Intelligent Machines 831 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 and Paris are here. Great interview. To kick things off, Vlad Prelevets is here. Founder of Kagi, the best search site ever. It beats Google all hollow. We'll also talk about the controversy over perplexity versus cloud flare, AI flight pricing and why it may cost you a lot more next time you fly, Plus an amazing new video generator from Google. That and a whole lot more coming up next on Intelligent Machines Podcasts you love. From people you trust. This is Twit.

This is Intelligent Machines, episode 831, recorded Wednesday, august 6th 2025. Nine seconds of Google. It's time for Intelligent Machines, the show where we talk about the latest in AI, robotics and the smart doodads and doohickeys surrounding you. Careful, there's one over there. This is always a pleasure to do this show because I get together with two of my favorite people.

0:01:08 - Paris Martineau
Paris martineau is here, investigative oh my gosh, there was an intelligent machine right underneath me it's everywhere, they're everywhere.

0:01:15 - Leo Laporte
She is a investigative journalist. That I am happy to say. Consumer reports. Parisnyc is her website. Hello paris. Is that a new painting over the fireplace, over the man?

0:01:25 - Paris Martineau
it's not, but I my office situation is kind of in flux right now, so the camera is probably we're gonna need an update later, it looks we're going to. Whenever the new desk gets here, I'll send you guys oh, you did.

0:01:37 - Leo Laporte
We shame you last week with your little red desk.

0:01:40 - Paris Martineau
Well, no, I love my little little red desk. I want to continue using it, but you are correct, it needs to be larger and the time is finally now. I've made some bold offers on an Eames desk on Facebook. Oh, it'll hopefully pay off.

0:01:56 - Leo Laporte
Well, next time you're visiting Petaluma you know, the new Eames Museum will be opening here in a few years, so if you're an Eames fan, this is is the place to be. Also with us. Jeff jarvis, emeritus professor of journalistic innovation at the craig newmark graduate school, author of the gutenberg parenthesis magazine, now in audiobooks and, of course, the web.

0:02:22 - Jeff Jarvis
I'm going to lunch with craig tomorrow and I would have loved to have gone to salt hanks, but I don't think there's any chance we'll get in he's just too popular folks he really is.

0:02:30 - Leo Laporte
New york times gave him a rave on friday which didn't help the lines amazing rave I want to rave about my search engine as you do often, you, you, you really do yes, I, uh.

I remember in the late 90s and there maybe even earlier than that, I used to work with uh computer columnist john c dvorak, who used to have a litmus test for geeks. He'd say what, what internet browser to use? I mean search engine to use, and if they said excite, they said yahoo. He would you say it's google. And I think when we for everybody has that experience when they first use google of like this is a revelation. This is how search should be.

But in the intervening years, a sad to say and has taken over google's search results, the latest being their ai has taken over Google's search results, the latest being their AI summaries, which are often wrong, often misleading, and something most people just don't want. I've gone on a search you've heard me talk about it for a better search engine for a long time. I've tried them all DuckDuckGo, ecosia, and I even paid for Neva back in the day, until Neva threw in the towel saying you can't compete against Google in search. Then I found something called Kagi K-A-G-I. Kagi had a crazy proposition. You want good search, you want no ads, you want reliable results. Pay for it. Pay for it, but what for it? Pay for it, but what. So? I've been paying for Kagi Search for a couple of years now and I must say it's worth every penny.

Joining us right now. The man who started Kagi with his own money, vrad Prelovac, who is the CEO and founder of Kagi, vlad, it's such a pleasure to meet you and have you on the show. Thank for joining us. Thank you, honor to be here. You self-funded.

0:04:28 - Vlad Prelovac
You put three million dollars in at the beginning to bootstrap it yeah, it turns out that um every we see in silicon valley thought it was a crazy idea and, uh, nobody would give you money.

in other words, yeah, I did what, what, what one's supposed to do when they live in Silicon Valley. When you have a crazy outlandish idea, you go to VCs and see if somebody would fund it. It turns out that they really thought it was crazy, and uh. So I ended up putting basically all of my money into the, into the project, because I so desperately needed something that I can rely on for my search and also that my kids can use without being surveilled and profiled all the time.

0:05:12 - Leo Laporte
So this is how kagi um was born in 2018, no less, which, yeah, I can kind of understand the vc saying what are you talking? No one's gonna ever beat Google. Uh, partly because Google gives a lot of money to companies like Apple, samsung uh Mozilla to make sure that they're the default search. In fact, good luck trying to change your search on your iPhone to Kagi, they don't. It's not an entry.

