's it doing there and what can it do, what's it got? And they might even try to break into it. I could, by the way, I could also use that Thinkst Canary to create what we call Canary tokens. These are little files I can distribute around, even on my. I have some on my Google Drive. These tokens can look like anything a PDF, a Word file, an Excel spreadsheet. They can even be things like tailscale config files. They can be a DNS host name that alerts when queried. It can be a credit card that alerts when somebody uses it. See, I can set this up as a little credit card and if somebody gets this credit card and uses it, I'm going to get an alert. That's the whole point of the Thinks Canary, just the alerts that matter, the alerts that tell you somebody is inside your network, somebody got that credit card, somebody's accessing files on your Google Drive. And the alerts can come any way you want them they can be text messages, email, they could be Slack. Of course it supports syslog. There's an API that supports webhooks.

The Things Canary is genius. Created by a team of people who have spent years teaching governments and companies how to break into systems, they know exactly what bad guys are looking for. They designed this honeypot to be exactly what it says it's going to be, with the MAC address, with all of the details. When it was my Synology NAS, it had a Synology MAC address, it had the DSM-7 login. It looked exactly like the real thing. But the minute someone accesses, say tries to brute force, this fake internal SSH server or my NAS server, or somebody accesses that lure file I put out there, I'm going to get the alert that says there's somebody in your system. Choose a profile for your Thinks Canary device, register it as I just did, with a console for monitoring and notifications. Then you just sit back and you wait.

Attackers who've breached your network can't help themselves. They don't look vulnerable. They look valuable, malicious insiders snooping around. They're going to tell you, they're going to make themselves known by accessing your things to Canary. I think every network needs these. A big bank might have hundreds of them spread all around. In fact, if you go to canarytoolscom, you'll see some of the big companies and the CISOs that you know who use Canaries and wouldn't do anything without a Thinks Canary, who say this is it. You can have all the perimeter defenses in the world, but on average, a company that gets breached doesn't know there's somebody inside their network for 91 days. Know there's somebody inside their network for 91 days, three months for an adversary to browse around, steal stuff, plant time bombs. You don't want them to do that. You need a Thinks Canary. Go to canarytools. Slash twit. 7,500 bucks a year will get you five of them. As I said, you might have hundreds, you might have a half dozen, you might have one or two, but just as an example, five Thinks Can canaries, your own hosted console, you get upgrades, you get support, you get maintenance for that year and, oh, if you use the code twit and how did you hear about us bucks you're going to get 10 off the price for life.

Now, maybe, after hearing all this, you're still a little skeptical. What I want to tell you, they understand. That's why they offer a two-month money-back guarantee for a full refund. 60 days Plenty of time for you to say I don't know. I got to tell you, though. We've been doing these ads for Things Canary for eight years, and no one has yet ever, ever, asked for their money back. Not one, not one.

Because once you get this, you go oh, I needed this. All along, you know, you realize this is brilliant. You could do your own honeypots. Trust me, it's murder, it's hard, it takes some real sophistication. These things are done right. Canarytools slash twit. Don't forget the twit offer code, so you get the 10% off.

Canarytools slash twit. Use Canarytools slash twit. Use the code twit and don't forget you got 60 days to change your mind, but you would be the first and I think it'd be a mistake. Thank you, thanks, canary. We appreciate your support. It was only a matter of time. I've been reading sob stories from college professors who say oh God, now, instead of grading and reading essays, I I spend most of my time figuring out if stuff was generated via ai. It's a, it's a. I think it's kind of a uh, a whack-a-mole because well, here's walter writes humanize ai text and bypass any ai detector with walter writes ai. So, uh, I guess you subscribe to the trusted by the way, you should read this by a hundred thousand students and professionals to bypass ai detectors using Walter Wright's AI humanizer tool. It was just a matter of time Humanize.

1:34:53 - Benito Gonzalez
AI text.

1:34:53 - Leo Laporte
I will say Take the em dashes out.

1:34:56 - Paris Martineau
That's the thing I was about to say. I'm so annoyed that AI has made it more problematic to be using em dashes Because, as a journalist, I feel like something a significant portion of journalists have is we can't quit em dashes and parentheticals. You got editors being like do you need this? And I'm like you can pry my em dashes from my cold dead hands.

1:35:17 - Leo Laporte
All right, I'm going to say to perplex me, write me a paragraph love story in the style of Danielle steel how about that? And I'm going to see, I'm a first, so we're in other words. I'm going to see, I'm a first, so we're. In other words, I'm going to generate some AI garbage. I'm going to make it 300 words.

1:35:34 - Benito Gonzalez
Yes, anthony has the right answer here, because most of the LLMs are trained by journalists, so there are M-dashes in there. There's all the stuff that journalists use.

1:35:41 - Leo Laporte
That's why the M-dash is not there. It's not regular people.

1:35:44 - Paris Martineau
Oh. In the golden haze and all the, all the bullet points are probably because they were trained on Business Insider.

1:35:50 - Leo Laporte
They've got the Axio style down. In the golden haze of a San Francisco evening, alexandra stood on the balcony of her Pacific Heights apartment, the city lights twinkling below the scattered diamonds. She'd always been fiercefully independent, building her career as a renowned architect. Of course, brick by determined brick.

1:36:11 - Jacob Ward
All right, let's see what Walter Reitz does with that, danielle Steele should sue the bejesus out of her. You know, right, like that's the thing. That's the whole basis of the lawsuit, right.

1:36:21 - Leo Laporte
Don't you think, though, that she probably by now is using AI?

1:36:29 - Jacob Ward
You know, the line that bothers me most on the on the marketing of that site is that it says that it's plagiarism free, as if somehow the detection of plagiarism is the problem, not the plagiarism well, now let's be fair, jacob.

1:36:43 - Leo Laporte
If you're using an ai to write something, that's not plagiarism, is it?

1:36:47 - Jacob Ward
yeah, I yeah. I guess that's right. I guess that's right, although you're plagiarizing the AI. Well, I suppose what you're doing is you're plagiarizing the AI yeah.

1:36:54 - Leo Laporte
I'd argue that's plagiarism. Can you plagiarize an AI? Really? You know what this?

1:36:57 - Jacob Ward
reminds me of. I had a conversation with a professor at UC Berkeley who did a class on how to read literature deeply and they read like three books over the course of the semester and they try to get deep inside them. And he says that the number one problem he has is convincing these kids who have just shown up there I think they're freshmen that there is an intellectual difference between having read the book and having read chat GPT summary of the book, that they don't think there's any difference there. And then I was just talking to my wife.

My wife just had a conversation with another academic yesterday who said that the problem he's having is that his students actually believe they have written the thing when they get chat GBT to do it for them. I wrote the prompt. Yeah, I wrote the prompt, and so they feel ownership of the product.

1:37:45 - Leo Laporte
That's like saying I made the Ikea sofa, I put my dollar in.

1:37:51 - Paris Martineau
I paid my friend Billy 500 bucks to write the final exam.

1:37:56 - Jacob Ward
That's my final exam for you my 500 bucks. I mean, hey, okay, okay, we're going to be fine, you guys, we're going to be fine. I have to say this is not anything new?

1:38:07 - Leo Laporte
because, okay now Jacob, you tell me you never used Cliff Notes in your entire academic career. Oh, I'm sure I used.

1:38:14 - Jacob Ward
Cliff Notes, no question, but at least I read them. I read the Cliff Notes. I read the Cliff Notes. You know, like I didn't just say, you know Cliff Notes didn't write it for me, give me the Cliff Notes for James Joyce, ulysses. Yeah, the comic book version of the classics. Sure, I read those, but I read them. I don't know.

1:38:36 - Leo Laporte
It's a different world. Yeah, I don't know. I'm glad I'm not a college professor. That's all I can say.

1:38:43 - Paris Martineau
If you've been using notebook go ahead, please I was going to say I was interviewing a college professor. I I was going to say I was interviewing a college professor, I believe, teaching a course on tech ethics, and it was a partially online course. This was about a year ago and as part of like one of the class, and in that first assignment, almost every single uh student used ai, which you could tell because their intro messages introducing themselves were word for word like the same message. And I'm like, if you cannot even write a three-sentence introduction message in your tech ethics and humanities course, like, what are we doing?

1:39:33 - Leo Laporte
That's pretty funny, I was just checking, you can still buy Cliff Notes by the way for everything, so they're not out of business.

1:39:42 - Paris Martineau
They call them study guides.

1:39:45 - Benito Gonzalez
These are AI-generated at this point, right, come on, yeah, maybe so oh, that's interesting are a generated at this point, right, come on, yeah, maybe, so, maybe.

1:39:54 - Paris Martineau
So that's interesting. I will say I met someone a week or two ago who's a phd in philosophy at nyu and part of one of his like side gigs is he writes those, but for like philosophy related topics, and I was like that can't, that job can't exist. There's one of them, right, there's one of the white-collar jobs it's gonna disappear that's going away.

1:40:13 - Leo Laporte
By the way, at the bottom of the cliff notes web page, chat, pdf, homework and exam help. Get instant solutions and study help. Yeah, you're right, ai has come to cliff notes. Even cliff notes were too much work. That's the problem. Yeah, too much, it took too long. Took too long, that's right. So finish assignments faster with ai, get solutions, summarize documents, explain concepts in seconds. I have to think, if I were a student, I I wouldn't resist this. I mean, if I had a professor's lecture notes, I would consider I would still do the work, but I would consider using these tools I mean, I also think that you, you gotta, you got I'm.

1:40:50 - Jacob Ward
So I had a conversation with a high school senior who's like a valedictorian, super smart kid, and he was saying, as judge, if he was coming out. He said you know, he's like zoom during COVID was one of these things that taught a whole generation of kids that the only purpose of school is to get through it so you can have a house and a Tesla. You guys taught us that a degree is about like getting out there and making money, that it's transactional and so why wouldn't we use any possible thing we can to get the better of that deal? Basically and I was like kid, I can't argue with you, that's it.

1:41:26 - Paris Martineau
No, I mean, I do think that that's ultimately like. The core of it is the what we've seen over the last couple of decades, and maybe this has just been the way it is always.

Maybe it's foolish of me to ever think there was a world where college was thought of as something you did because the intellectual pursuit is what you're going there for, but that's how I thought of college when I was there, so I was like I'm, I obviously want to get a job out of this, but also what I am getting from this is the ability to critically think, yes, to be exposed to new ideas, and I do think that gets lost in kind of like the neoliberal effication, the commodification of college and degrees, but I think that's also probably a very uh rational output and effect of college degrees being as expensive as they are totally.

1:42:17 - Leo Laporte
I also think it starts earlier than that. Paul graham uh who is with founder of y combinator and uh a list packer and uh, I always like reading his stuff wrote a great book called hackers and painters Painters, and the first chapter talks about high school and that every kid in high school pretty quickly gets. They're not there to learn, they're being housed so their parents can work during the day and they will be kept out of trouble. Teachers aren't there to teach, the students aren't there to learn. It is a form of um, it's a prison, basically, and they and they learn it now.

This is interesting because he wrote this 20 years ago. He wrote this before not only ai, before social networks were really prominent, before twitter, and he he talks about why kids in high school are suicidal and depressed, why they're rebellious. He says and it's nothing new, this has always been a problem as long as we've been housing teenagers dangers, uh. And he said you don't in in. You know, in the renaissance times, when a kid at the age of 12 became a? Uh, an apprentice and was started to learn a skill, you don't read about these same kinds of uh enemy. And anyway, with these kids, with the kids in the renaissance because they were doing something, they were working, but in school it's just this empty, meaningless exercise that everybody knows. He says the real problem is the emptiness of school life. We won't see solutions till adults realize that the adults who may realize it first are the ones who are themselves nerds in school.

This is the. The chapter is really about why nerds are unpopular, he says because they're smart and they understand that being popular is not the goal of high school. Do you want your kids to be as unhappy in eighth grade as you were? I wouldn't. Well then, is there anything we do to fix things? Almost certainly there's nothing inevitable at the current system. It's come about mostly by default. Nerds in schools should not hold their breath. Maybe one day a heavily armed force of adults will show up in helicopters to rescue you, but they probably won't be coming this month, I think. In a way, reading this reminded me this was a. Kids were bored and depressed even before there was social media, even before there was ai. This is not anything new. This is what we've done with our education system I mean in some ways, the.

1:45:00 - Paris Martineau
Obviously the education system is an improvement from what we had previously, which was wide scale child labor, yeah, that's right, but I also, I don't think but at least a kid. A kid at the age of 15 who's laboring at something and making something is doing something of value that kid may not have any of their fingers, but they're doing something that's true they may be working johnny may be working with only one lung, but he's gonna survive to the ripe old age of 19 and father about 17 children, and that's beautiful smoke.

1:45:30 - Jacob Ward
He gets two smoke breaks a day.

1:45:31 - Paris Martineau
You know it's and you know sometimes he goes through his five to six packs of cigarettes during those two smoke breaks. But he deserves it.

1:45:38 - Benito Gonzalez
He's a hard worker, he deserves it that's during the industrialization though I think he's talking about pre-industrialization when it was working for their fathers, right, yeah, yeah, it was a painter. Now you're gonna learn how to paint like your dad painted, because that's just how it worked I mean, I think there was.

1:45:54 - Jacob Ward
There was this moment when it was supposed to be as, as paris says, you know this, this quality it was supposed to be to create a more informed you know citizenry that we were, that everyone was going to be a better smarter citizen it was. It was this great democratic ideal of free education was all about that, but yeah, but, but you know, yeah it hasn't really well.

1:46:12 - Benito Gonzalez
College became a. College became a commercial enterprise is what happened I completely agree with you.

1:46:17 - Leo Laporte
The colleges, at this point, are for-profit institutions, you know, and not really designed to deliver pride themselves on how many people they turn away.

1:46:26 - Jacob Ward
Yeah, you know how much of their endowment they don't spend.

1:46:30 - Leo Laporte
So I guess my point is a smart kid today is probably not going to learn in school, or even in college. They're going to, but they have opportunities to learn on their own. It started with things like Khan Academy, but now with I mean, notebook, lm, and there's a lot of tools out there that you could become really good at something on your own. I think AI can make excellent teachers.

1:47:05 - Paris Martineau
Ostensibly yes, and being socially pressured to go through something like college is learning, and developing critical thinking skills is not intrinsic for most people no I agree it's an inherently uncomfortable process and I did you learn that from school, or did you learn that on your own? I? I mean, I learned that from school, but I'm a dork, but I was always the sort of person?

1:47:28 - Leo Laporte
were the kids around you learning?

1:47:30 - Paris Martineau
that Some. I mean I went to high school and other school in Florida so it was kind of grim, but there were some people who were actually. I know it's grim everywhere and there are handful.

1:47:40 - Leo Laporte
I got a lot out of school too.

1:47:52 - Paris Martineau
Just because the system is deeply broken and profoundly messed up doesn't mean that the idea of fostering a community of knowledge and some sort of large systemic force to push kids outside of their box, their own bubble and introduce them to concepts they normally would not come across of their own volition is bunk like. I think that's a worthwhile endeavor and we shouldn't abandon that just because the reality is.

1:48:09 - Leo Laporte
I completely we know we shouldn't abandon we just because the reality is. I completely. We know we shouldn't abandon it, we should fix it. Yeah, because I agree with you Free education is a vital, important part, totally, and I worry that Khan Academy moves us in the opposite direction. Really you think so?

1:48:23 - Jacob Ward
I just worry that it makes it just will hew to the same thing that all these other tech products wind up producing, which is the effect of saying, well, we don't need to spend the money on that because, look, it can be done digitally. And even though it's not proven that that is the right outcome, people want to, just they want the, they have the shareholder perspective on it. I just think we got to get back to this idea that there's just some fundamental things that cost money and don't pencil out in the short term, and education's got to be one of those.

1:48:55 - Leo Laporte
I also think, though. I mean, I look at my son, who didn't was not a great student, but he taught himself to cook from YouTube videos, and he's quite an accomplished cook now. He's a celebrity chef, in fact. I think there are a lot of kids.

1:49:10 - Paris Martineau
Your son has a very unique uh experience.

1:49:13 - Leo Laporte
I know, I think there are a lot of kids who learn things on youtube. They watch, you can look at, you can go to youtube and learn how to do almost anything. Yeah, absolutely. I don't think we should abandon a public education because of that. Yeah, but I do think the opportunities for a smart kid, a kid like you, paris, are much broader than just what they can learn in school yeah, yeah, certainly yeah, and that's a good thing yeah, I, the only part I worry about is just is rewarding kids who are more comfortable alone and rewarding them for I think, that think, I think

1:49:50 - Paris Martineau
that I mean especially but I think it's also something that we need to keep in mind, especially given that we have a large amount of kind of coalescing social, political, social crises that kind of all stem back this loneliness epidemic. You both see it in men with masculinity. You see it with what you guys were just talking about in terms of whole generations of students that kind of came of age educationally in a completely isolated space where your main social interaction is entirely digital. Your main experience of school and interactions with authority figures is entirely through a screen in your bedroom. I do think that there is, I don't know, some version of that, like we should be pushing more pro-social behaviors just as a society, because I think people end up I think everybody should do square dancing and join a ski ball club.

I would. I do think that would solve a lot of America's issues.

1:50:47 - Leo Laporte
No, it's funny because I was talking about this earlier today. I walk around my neighborhood, typical suburban neighborhood. When I was growing up, there were always kids in the street, in the front yard playing. It's it's like the twilight zone there's nobody around, it is dead quiet. Kids do not play outside anymore.

1:51:07 - Benito Gonzalez
Also, that's a very good point that this is a very american conversation as well, like oh, yeah, yeah paul was saying that when we were having this conversation earlier today.

1:51:17 - Leo Laporte
He said I see that in canada there's kids playing the street all the time. We what do we do wrong in america? Was it scaring people with stranger danger?

1:51:24 - Paris Martineau
I do think that's a big part of it because so I'm still in a lot of um parenting groups, not because I'm a parent myself, but because of my previous beat on Kids Online Safety. I joined all these different chat rooms and groups so I could just kind of get the pulse, and now I'm kind of addicted to just seeing what's kid that wants to play outside. Now there is this like strange social pressure from other adults that like they see a child out in the street and they start posting on next door, they start calling yes, they call the local police.

They're like who is this child and who has abandoned them?