0:05:45 - Vlad Prelovac
Yeah, pretty much. So. The main difference is that I never had a plan to kill Google or dominate the world. The plan was always only to create a search engine that you can trust, that will have your best interest in mind and that can align incentives to a paid business model, meaning people pay for the search results. If they're not good enough, they walk away with their wallet. So that creates this positive feedback loop that created a lot of innovation, I would say, in Kagi as a product in search experience and just is a, you know, wonderful thing to use, knowing that a tool that you need and rely on so much through the day is designed with your best interest in mind. And that's some random advertisers.

0:06:35 - Leo Laporte
You're a public benefit company, a corporation, as of last year, so that means that it's not a nonprofit, but it means that you have to. There are some rules regarding how much money you can make. Does it also mean you can't sell out?

0:06:54 - Vlad Prelovac
No, it means that we are not only beholden to our shareholders but also to our public benefit statement, to a public benefit statement. So, you know, normally corporations in the US are, you know, at will of the shareholders and optimized for shareholder value. A public benefit corporation also can take into account their public benefit mission. So that makes a great deal of difference in terms of how we think strategically, how we make decisions, how we, you know, create initiatives.

0:07:30 - Leo Laporte
The uh. The public benefit statement is, quote Kagi is committed to creating a more human centric and sustainable web that benefits individuals, communities and society as a whole, with a transparent business model that aligns the incentives of everyone involved. So now you're kind of bound to live up to that.

0:07:50 - Vlad Prelovac
Yes, which was the goal from day one. So I only learned about PBC concept last year, so this is when we decided to transform to a PBC. But it is basically a structure that allows companies like like ours to to really emphasize their mission.

0:08:07 - Leo Laporte
Are you profitable?

0:08:09 - Vlad Prelovac
As of this time, we are breakeven profitable, so Is that?

0:08:14 - Leo Laporte
enough for you? Is that sufficient? Are you happy with that?

0:08:17 - Vlad Prelovac
Sure, I'm, I'm, I'm beyond happy.

0:08:19 - YouTube
So we have a decent search engine.

0:08:22 - Vlad Prelovac
We are, you know, providing stability for 50 or so employees, and you know we have the world's best search engine. So that's, you know, a lot to be happy about.

0:08:36 - Leo Laporte
One of your most recent employees is a good friend of the network. Mike Elgin's son, kevin Elgin, has joined you on the team. Now Kagi's not free. I mentioned that there is a free tier which gives you 100 searches. I think I started with the $5 a month 300 searches a month and then I just skipped right to the 25 bucks a month, Not so much because I needed the number of searches because you can get unlimited searches for 10 bucks a month but because I really wanted to support you. I really believed in what you were doing and I thought if anything is going to succeed against Google, it's going to succeed because we, its users, want it to and support it that way.

0:09:17 - Jeff Jarvis
Tell more about the assistant that you've got at that level.

0:09:21 - Leo Laporte
Yeah, so that's actually that's a little bit controversial. I'm going to uh mention uh real quickly because it's, it's, it's. The timing couldn't be better. If you go to ours technica, the front page right now has lee hutchinson's uh article. Enough is enough.

I dumped google's worsening search for kagi. I like how the search engine is the product instead of me and, uh, he. If you haven't, even if you use Kagi, you should probably read it. I actually learned about some amazing things Kagi does. That's very, very cool. You can really control your search. You don't need to, but you can totally control the search results. But a little controversial because one of the things people are moving to Kagi for is the fact that this Google AI summary that they're foisting on us a lot of people dislike. In fact, lee refers to that. He says you know I'm getting incorrect information from Google at this point in this search summary and I don't like that. So he, I think a lot of people say I'd like to go somewhere without AI. What is your plan with AI? What are you doing with the AI?

0:10:35 - Vlad Prelovac
Yeah, I mean, that's a great question. First of all, ai has been part of search for decades, and we are talking here about generative AI, the latest breed, the LLMs, which have been controversial mostly because of their tendency to hallucinate, so using them in information consumption scenarios is not really optimal all the time. And from day one when LLMs became a thing in late 2022, we had this thing called the AI integration philosophy. It basically said we're going to treat AI as a tool and it's going to be on demand only, so we're not going to push it down user throats. And if people want to use it, they certainly can. And that means that when you use Kagi, the search engine, it's going to work with like any other search engine Well, I guess not anymore, but you're going to get search results. But then we built some clever little things, like if you add a question mark to the end of your query, you can get an AI summary of the results, which is very useful. That's cool, all right? Yeah, you just get a very quick summary.