And it's like, no, that child is playing hide and seek with their friends. Their house is two blocks away, but we've somehow, at least in these pockets of America, created this incredibly antisocial, paranoid culture that I think is really hard to unwind. There are these groups that I've covered in my previous journalism that are kind of like wait till all these different parent movements to get kids off screens have more of a return to a traditional outdoors childhood, and one of the problems they have, uh, come up across at least one of the chapters here in brooklyn and new york is like they're like we don't have places for the kids to go. That's like we had to go and they had to like go and meet with coffee shops or like libraries or bookstores and be like will you allow children to hang out like teenagers, to hang out here, maybe only buy one coffee or maybe not buy any books but just socialize? And they had to be like yes, and then they put up little stickers on their wall like window it's like children.

1:52:57 - Benito Gonzalez
You can socialize here, it's because you lost them all, you all lost them.

1:53:01 - Leo Laporte
Yeah, there's no mall, so I think maybe that's something we as a society can address and we can kind of foster that kind of thing, and I think that's probably something we should do. Um, you know, I know, as an adult, I go, I, you know, we, lisa and I go out and we do swing dancing, I, we take tai chi, we do stuff to get out in the world you're learning piano in our group chat this week that he's.

1:53:24 - Paris Martineau
He's a maestro on on the keys. I'm not, I'm a very early learner.

1:53:30 - Leo Laporte
But uh, it's. It's a great thing to do and I think it. I think it's important to go out and be with people and socialize and go out and do things um, but I don't think the society fosters that really. You know, maybe it does with you young people.

1:53:44 - Jacob Ward
I don't know, I don't want to uh, this is going to sound like a weird prescription. But uh, this is going to sound like a weird prescription. But the keeping separate of kids and booze is looked upon as really weird in other countries.

1:53:56 - Paris Martineau
Jake's coming out in favor of getting the kids drunk. You have to hear him first. Here's the thing.

1:54:01 - Jacob Ward
Here's the thing we're not saying get the kids drunk, but what we're saying is that the parents should be able to drink socially with their kids around.

1:54:07 - Leo Laporte
Oh yeah, this is a thing in the rest of the world, in the rest of the world, in france, everybody's sitting around the table. You put a little water and wine in the glass. The kid can drink that?

1:54:16 - Jacob Ward
well, not, we're all literally, literally. I'm not saying that the kids should do it, I'm saying the the uh, they should be exposed to it. Parents shouldn't be separated, be allowed to drink a little bit and I was talking about this the other day I was.

1:54:27 - Leo Laporte
we got lovely French restaurant, the pub culture allows your kids to come in.

I was at a French restaurant lovely, wonderful kind of bistro place Family. They were having a birthday party or something. A bunch of adults, there were three or four kids. None of the kids were listening to what the adults were saying. They were either on a screen or the I mean the. The kids weren't part of the group. How are you going to learn how to have a conversation If you, uh, if you aren't in a group? Anyway? We got to take a break. We got to take a little break. We're getting way behind. I only have five minutes left in this two hour show and we have 400 stories still to do. Uh, jacob Ward is here. It's great to have him from the ripcurrentcom, paris Martineau, looking for work. She's such a talented investigative journalist.

1:55:20 - Paris Martineau
Listen, I'm out there I'm doing stuff. Listen, if you want to hire me, that's also interesting. But if you've got an interesting story of corporate fraud, oh yes, I'm working'm. I'm working, I'm working on oh good, um, but if you have an interesting story of corporate fraud or honestly, just an interesting story in general, be it tech related or otherwise, hit me up.

1:55:42 - Leo Laporte
My signal is martin, 0.01 I'm a big proponent of going your own way, of not working for the man or, in your case, the woman but doing your. You know, being an independent freelancer. Go for it. This is a good opportunity for you.

1:55:56 - Paris Martineau
I mean, yeah, no, I think it's just very interesting to be able to suddenly write about anything for anything.

1:56:00 - Benito Gonzalez
Yeah um, you're free. Excited about it free at last.

1:56:04 - Paris Martineau
Free and lance, as they say. Free and Lance.

1:56:08 - Leo Laporte
Our show today brought to you by Storyblock. You know if you have worked anywhere a publication as a marketer, anywhere with a website really, the pain of legacy content management systems. We've all been subjected to them. They promise enterprise grade features but they deliver slow, clunky systems that need dev support even for the smallest little update. And, of course, if you're trying to move fast and who isn't? That's a nightmare. Well, storyblock changes that. I love this.

Unlike those monolithic CMSs, storyblock does something that I am a big believer in. It's called Headless CMS's. Storyblock does something that I am a big believer in it's called headless. It's headless. It completely decouples your backend from your front end. That means you've got a robust, solid front end with a strong API and a great content management system and, by the way, a great editor built in. But the front end can be whatever you want. Developers can build in any framework React, astro, vue. Marketers get that intuitive visual editor to create and update content. No dev tickets. You want to move that image over three pixels? You go ahead and do it.

Storyblock scales whether you're a freelancer or part of a global enterprise. They have a global CDN, aws, data centers in the US, europe and Asia. It's built for performance at scale. And man is Storyblock Enterprise ready. It's got all the features Enterprise requires Role-based access control, enterprise SLAs, of course, top-tier security the things Fortune 500s demand. One global e-commerce giant switched to Storyblock and cut content update cycles from weeks to to ours. Another major brand, empowered Marketing, launched campaigns independently. Yeah, you can do it, freeing up the devs for bigger projects. If you've been there, you know what I'm talking about. Storyblock content loads fast anywhere in the world. Better UX, higher engagement, improved SEO and with that beautiful real-time visual editor, marketers see exactly what their content will look like before publishing and no more endless back and forth over minor tweaks.

Storyblock has an API-first approach. That means your devs get fewer interruptions. Marketers get more autonomy. Everybody's happy. If you're an agency, oh, you'll love it. Storyblock offers multi-client workspaces, flexible permissions and seamless collaboration tools. You can manage multiple projects without disrupting development workflows. So whether you're a startup, an enterprise or an agency juggling multiple clients, storyblock gives you the power and flexibility you need. Try it today. Storyblockcom slash TwitTV dash25 and use that code twitt25. Twit listeners get 20% off for three months on the Growth or Growth Plus plans. That's storyblockcom slash twittv-25, the offer code twitt25 for 20% off the first three months on Growth and Growth Plus plans. First three months on growth and growth plus plans. S-t-o-r-y-b-l-o-k. Dot com, slash t-w-i-t-t-v. Dash 25 and the offer code is twit25. Thank you, story block.

1:59:26 - Paris Martineau
thank you, do we lose jake consumed by the zoom.

1:59:31 - Benito Gonzalez
Gods yeah, he's texting. You got this connect. I don't know what happened. It might have been, it might be power, it might be internet, I don't know he's not on the horn, well, let's we'll continue.

1:59:43 - Paris Martineau
I've got a. I've got a silly story for us, line 111.

1:59:49 - Leo Laporte
There's a lot of stuff in the there are so many silly stories here actually it's in the AI gone wild section.

1:59:55 - Paris Martineau
This is a advice column submission from slates.

2:00:00 - Leo Laporte
Uh, dear prudence oh, I love dear, prudence helps.

2:00:04 - Paris Martineau
My husband's best man made a stunning admission during his wedding speech.

2:00:07 - Leo Laporte
I might never get over it, so I'll why does he have the chat gpt katanis on his head who who's to say dear prudence?

2:00:15 - Paris Martineau
my partner of five years and I just got married after two years of extensive wedding planning and preparation. We had a very large guest list, all this stuff. Your husband was very intentional about making sure the labor of the wedding planning was split. As equally, we agreed that we wanted to write our own vows because we thought that it would be more meaningful than using traditional ones. As a self-admitted perfectionist and English major, I spent an immense amount of time on mine. I was really looking forward to hearing what he wrote.

This person writes the ceremony things without a hitch. The vows he wrote were beautiful and made me tear up. During the reception, however, his best man gave a, I believe, slightly drunk toast where he mentioned my husband using chat GPT to write his vows. Everyone laughed, including me, until he emphasized that it wasn't a joke and that my husband actually did use chat GPT to write them at the last minute, apparently to emphasize how lucky he was to find a wife like me. My husband was laughing nervously and I was taken aback. As soon as the toast roped, I ran to the restroom and cried, feeling extremely hurt that not only did he use ai to write something so intimate but mostly that he presumably would not have told me had this not been revealed during the toast. He followed me to the bathroom and apologized. Basically goes through that this person's whole wedding night was ruined because she realized that not only did her husband use chat gpt to write the vows, but he lied about it and everybody else was planning on lying it too.

I don't know. I think it's somewhat huh.

2:01:39 - Leo Laporte
From the moment our paths crossed, you've brought light and laughter into my life in ways I never imagined possible. Today, standing here with you, I feel the depth of my love for you more than ever. I vow to always cherish your heart, to listen with patience and speak with kindness. I promise to celebrate your victories and lift you up in your moments of doubt. I will honor your dreams and respect our differences, knowing that together we are stronger. I think that's a pretty good wedding.

2:02:14 - Paris Martineau
I'm sure that they were great. I just do think it's a little skeevy to write your wedding vows using Chachapiti and not tell your wife.

2:02:22 - Leo Laporte
You are my home, my adventure and my peace. I love you more deeply than words can express, and I promise to spend the rest of my day showing you just how much this is good.

2:02:35 - Paris Martineau
It's good stuff. You're going to redo your. You could renew your vows with this.

2:02:39 - Leo Laporte
I you know, I don't think we made up our wedding vows, actually, I think. I think. I think the the preacher made him up. Is that which is worse, having the preacher make it up or an? Ai? Jacob ward has returned. He is now on his cell phone, I think.

2:02:56 - Paris Martineau
Yeah, dude this is now you're doing the real jeff jarvis lifestyle, which is there was once that jeff's internet went out and he had. We had to wait like an hour. Then he did a podcast from outside. It was.

2:03:08 - Leo Laporte
It was a nightmare oh, that's right, because he had a big old power. I don't know what's going on.

2:03:13 - Jacob Ward
Yeah, I'm having some kind of internet outage in my neighborhood. I'm'm seeing him on Zoom.

2:03:17 - Benito Gonzalez
That's really terrible, his camera's not coming into the e-cam. I see it, I see it too, but it's not coming into the e-cam.

2:03:22 - Paris Martineau
Is the audio coming in Chat?

2:03:24 - Benito Gonzalez
can you hear him? The audio's coming in, but I see him. He's on the Zoom. I hear you guys.

2:03:28 - Jacob Ward
just fine, he's on the Zoom, my whole. So my, yeah, my whole system has gone down the whole. I don't know if it's a neighborhood thing or what's happening. Anyway, I'm going to try and come back in, but if not, you guys, this has been a real pleasure.

2:03:41 - Leo Laporte
Jacob, where do you live? So I know, are you in the Bay Area or I am?

2:03:45 - Jacob Ward
I'm in the Bay Area. So you know it's the capital of technology. So what the heck you know?

2:03:51 - Leo Laporte
but I it drops out a lot.

I find we actually have Starlink on the roof just in case, and we've had to flip over to it. Jacob, it was so nice to have you on TheRipCurrentcom. If your internet comes back, call us back. Okay, parallel to the shore. All right, I will. Thanks guys, appreciate it. Thank you, jacob. Take care, jacob Ward, everybody. He missed that whole wedding vow thing, though I was gonna say he must have come in very confused paris. It's been like wow, they're finally doing it. I vow to always cherish your heart, to listen with patience and speak with kindness. I think these are, these are as good as anything I would write yeah, I mean I'm sure it's ultimately not a big deal.

2:04:34 - Paris Martineau
I think the thing about this that I would argue the advice replier on this, the advice giver messed up with. She basically said like oh, you're overreacting, not a big deal. Everybody is chat GPT for this stuff. He's maybe not a good writer and I'm like that's not the problem. The problem is he lied to you about it.

2:04:53 - Jacob Ward
Yeah, because you guys both said we're going to you about it.

2:04:55 - Leo Laporte
Yeah, because he was trying to preserve the romance when you guys both said we're going to write these by ourselves. That's well, you're married. That's what you do you preserve the romance.

2:05:03 - Paris Martineau
Yeah, but like maybe you shouldn't start it off moment one of the marriage. Maybe you should wait like a couple of weeks before you start lying left and right.

2:05:12 - Leo Laporte
It could have been a very big bonding moment with you. Know, honey, I just have to tell you something before we get up. There is, I'm a terrible writer and I asked for some help and I think the vows are really meaningful. Have you ever bought wait a minute have you ever bought a greeting card with a pre-written message?

2:05:31 - Paris Martineau
I have yes.

2:05:34 - Leo Laporte
How is this different?

2:05:36 - Paris Martineau
Well, because when you give someone the greeting card, they know that you didn't write it because it's printed and you should go.

2:05:42 - Leo Laporte
I mean this sign paris I actually wrote this.

2:05:47 - Paris Martineau
I got them to print it I made them print it.

2:05:51 - Benito Gonzalez
Um, there's also this weird thing though that people have that like I don't, I would rather like my best friend give me a poorly written speech written by him.

2:05:59 - Paris Martineau
That's the thing it's like I want to hear.

2:06:03 - Leo Laporte
It's really heartfelt yeah.

2:06:04 - Paris Martineau
Like yeah.

2:06:05 - Leo Laporte
But some people can't get their words out in any heartfelt way, you know, I mean.

2:06:12 - Paris Martineau
I feel like the effort. Look who just came back.

2:06:13 - Leo Laporte
Jacob's coming back? Yeah, I'm back. How about that? I know what happened. The internet came back. Yeah, it came back, I don't know why. So I'm going to make a confession. We're sitting around the table with my in-laws and I used ChatGBT to make a picture of us sitting around the table While we're at the lunch. I I found an application that I could send that picture to, and I had the AI in the application write a heartfelt message from Lisa to her parents.

2:06:50 - Paris Martineau
And I mailed it and uh, so you printed it out and then this company does the whole thing. You just do it on my phone while we're at lunch and I said and I pressed send and uh, wait.

2:06:55 - Leo Laporte
So you printed it out and then no, this company does the whole thing. You just, oh, it's like frames I'm doing it on my phone while we're at lunch and I said and I pressed send and uh, it was really good. And uh, last week lisa said um, can I ask you something? My mother was so grateful, she was so touched, so moved by this card.

2:07:14 - Paris Martineau
She got a sentiment did she have to play it off as if she wrote it? You didn't tell her you were sending this lisa's good.

2:07:24 - Leo Laporte
I don't think she said oh mom, I don't know anything about that. I think she played along, but she did that she's so thoughtful.

2:07:30 - Jacob Ward
I know?

2:07:32 - Leo Laporte
well, yeah, it was, it was such BS. It was so untrue, but it made her mom feel good, and that's what really matters. I hope she doesn't listen to the show.

2:07:43 - Paris Martineau
Please, mom, please, don't listen.

2:07:46 - Leo Laporte
I should have brought this up actually when Daniel was here. More than half of the top 100 mental health TikToks contain in misinformation. That's not surprising to me Questionable supplements and quick fix healing methods, but of course it's not surprising. But this is the point is people are going to TikTok to get their mental health advice Probably better to go to an AI.

2:08:10 - Paris Martineau
People are going to TikTok to have somebody tell you that everyone who's ever done you wrong is gaslighting you and that you're actually the smartest person in the world and that all of your opinions are valid. It does tell you that everyone who's ever done you wrong is gaslighting you and that you're actually the smartest person in the world and that all of your opinions are valid. It does tell me that it does, and it's right for you, everybody else, it's right.

2:08:26 - Jacob Ward
You're the special one.

2:08:30 - Paris Martineau
The stripper really likes you and it's really going to work out.

2:08:33 - Leo Laporte
Don't say that she loves me. She said I think you should do. Do you should have chat dbt? Write your vows for her? I think she'd really appreciate it, honey, when I saw your tassels on stage. Uh, let's see. Uh, let's quickly change the subject. Um, uh, oh boy, there's so many good ones and we don't have a whole lot of time left. Oh, they're making a movie out of the open. They are board drama.

2:09:03 - Paris Martineau
No, out of the open air board wait which one from which book?

2:09:06 - Leo Laporte
karen howe's book oh, that's a good question.

2:09:10 - Paris Martineau
That I don't know because that is the reason why I mean the reason why people are writing, yeah that's right that are coming out right now are because, uh, publishers are interested. Who would you cast as sam altman? Are you ready? Again, I'm not the person to this because I know no actors jesse eisenberg, jesse eisenberg, yeah you can't do him too.

2:09:27 - Jacob Ward
You can't do him again.

2:09:29 - Leo Laporte
Better than that? Yeah, that's right, wasn't he? Uh, mark zuckerberg? Nope, better than that. This guy here, andrew Garfield.

2:09:39 - Jacob Ward
Oh, I was just going to say Andrew Garfield, that is a good call Perfect. Sam Altman, that's a good call.

2:09:46 - Leo Laporte
Yep.

2:09:47 - Paris Martineau
Nice.

2:09:53 - Leo Laporte
The movie will be. Let's see.

2:09:54 - Paris Martineau
I'm going to zoom out and tell you some more about this. Let's see. I need to see what book it's based on luca guadagnino is in directs to direct.

2:10:02 - Leo Laporte
Is in talks to direct. Artificial, a dramatization of sam altman's dramatic firing and rehiring at open ai in 2023. Oh, get this, it's an amazon film um I'm curious, why now? Oh, and of course, the oscar-winning yura borisov from onora will be the lead roles he'll be playing. Uh, probably suits giver right simon rich wrote the script.

2:10:36 - Paris Martineau
What's it from those that's?

2:10:38 - Jacob Ward
that's a great question, maybe it's just a script unrelated to a book, because I don't think this is, and this is different than the, this is different than the, than the play wasn't there.

2:10:47 - Paris Martineau
There's a, there was a play in uh in san francisco or maybe it was in san francisco too. There was a play about this that every review I heard of it made it sound inscrutable to the average person. But like it was perfectly targeted.

2:10:59 - Jacob Ward
This is the thing.

2:11:00 - Paris Martineau
Tech reporters that spent every waking minute thinking about that story.

2:11:03 - Leo Laporte
I saw a play and I loved it in New York, but Jeff found all the bad reviews and really dissed it.

2:11:10 - Paris Martineau
But it was his nature.