0:11:53 - Leo Laporte
You don't get it unless you put the question mark.

0:11:55 - Vlad Prelovac
Yeah, you have to put a question mark in and then you get this little AI summary, so it's pretty useful. I use it sort of 50% of the time. It's always providing references and sources and citations. And there are also some people who want to use these LLMs and just talk directly to them. So we build a product around that called Kagi Assistant, which basically gives you access to all the top LLMs in the world grounded in Kagi Search. So you get benefit of both accessing all the models out there and the accuracy of Kagi Search results, and you can do a lot of cool things with it. My favorite one is I created a custom assistant called Athletic Coach, so you know, whenever I run, do my cycling, whatever you know, I dumped the screenshot from Strava into it and you know I asked to scrutinize my session and things like that.

0:12:56 - Leo Laporte
So there's definitely models you can choose from, by the way, including Google's own models. I mean, this is, this is a great way to access all of these incredible models. Qn is in there, DeepSeek Grok, of course, and it looks like you keep up with the latest.

0:13:10 - Vlad Prelovac
Yeah, we pretty much do update this almost daily and you get this even on the $10 plan. But on a $25 plan you get the high-end models. So we are approaching AI. I like to think very mindfully, like not really rushing into it, and trying to be very thoughtful about how we use it, how we integrate it in product, how to add value. For example, in Kagi search, I don't know if you knew any search results that you can, that you see you can, for example, translate them, you can summarize them, you can ask questions about that page. So that's all sort of one click away, but still on demand.

0:14:01 - Leo Laporte
So by the way, you can also exclude ai generated images from the image search, so you, you can have it both ways. If you don't want to see ai images, you can. I have that still turned on, but I mean I like seeing all the images. Yeah, this is the most customizable search engine you've ever seen with these lenses, the personalized results I mean there is. There's nothing you can't do in this search engine. It's pretty amazing.

0:14:27 - Vlad Prelovac
I think what we are showing here is what a search engine looks like when you optimize for the user and Kagi is all about user agency, so putting user in control, not trying to figure out for them what they want to see, but actually giving user the tools to personalize their search feed to find the results they want and things like that. And this is what people pay to have someone for.

0:14:52 - Leo Laporte
It is an issue, I think, for a lot of people. You can't just choose Kagi on the iPhone, but you do have instructions in your Kagi documents of how to make that happen, cause it is possible to make that happen. You also offer your own browser browser, the orion browser, which is a web kit based browser for mac, iphone and ipad and that, of course, can use and does use kagi as its default search. So if people want to switch to kagi, there are ways to do it, it's just there's a few extra steps. This is where google's you know the power of the default has made google so dominant. I think it's interesting that people are starting to turn away, though, even though it takes extra steps, which where do you get your search results? That's the other thing people want to know. Is this just being recast?

0:15:35 - Paris Martineau
Yeah, could you talk a little bit about your like web index and your news index?

0:15:40 - Vlad Prelovac
Sure. So we, we are a small company, bootstrapped and all that. So I realized from day one you cannot, you know, we will not be able to index the entire web and things like that. It's going to be billions of dollars. So we are getting search results from every other search engine that we can pay the results for.

And then we also build our own index which focuses on non-commercial content, which is, a the content I enjoy the most reading and B something that's missing in other search engines, and by this I mostly mean personal blogs and forums and things like that Sort of unmonetized human writing, so to speak. And then we mash all that together, we apply our algorithms, we do some nice things like, for example, we measure how many ads and trackers are there on any page and if there are too many, we downrank those kind of results because they usually correspond to low-quality content. Little signals like that that other search engines cannot do, because they usually correspond to low quality content. I mean little signals like that that other search engines cannot do because they're, you know, ad based and would be admitting that their model doesn't work. But but we can and this helps tremendously propel high quality, you know pages up and and downrank those, those bad pages. You know, with lots of ads, trackers, affiliate links, you'll see them.

So, yeah, we do our own stuff on top of search results. But then, most importantly, we give users the tools so you can block a website and cut your results if you don't like it. Lots of people do that for political content, for example. You don't like this website or this news source, say, block it in my search results, I don't want to ever see it again. And similarly, you can promote sites you like, and we actually have a public list of sites that Kagi users block and promote the most, so you can quickly Like the old blog roll. Yeah, basically, if you go to kagicom slash stats, there's something called domain insights and this gives you an idea of what Kagi users, at least, are thinking are good or bad websites on the web.