2:11:13 - Leo Laporte
I saw it on Lincoln Center. It was so good. It was about a Nobel Prize winning novelist who, it turns out, used AI to write the novel, and it starred Robert Downey Jr. Okay, it's called McNeil. I really liked it. I think it sank without a trace. Here is the actress, monica Barbaro, who plays Mina Marotti. Okay, cool.

2:11:34 - Paris Martineau
Yeah, she looks kind of like Mina Marotti.

2:11:35 - Leo Laporte
Okay, cool. Yeah, she looks kind of like Mina Marotti. So we got Garfield, we got Barbaro.

2:11:40 - Paris Martineau
Let me see who else I can tell you that I interviewed Mira Marotti before opening up. Like there was a time right before the pandemic hit, like it was like a week or two before stuff started bubbling over in New York where I got a last minute invite to be Diane von Furstenberg does this like women in charge festival event, like thing that's like women interviewing women for like some audience of like some powerful women and I was asked to interview this AI founder about AI. Like last minute went there the. The real boon of it for me is because the pandemic happened right after they couldn't send a courier to get my DVF very expensive dress back, so I got to keep it, but as part of it- you got a DVF dress for doing it.

I did for free from doing it and they can't be taking that back because that was five years ago. That's mine now, but I was interviewing her and I wrote my questions, as one does. But I go there and they hand me a card. They're like oh, since we asked you last minute, we just wrote some questions for you, and they were all stuff like what it like to be woman in AI. Is it hard to be girl? But tech, femininity, masculinity, technology and so I was like we can't do this.

2:12:52 - Leo Laporte
What's your favorite fit for a board meeting and I was like we can't do this. What's?

2:12:54 - Paris Martineau
your favorite fit for a board meeting and I was like we can't do this. And I was like ask about like I don't know bias and like AI training Stuff. That was on the theme but not Good for you and then listen, respect the hell of Diane von Furstenberg. Really cool person, really smart thinker. My nephew dated her daughter.

2:13:08 - Benito Gonzalez
As.

2:13:12 - Paris Martineau
I'm going through the questions. Diane stands up in the audience, raises her hand.

2:13:22 - Jacob Ward
She's like I've just got to ask what's it like to be a woman in tech? They were her questions.

2:13:26 - Leo Laporte
I wrote those questions and I really wanted you to use them, she answered them.

2:13:31 - Paris Martineau
She did quite good. Mira was very nervous, though Was she smart. Yeah, incredibly, I was like listen, you got this.

2:13:39 - Leo Laporte
It's a kind of interesting story. It's going to be a great movie because, of course, she was the CEO. Interim CEO took over from Sam Altman when his board fired him for five days. But what we didn't know at the time but we have learned from Karen Howe's book is she was actually one of the instrumental people in getting him fired. That ilia suits giver and mira marati teamed up and most of the slack messages that suits giver showed the board were mira marati slack things that sam holman had said to mira marati. And, of course, uh Marotti left and started her own AI startup and has already raised a huge amount of money for it because she's Mina Marotti. Of course. Here's good news and bad news for podcasters. Uh-oh, watch out, jacob, we're in trouble. Google's Notebook LM now lets you share your notebook and AI podcasts publicly. Get ready for an onslaught of notebook lm podcasts, right?

2:14:41 - Jacob Ward
this has totally been like the thing standing between us and and uh, being just washed away no, no, no, the just like, yeah, it's been, like there's been at least a little tiny touch of barrier to entry, and now all that's going away. So you know, between that and all the tiktokers who are but like, yeah, it's been, like there's been at least a little tiny touch of barrier to entry.

2:14:58 - Leo Laporte
And now all that's going away.

2:14:59 - Jacob Ward
So you know, between that and all the tick tockers who are, you know, just crushing, just, you know, crunching out, you know, 100 pieces of content a day, because now you can write those of us who insist on voicing our own stuff, that's it for us.

2:15:13 - Leo Laporte
We, we will have the creators of Notebook LM, the Google guys, on the show. We, uh, we will have the creators of notebook LM, the Google guys, uh, on uh the show. We've had them on before, we'll have them on again. They're friends of uh, of uh, jeff's uh, in the coming weeks I don't see a date yet, but that's coming up. Notebook LM actually is really cool. I don't know if you've used it. That's the one where you upload all your content and you say make me a podcast out of that. That's not the only thing you can do. Actually, it's a very useful tool if you have, I would think, for researching a story, for instance. Upload all this stuff and then have it be able to ask questions of it about the story Seems like that would be pretty useful.

2:15:50 - Benito Gonzalez
I'm curious, though do people actually listen to any of these Notebook LM podcasts Like, is there anybody who actually does this?

2:15:55 - Jacob Ward
like, oh, I do this, I've never, finished one that's what I know, yeah, I know people who use it as a research thing. I don't think people use them as entertainment yet, but uh, you know who does?

2:16:06 - Leo Laporte
satchin adela, the ceo of microsoft, in an interview uh, was it with vanity fair, I think it was he uh said I listen to podcasts. And then he admitted well, I don't actually listen to podcasts.

2:16:17 - Paris Martineau
I feed podcasts into co-pilot and then when I'm driving to work I have co-pilot summarize them and I can ask questions I look through a series of mirrors that are reflected through a pool of water, and then a man interprets them in front of me. That's reflected through three mirrors, and then I watch that, and that's how I can listen to your show many, many times.

2:16:42 - Leo Laporte
Yeah, I'm a long time listener this is I'm going to give you some ammunition, paris, because I'm a generous person. Wow, ai, hallucination cases database. This database attracts legal decisions. You know. You know that lawyers continue to use ai generated pleadings. These are the database databases legal decisions where generative ai produced hallucinated content, typically fake citations that were then submitted to the court 138 so far. Wow, yeah, lawyers can't stop using it. They cannot stop.

2:17:22 - Paris Martineau
I mean, even in the FDA and I think HHS had recently released like guidance documents that had a bunch of yeah, that was a formatting issue.

2:17:32 - Leo Laporte
Yeah, yeah, yeah.

2:17:34 - Paris Martineau
That's yeah.

2:17:34 - Leo Laporte
Robert F Kennedy Jr in his quest to produce a document that made sense. And they apparently write it and it had all sorts of phony studies and all sorts of things. You'll like this database because one of the columns is what was the court's reaction to the fake?

2:17:55 - Jacob Ward
pleading Right. Right, that is the big one. Yeah, that's the big one. Yeah, go for it or not.

2:18:00 - Leo Laporte
There were sanctions occasionally. In one case this was in Israel. Ai use was noted by the lower court but no specific actions was taken. Most of them though warnings right Most of them.

2:18:14 - Paris Martineau
Yeah, my former colleague, Gabby Delvai, who's now at the Verge, wrote a good piece on this kind of like, summarizing a lot of these things, and I don't know she's had so many interesting details in this. There was one attorney who was sanctioned in 2023 for using it like this, but he said that he thought ChatGPT was a quote super search engine and it took filing with fake citations to reveal that it's more, not that.

2:18:41 - Leo Laporte
Right Today in the Washington Post, jeffrey Fowler took five AI bots and gave them tests to see who was the smartest. And actually it was an interesting way to do this. They gave a novel to the AI, a novel called the Jackal's Mistress, and asked the AI to answer questions about the novel. The judge was the author of the novel, which is probably a pretty good judge of. He actually got a permission from the author's guild uh, who is suing ai right now about all of this to to do this, uh. And he said uh, chat gpt did probably all the bots did poorly in literature, but chat GPTT probably did the best. Only Claude got all the facts right about the book. Gemini, which wrote very short responses to our questions, was most often guilty of what the author called inaccurate, misleading and sloppy reading. In one summary, gemini described a man who had just had a leg amputated. It's a Civil War novel appearing on another character's doorstep Leaning, I guess I don't know. He did say and I thought this was interesting that they asked the AIs to describe how the book's epilogue quote made them feel, and the author said both Chappie GBTPT and Claude got appeared to have. Quote all the feels. Their responses express precisely what I was trying to convey.

In law, claude was the best they used a corporate lawyer to. Claude was the best they used a corporate lawyer to. And they gave two common legal contracts to Chet GPT and asked a lawyer to judge the response. Claude won overall by offering the most consistently decent answers to our questions. It did its best work on our most complex request, suggesting changes to our test rental agreement. Uh, the judge the lawyer said Miller's. The Claude's answer was complete, picked up on nuance and laid things out exactly as he would. He said this in this case. It would be a good substitute for a lawyer. Health science Claude was the best. Um, so what was interesting is that the every one of these ais hallucinated. Except for claude, not one hallucination in any of the five tests all right, go off look, I gave you one, now it's my turn.

That's great. We, you know, we go back and forth no, no, go off.

2:21:38 - Paris Martineau
To be clear in uh the kids.

2:21:40 - Leo Laporte
Parlance means like good job, you got it oh, I thought it was like you were saying go off yeah, yeah, I realized.

2:21:47 - Paris Martineau
No, it's like go off, king, like you're go off, king.

2:21:51 - Leo Laporte
If you said king, I would understood. Just I put the fries in the bag. I'm just saying, jacob, we're going to let you go. Thank you so much.

2:22:01 - Jacob Ward
You have a Brazil story before you go. Well, I was just going to say one thing that I thought you guys would be interested in is I just came back from a trip to Brazil and it was so interesting because Was it for work or pleasure?

Well, my wife got to go for work. I went for pleasure, I met her there and we got to spend a week together in Rio. And, as is always true on every vacation, I can't help but offer up my like nerdy, dystopian observations in any place and ruin the vibe. And in this case, what was fascinating about Brazil and this takes us all the way back to our conversation about Palantir is how much Chinese surveillance technology is in the country. Right, they've got this deep relationship with the BRICS nations right.

Brazil, russia, india, china, south Africa and China has been exporting their safe cities like surveillance programs to all these different capital cities Quito, ecuador, karachi, pakistan, all these places. Rio has tons of that. And so when you're walking around in the tourist areas, it's just cameras everywhere and the combination of that and a lot of military police sort of floating around. People locally say it's actually much safer than it used to be and it's so interesting to be in that. But it's not. That's not helping people in the favelas, it's not helping people in the rural areas, necessarily, but there is this bargain being struck there between doing away with privacy and using technology for surveillance. That I just thought was so interesting and I just think anybody, any listener of this show or you know, and you guys especially like if you ever get a chance to go to that country and just see this like alternative universe of it's a democracy. And they also had a president who tried to inspire a kind of Bolsonaro.

2:23:50 - Leo Laporte
Yeah.

2:23:50 - Jacob Ward
Yeah, bolsonaro, he's now on trial for that. There's a really robust misinformation problem, but there's also a really robust journalism world there, and then this world of chinese surveillance everywhere. It's literally like going to an alternative timeline. It is really cool and interesting, so anyway highly recommend I'm jealous did you go to ipanema?

2:24:12 - Leo Laporte
oh yeah, walk down the beach.

2:24:14 - Jacob Ward
We did the whole thing. We did the beach. We went to a soccer game, which was a crazy experience. Oh man, we got to see one neighborhood against another neighborhood. It was like watching yankee's red socks. Oh, it was crazy oh, I bet that was crazy you know you're like cool and you think it's like a national championship and it's literally like this neighborhood, two blocks, yeah, yeah, crazy. So anyway, highly, thank you, jacob warp the I really appreciate.

2:24:39 - Leo Laporte
It's always great to talk to you. Come back soon, sure will, thanks. You guys appreciate it. Thank you, jacob. All right, yeah, I didn't realize you had just been on his podcast.

2:24:47 - Paris Martineau
That's hysterical all right, I got one for podcast. I think the week after he was last on our podcast, but it was recently released it was delightful.

2:24:55 - Leo Laporte
He's, he's. I just love jacob. I really love jacob.

2:24:59 - Paris Martineau
Yeah, I love his fun background that he has.

2:25:01 - Leo Laporte
Yeah, it's great uh story from the wall street journal that might I'm not sure, but it might have something to do with one of our uh club members. Um, let me see, let me pull this up here the wall street wall street journal.

There we go. Uh, morgan stanley you've heard of them. Legacy code nine million lines of legacy code, to be exact launched their own devgenai and ai tool, built on the gpt models, to translate the code into modern languages. They say it was a success. It it saved developers an estimated 280,000 hours. The tool translates code into English specs, enabling any developer to rewrite it in a modern language. So it didn't write the code exactly. Darren, did you have anything to do with that? I think Darren works at Morgan in AI, so I'm just wondering if he had anything to do.

2:26:19 - Paris Martineau
So we assume that you have to be responsible for everything we read? Oh that's easy Not is he not more?

2:26:25 - Leo Laporte
Isn't Morgan Stanley JP? Isn't that related?

2:26:28 - Paris Martineau
No JP Morgan in the U S 15 years ago.

2:26:31 - Leo Laporte
Oh you left 15 years ago.

2:26:34 - Paris Martineau
Well when you like to be there now. Yeah, they're doing your favorite thing see, this is the reason why you listeners need to get in club twit, because then you can have leo be spreading misinformation about your job.

2:26:49 - Leo Laporte
I remember telling stories about morgan stanley. I just assumed they were current. I didn't know. They were really old, so probably what they were doing is taking your old code, darren, and rewriting it in a modern language. That's what was going on. It's interesting.

2:27:06 - Paris Martineau
They were like we've got to update this stuff.

2:27:07 - Leo Laporte
This is an example anyway of not exactly vibe coding but the value of AI in analyzing code bases. And you know I'm not surprised by this because it's a computer. It's talking to another. It's like of course a computer can understand computer code. That makes sense to me, anyway, it's interesting.

2:27:27 - Paris Martineau
I mean they say in the article, when it comes to full translation, the technology still has some room to mature, which is of course true, like it can technically rewrite code from an old language like Pearl in a new one like Python, but it wouldn't necessarily know how to write it as efficient code that takes advantage of all of Python's capabilities, and that's one big reason humans are staying in the loop. The guy from Morgan Stanley said, which I mean, I'm sure eventually that's intelligent.

2:27:53 - Leo Laporte
Yeah, that is intelligent Keep humans in the loop. I think you, I think you're always gonna have to do that. We are gonna do a little vibe coding session on a friday I think darren will join us? I don't know. I've been trying to decide. One of the things I've been doing with claude code, which is my current favorite tool that one of harper reads recommendations is uh, solve the advent of code problems from last year that I got stuck on.

2:28:18 - Paris Martineau
And so far isn't the part of participating in advent of code. Have a sense of pride?

2:28:23 - Leo Laporte
No absolutely, absolutely. I mean, I it was more to see if it could do it then, because that's how I'd want to do it, although Darren, who is very good at this. We have a number of people in our club who take this coding challenge every December. It's an advent calendar, a different puzzle every day. We all do it together. It's fun. But Darren said a couple of days ago I think this is going to take all the fun out of it because everybody's going to be able to solve them with AI instantly.

2:28:55 - Paris Martineau
Yeah, and isn't part of this whole thing, like you've got a leaderboard of how fast people do. I remember you guys showing me this.

2:29:01 - Leo Laporte
It was already problematic this year because there was somebody using AI who solved it in seconds and it's like well that you got to take that off the top. I think they'll probably solve this by having an AI division and a human division, or like a vibe coding division and a hand coding division. I you know, here's my attitude towards that. There will always be people who do woodworking because they enjoy the process. Doesn't mean your furniture is going to be made by a little old man in North Carolina.

2:29:31 - Paris Martineau
Well, no, because you want your cheap furniture to be shipped to your house in two days From.

2:29:36 - Leo Laporte
China, of course you do, For you know, $25.

2:29:38 - Paris Martineau
Right.

2:29:40 - Leo Laporte
But there will always be people who will like to do that and I will be the kind of person who will hand code. There's no point in using a chatbot to do it. Anyway, we're going to play with it. I hope we get some interesting things. I like to imagine you out there with a chisel and a little hammer coating your little heart out. You know what. You say that and I know you're mocking me. It's okay, I'm not mocking. Yes, I know.

2:30:02 - Paris Martineau
Brassing is my love language.

2:30:03 - Leo Laporte
It's exactly how I do it. To me it's like woodwork, it's like I'm building a little.

2:30:10 - Paris Martineau
No, I mean, I think that earnestly.

2:30:13 - Leo Laporte
No, it's really true. Like yeah we were talking we were talking before uh at the beginning of the show, about reddit and uh, how they analyze the mi, the a-hole reddit subreddit. Reddit has sued anthropic because they say anthropic accessed their site more than a hundred thousand times after saying it had stopped. Because, you see, reddit has a deal with the uh, the competitors, with open ai you can't be having anthropic do that?

2:30:44 - Paris Martineau
that's not fair at all, and reddit has its own kind of internal. Do you know what reddit's internal chat bot is powered by? Is it an open ai?

2:30:52 - Leo Laporte
oh, that's a good question. I I do not know.

2:30:54 - Paris Martineau
For anyone who doesn't know, on Reddit, now you can, I think, on both the app as well as the website. Instead of going to your various subreddits or using the search bar, you can basically search into a little chat bot window about your question and then it gives you responses, I guess powered by Reddit queries that then have little links out. I found it not particularly useful because when I'm looking for something on Reddit I usually actually want to read real people's responses, not just Reddit summary, but occasionally One of the reasons this exists is because Google it's been long held that the best way to get an answer from Google is to type site colon redditcom.

Yeah, that is true.

2:31:37 - Leo Laporte
So I'm going to ask it this is Reddit answers what's the best AI for writing wedding vows? Let's see if it knows.

2:31:45 - Paris Martineau
I mean that's an interesting query because I feel like probably what this is useful is. I like that, I like the extent of all these.

2:31:53 - Leo Laporte
Oh, there's an actual yeah, of course, of course, wedspeechai.

2:31:57 - Paris Martineau
Of course, of course.

2:31:59 - Leo Laporte
Oh my God. And somebody says it didn't just throw out cheesy one-liners. It asked real questions about how I knew him, what kind of tone I wanted, what stories I might include.

2:32:08 - Paris Martineau
Okay, it's directed that one looks like it's directed, so it's not likeTurfing going on, and this is also one of the reasons that I believe a new announcement that I have somewhere on the rundown from Reddit this week is that they're now going to allow profiles the ability to turn off Normally if, say, you see that comment come up being like, actually this wedding site is perfect and good and wonderful and you should use it. I'm a real person.

You could go click their profile, go read all of their comments and see if maybe the only thing they've ever commented about is how this wedding site is really good and you should use it. But now Reddit has a feature where you can turn off history on your profile, which will make it really hard to tell when someone's astroturfing or uh just karma farming.

2:33:01 - Leo Laporte
I think it's wait a minute. What's karma farming and astroturfing?

2:33:04 - Paris Martineau
oh good, okay, yeah, that's a show title for you, so karma farming is um, whenever you'll see, it'll often be with like reposting. You'll see, someone will uh post something on like uh, mildly interesting about like a uh, let's say like a safe they found buried in their ground, which is huge reddit bait, and then everybody upvotes because they want to see what's inside the safe. The person never responds and someone reverse image, searches that safe photo and they're like oh, somebody posted this on reddit like four years ago and actually it was their safe and here's what happened with it. This person's just taking that post in order to get the reddit karma.

2:33:41 - Leo Laporte
The upvotes so they have a higher value reddit account can I just say that was an oddly specific anecdote, paris, you have any experience with anything like that I know, I just I realized recently I am a compulsive reddit user and it's really embarrassing for me.

2:33:57 - Paris Martineau
It's my social network, okay, so now they have the little uh awards on your profile.

2:34:03 - Leo Laporte
Oh yeah, I don't want some gold. I'll give you some gold.

2:34:05 - Paris Martineau
Listen, I would like gold cause. I think that's fun, but I I got an award the other day. That was like 365 days, no, like the internal Reddit achievements. It was like you've logged into Reddit, you've opened the Reddit app 365 days in a row and I was like I don't want to know this Reddit. Don't tell me that.

2:34:26 - Leo Laporte
Do you have a custom Reddit avatar?

2:34:29 - Paris Martineau
Only for my work account. I don't entirely believe in personalizing a Reddit account if I want it to be anonymous.

2:34:36 - Leo Laporte
That's a good point and, by the way, they're expensive $10?.

2:34:41 - Paris Martineau
Oh yeah, this is their whole NFT sort of thing.

2:34:44 - Benito Gonzalez
I'm not doing that.

2:34:46 - Daniel Oberhaus
I can create one.

2:34:47 - Leo Laporte
Yeah, like a little. Avi, like that I do. That's me right there, you see that that's you. I'm not sure why I'm carrying a cleaning bottle there, you see that that's the one.

2:35:00 - Paris Martineau
I'm not sure why I'm carrying a cleaning bottle. Why are you wearing a?

2:35:02 - Leo Laporte
tanuki hat, I don't know, I just do you're carrying where do you see your awards? Where would I see, oh here's my achievements let's see what I got I'll tell you things about yourself I'm a banana aficionado you're a banana baby I've been all the bananas. What does that mean?

2:35:15 - Paris Martineau
that's the amount of scrolling you do, oh I'm a big time banana.

2:35:20 - Leo Laporte
Cool comment. Okay, I've made a cool comment.

2:35:24 - Paris Martineau
I'm a hometown hero oh, you've got a five-day streak.

2:35:28 - Leo Laporte
How cute oh, what was yours?

2:35:31 - Paris Martineau
I'll look it up. It's gonna be something grand.

2:35:33 - Leo Laporte
I am an elder I'm an elder elder, that's good.

2:35:36 - Paris Martineau
That's true.

2:35:40 - Leo Laporte
They are accurate in that.

2:35:41 - Paris Martineau
Please Three years anywhere is impressive 391 day streak. Oh, you got me beat. Okay, it's just because whenever I'm this is even sadder. Actually, whenever I'm going to sleep at night, I'll like scroll through some subreddits to bore me to go to sleep, and it's just a very relaxing nighttime habit that I have.

2:36:00 - Leo Laporte
See, I only follow interesting subreddits. What's a boring subreddit?

2:36:03 - Paris Martineau
They're interesting, but then the more you get into it, the more your mind kind of glazes over. Enjoy gme meltdown, which is a like uh subreddit chronicling all of the meme stock. Uh oh, games in this case is the term used for people who are continuing to hold the bag for game stop, bed bath and beyond amc other things like that but it chronicles all the different kind of characters in this universe and it's absurdly detailed and ultimately meaningless.

But it's the perfect thing to kind of glaze your eyes over as you fall into a deep, dark and dreamless slumber.

2:36:46 - Leo Laporte
By the way, this web speech is real. They have written 1794 successful wedding speeches not just vows, but speeches. You can create a speech. This is good. Actually, that's the worst thing to be a best man and have to write the speech.

2:37:07 - Paris Martineau
This would be good, I wonder if yeah, I had to do that oh man 15 bucks.

2:37:10 - Leo Laporte
What For an 800 word speech. I just did it on the airplane flight to my friend's wedding. Did you have to give the best? What do they call it? Made of honor speech?

2:37:22 - Paris Martineau
I did I crushed it, made people laugh Made people cry.

2:37:25 - Leo Laporte
Did you tell anecdotes? I had a little improv.

2:37:27 - Jacob Ward
I did.

2:37:30 - Paris Martineau
I told anecdotes from more recent times. But then I also it was the sort of thing where I was, I don't know, worried about it, just because it was like a friend I'd had since like childhood growing up, and obviously we've been close ever since. But, like over the last couple of years she's lived in a totally different side of the country, so I wouldn't say like I'm the closest with her and her now husband, but I was like dang, how am I going to get in this? But made it work, it was good.

And then midway through the wedding, right before speech times coming up, my parents were there because it was like a hometown friend wedding and my mother, bless her heart, always doing journalism for me, comes up. She's like Paris. I've been talking to Jessica, my friend's, like her family, I've got great news for you. I'm like what she's like. I've checked with multiple sources, my friend, she lost both of her parents in college and she's like during, like right after the vows concluded, after the ceremony, jessica's aunt got two phone calls from Tim, jessica's late father. It somehow showed up on her phone and she's like shown it to other people. It's like I don't know how someone with Tim's number is calling or right now, and she was like I've got multiple sources that witnessed it. You should reference this in your speech and I did, and people loved that.

2:38:43 - Leo Laporte
And I was like that's good on mom for coming up with that juicy tidbit.

2:38:47 - Paris Martineau
Listen, I was like you got to have people doing the ground. I was like you got to have groundwork going on for the spook, for the spooky vibes to incorporate in it. Fabio martineau was there. The chat asked he was. He was there who?

2:39:02 - Leo Laporte
oh, fabio martineau. My father was there. Yeah, looking good too. I saw the pictures. It's true. We're gonna take a little break. Final words and our picks of the week coming up.

Before we do that, though, I do want to give a little plug for our club. Yes, we raised the price. We've had the same price for four years, the whole time the club's been around and, of course, every one of you who are already club members at that price will keep that price. That's our promise to you. But we hope we can get some new members at the ten dollar a month price. I still think it's a good deal. You get an awful lot for that. Access, of course, to our discord. That's where we're going to do the ai user group on friday at 2 pm, pacific, and it's where we're going to do our streams of keynotes from now on, so we avoid getting taken down. Monday, micah sargent and I are bringing our lunch boxes because we're going to do the apple wwdc keynote at 10 am and then at 1 pm we're going to do the State of the Union keynote. We've never done that before, but hey, you know, we're there, we're going to stick around, and one of the good things about doing these events in the Club.

Twit Discord is you can participate. We want to get your comments as well. Stacey's Book Club's coming up. We've got a photography episode with Chris Marquart. You still have about a week to take a geometric photos photo to submit Micah's Crafting Corner.

You also get ad free versions of all the shows. You get animated gifts with all the prices. You get access to some of the fun best people like Darren and Dr Dew and Wad Fan and all the folks in the discord and, by the way, it's not just during the shows. They they're chatting and doing things, talking about the things that geeks care about, all hours of the day and night, cause we have people from all over the world.

The most important reason to join the club, though of course it supports us about 25% of our operating costs come from the club. Without that, we operating costs come from the club. Without that, we'd have to cut back on shows, cut back on people. We don't want to do that. So if you want to hear more of what you hear here, if you like what we're doing, if you want to support it, twittv slash club twit, and we thank you so much for supporting everything we do. We really appreciate it. Well, jeff isn't here, but he did leave us, uh a pick of the week. He left us two, one from himself. Most new cars in norway are evs, and actually that's starting to be true all over uh. Europe. Not teslas, interestingly, uh, but evs for sure, and I'm an EV fan.

2:41:35 - Paris Martineau
It is very interesting from Norway, given how cold it is that you would, I'd assume, have like more pronounced range anxiety.

2:41:45 - Leo Laporte
Yeah, I mean you get less mileage when it's cold. But you know, range anxiety is over.

2:41:51 - Paris Martineau
I think overhyped Norway is also a very small place. I guess, you can't really drive that far. No offense, Norway.

2:41:57 - Leo Laporte
You know there are people who have a lot of driving to do, but most people don't drive more than 25 miles to work and back, so there are very few cars that wouldn't do that. And especially since you fill it up every night if you want, if you have a charger in your house, and it's not like the climate's friendly to gas cars either.

2:42:17 - Paris Martineau
Yeah, that's a good point yeah it's just I will say, when I was there, like two or three years ago, I did notice that a significant portion of the cars were evs yeah, I mean, actually it might be better because there's more liquids.

2:42:26 - Benito Gonzalez
There's more fluids in a gas car, so that's all the stuff that gets frozen, right that's a good point.

2:42:31 - Leo Laporte
Yeah and uh. Jeff's son, jake, wanted us to know that david cope, the father of ai music, passed away not the ai music that we're used to, you know the, the suno beats. He created an algorithm in the 80s called the emi algorithm. But what's interesting it it? It's spawned the same kind of questions about whether computers making music are a good thing or not. His, uh, his algorithm was able to make music in the style of bach and mozart and other classical masters. David cope passed away on may 4th at the age of 83. He created and even created the term generative music. He figured out how to program a computer to write classical music. The reason he did it he was commissioned to compose an opera and was struggling with writer's block.

2:43:27 - Paris Martineau
He was desperate to find a compositional partner. New York Times writes he found one in a floppy disk. Isn't that amazing.

2:43:34 - Leo Laporte
Yeah, I haven't. I see if we could find the music. It'd be interesting if we could find it. I don't know if they have it on this New York Times. I guess not.

2:43:43 - Paris Martineau
Oh, there is, If you scroll down. Emi composes, in the style of Mozart, an excerpt from a sonata.

2:43:49 - Benito Gonzalez
I wonder, can I play it? Depends who owns it yeah.

2:43:59 - Leo Laporte
The computer owns it. Let me let me here it is. This is supposedly Mozart, made by an AI in the 80s, Although it says it's performed by.

2:44:10 - Benito Gonzalez
Oh no, it was composed by AI, but a person still played it.

2:44:15 - Paris Martineau
Oh, so probably they own it somehow.

2:44:19 - Benito Gonzalez
I'm not hearing it, leo, I think you're.

2:44:22 - Leo Laporte
Oh, I have to turn the original sound on. No, sorry, that's terrible oh.

2:44:35 - Paris Martineau
I think it seems haunting.

2:44:37 - Leo Laporte
It's haunting because it's like off.

2:44:40 - Paris Martineau
I mean, but that's I feel like the tenor of it is it or tenor is the wrong word. Is it like a minor key or something?

2:44:46 - Leo Laporte
Yeah, it's very minor. Yeah, Huh, I have a pick, because I realized Jeff wasn't going to be here, so I should throw a pick of my own in. Do you ever, you know, is it time for you to go out? And you're running and you just don't feel like running it.

2:45:04 - Paris Martineau
Yeah, but then what will my friends think?

2:45:06 - Leo Laporte
Oh, that's a good question. That's why there's fakemyrun. Fake my run creates fake running routes. It allows you to draw custom realistic running routes. You can draw them anywhere and upload them to your strava or your running application. And why then people think you actually worked out today?

2:45:26 - Paris Martineau
oh, I love this isn't this brilliant? They should use this to seed fake. You know? Remember how um strava data was used to dox a military base somewhere, and I think the middle east yes someone should do a reverse version of that, like a um a uh alternative influence campaign of fake military people running in fake places I'm gonna run around paris and uh, look, I went out to the cemetery, I'm crossing the pont neuf back again.

2:45:56 - Leo Laporte
There you go. Nice little run of uh 10 kilometers took me an hour I like that you can include heart rate data yeah, let's, oh yeah, we want to change the pace. There we go, there we go, and oh, that's so fun yeah, yeah, isn't that good. Let's just get the pace up and down here.

2:46:14 - Benito Gonzalez
Oh, I don't know how to do that the sad part is that this brings into question all those people who like draw pictures when they're on their runs, you know yeah include heart rate yeah, it's gotta be, this is all very realistic and I'm gonna call it.

2:46:29 - Leo Laporte
I love running in paris, france. All right, go. Oh, wait a minute.

2:46:38 - Paris Martineau
You're going to pony up if you want to fake that run.

2:46:40 - Leo Laporte
I got to pony up at six bucks for ten tokens. Well, nothing's free.

2:46:44 - Paris Martineau
I'm just imagining being like well, I'm too lazy to go out so I'm going to pay six dollars to make people believe I was Pick of the week. Paris, my pick of the week is a very classic Parisis pick. I was at my weird mall, uh goth bodega cinema, this weekend and they started something called hot roach summer, which is their uh summer scent like uh screening series. That's all roach themed.

2:47:12 - Leo Laporte
Kind of horror god, and there's more than one movie with roaches.

2:47:16 - Paris Martineau
There's many.

And this was a double feature with something called Roach and the Nest. But the one I really want to highlight is called the Nest, which if you play this Vimeo thing you can see some clips from it. It is a fantastic 1988, like B movie, horror movie with like incredible, like phantasmagorical SFX stuff. They have the craziest roach monsters. It's basically a small New England island town. Everything's good and sleepy and quiet until all of a sudden people start noticing more and more roaches and they realize, oh, could it that, uh, that mysterious company that's moved in to do development?

oh, and it's if you love a like a b horror movie with like really interesting practical effects, I would recommend the nest. It's available on 2b for free right now.

2:48:11 - Leo Laporte
Um, it's basically it's not for the squeamish it's what stranger things was taken after.

2:48:18 - Paris Martineau
It was inspired by oh look, basically, yeah, at one point the roaches light a man's house on fire look at it gleefully tweezing its twisters.

2:48:30 - Leo Laporte
Yeah, yeah, yeah, oh man roaches.

2:48:33 - Paris Martineau
They do a lot. There are roach zombies. It's a fantastic time.

2:48:39 - Benito Gonzalez
It's only horror movies. It's only horror movies. So no Joe's Apartment so far. No Joe's Apartment basically.

2:48:46 - Paris Martineau
That's a great. So the movies in the series are Roach, which was a small like dystopian, a short, dystopian kind of short film that was honestly in 3d. I went in and I got the actual 3d paper. It felt very retro. Then the nest, which is the one I just described, which is phenomenal, and when I haven't seen yet. It's called bug. It's a 1975, uh film. It says the final project behind B movie hype man William Castle, as well as one of the few he's given screenwriting credit. Bug is a gnarly under-the-radar entry in the natural horror canon. Bradford Dillman plays a level-headed professor whose logical world is set ablaze by a new species of cockroach with pyrotechnic abilities, building to a showdown between rational science and the cosmic unknown. Okay, I will be seeing this and giving you guys my full review last year spectacle did rat summer yeah, it's from the sicko.

It's billed as roach summer from the sickos. That brought you rat summer and that's why I knew I had to go see it what's next spider summer all right paris, I my nightmares on you this time.

2:49:54 - Leo Laporte
Thank you so much, paris Martineau, appreciate it. Always great to see you, my dear, always great to be here.

2:50:01 - Paris Martineau
Yes, and thanks again to Even we hold down the fort sans Jeff, sans Jake.

2:50:05 - Leo Laporte
Yeah, that was hard. Yeah, just the two of us. Wow, I'm glad I had those wedding vests.

We almost got married so what happens Thanks to Jacob Ward, author of the loop, and of course, rip. The rip current is his uh. A newsletter and podcast, the rip currentcom, and a certain paris martineau appeared last week, so you can listen to that there. At the rip currentcom we do intelligent machines every wednesday, 11 am, pacific 2 pm eastern time, 1800 utc. Next week, our guest will be stephen witt, author of the thinking machine. That should be interesting. Yeah, uh, I'm not sure what it's about, uh, but we'll find out are you?

gonna read this one. Oh, it's the history of nvidia. Yeah, yeah, I can't. You know what I? I don't understand how this computer thing works. Well, they never send me books, they send me.

2:50:58 - Paris Martineau
I do think that if you don't read the books, you shouldn't tell the guests immediately that you haven't read the book. I like to be honest, I do you would get along well with that guy who's chat GPT to write his vows and then told his wife that was a mistake anyway.

2:51:15 - Leo Laporte
Yeah, the thinking machine is the story of nvidia should be very interesting. It's actually quite an interesting story. Uh, stephen would be our guest. I hope you will join us next wednesday, 11 am pacific, 2 pm eastern. You can watch us live, as I mentioned, on eight different platforms, but, of course, after the fact, you can get a copy of all the shows at twittv I am. We also have a YouTube channel dedicated to intelligent machines. Best way to get the show would be subscribe, though, and your favorite podcast client. There's audio and video of every episode. They're free and they're wonderful and they're long as hell.

2:51:50 - Paris Martineau
Also, if you're out there, you're listening to this. You've waited to the end of the credits. Go out there. Go on Apple Podcasts, go on Google Podcasts. Leave us a five-star review and write something fun and listen. If you write something fun in the review and it makes me laugh, I'll read it on the show. So try your best People have been review bombing us and we got to get that up to a respectable 4.5 and above.

So do us a solid and the more people that say they want the craig newmark jingle there have been some entrants that have said no craig newmark jingle and I want their responses to be no. They've said not. I will think this is something I'll go into more when jeff's here. They haven't said no jingle, they're just saying there should be a different custom jingle for jeff. Okay, that's fair.

2:52:33 - Leo Laporte
That's fair we gotta, we gotta, get the choir on it.

2:52:36 - Paris Martineau
Let us know what weird things do you want on the show. What thoughts do you have about what we're doing?

2:52:41 - Leo Laporte
We're trying to think of some segments we could do on the show Fun little you know.

2:52:46 - Paris Martineau
Yeah, if you've got a segment idea, if you've got vows for Leo to use with his wife and to priests, put them there in the reviews for Apple or Google Podcasts.

2:53:01 - Leo Laporte
We might read them. Thank you, everybody for joining us.

2:53:04 - Paris Martineau
We'll see you next time on intelligent machines. Bye-bye.