0:18:07 - Jeff Jarvis
Vlad, I confessed to you before we got on that I'm late to Kagi. I'm already loving it in no time at all. I'm curious about your views of the web as a whole. People are complaining about Google being worse, and there's two theories about that that Google's doing things wrong. But also the web is worse, not just because of AI, because of all kinds of things. How do you view the state of the web itself? And you've already mentioned a few ways in which you try to clean some of the junk out, but where do you see the trend going? How do you fight I can't help but say it the enshitification that's out there?

0:18:42 - Vlad Prelovac
Yeah, I think it's pretty simple. Up until until 2004 or 5, the web was in pretty good shape and it was driven by, I would say, self-expression and, uh, you know, people just creating those quirky, funky websites. I had one that I would probably be ashamed of now, but uh, you know, it was the was it on geo cities, yeah something like that like all those crazy graphics and yeah, under construction I was saying how madly in love I was with meg ryan and things like that.

0:19:19 - Leo Laporte
So you know well, I certainly hope that hasn't changed.

0:19:24 - Vlad Prelovac
Stuff that one would do, but that sort of quickly changed as soon as then the most popular search engine, google, adopted ads as a business model. Larry, in their white paper describing PageRank, which basically led to, google said very clearly that ads will ruin any search because of misalignment of incentives, and they were giving examples of other search engines at a time doing that and saying how Google will not do that. And about six years from that paper, they did indeed embrace the ad-based business model and I think this is what fundamentally led to deterioration of the web over the next few decades, because first, the search got worse because the incentives were wrong and the search experience was optimized for search customers, which were not the same as search users. The customers were advertisers, obviously and the web consequently got worse because there were a whole new generation of websites built around ad monetization as their sole primary purpose, which led to a ton of low quality content and most of the web today I would say 99% of all pages, exist for the purpose of monetization through ads and they do not exist for the purpose of informing and educating the reader, and that unfortunately includes most of the news websites that we read.

So this process has been slow and we haven't really been noticing it, as it gradually happened over two decades, but it got into such a bad situation that many people today are waking up to this fact and you cannot basically browse the web today without an ad blocker.

If you're doing that, then good luck. But if you think about it, if everybody had an ad blocker, 99% of the web would disappear. 99% of the web would disappear. So this broken web exists basically and fundamentally only because there are still people out there, basically the most vulnerable, least tech-savvy people, who are still consuming these ads and being targeted by them. So it's sort of a sad story, but I would say the last 20 years of the web have been built on a lie and it's now sort of unraveling. With these AI or using Google results, those websites are suddenly realizing that there was no contract in place and that Google convinced everybody that exposure was payment enough, like sending traffic, and these publishers got addicted to that traffic. The traffic is there no more and Google sort of kept all for itself. So, yeah, it's been sad to watch.

0:22:53 - Paris Martineau
So, yeah, it's been sad to watch but yeah, I got sort of woken up from this in 2018 and decided to do something about it. What was the?

0:23:01 - Vlad Prelovac
tipping point for you, when did you kind of realize that this was a problem? Problem, yeah, there were two things that happened, um. One was I literally remember the query I was looking for git lab, which is like a github alternative, and google served me an ad for a git lab and then that same result below as an organic result, and I realized that I, as a user, am getting two identical results and I felt like a monkey. I felt like my intelligence was so insulted by that and for a trillion-dollar company to do that to me as a user, I thought it was outrageous. And at the same time, my oldest kid was starting elementary school and here in Palo Alto they all get Chromebooks with Google pre-installed and all that A fantastic distribution strategy for Google.

But I also realized that my kid will be exposed to this crap over the next 12 years. She will be profiled, targeted with ads and all sorts of bad things that I didn't want for my kids and I have three kids, by the way. And, yeah, at a time I could pay for YouTube premium, for example, to get rid of ads, but there was no search engine, which I thought was, you know, many times more important. And this is where the idea came from.

0:24:25 - Jeff Jarvis
Basically, you dropped your daughter out of the Chromebook. I mean, what did you do there?

0:24:30 - Vlad Prelovac
Funny thing happens. So you know, all my kids use Kagi. They install that on their Chromebooks. They go to school. Teacher sees Kagi and the teacher goes like what's this? And they turn it back to Google. Really, so there's a lot of wow education to be had. Yes, I mean, the teacher is acting obviously in. You know they're thinking they're acting best interest, thinking the kid got you know malware or something right, um, and you know it will take time for for these ideas to propagate, I feel. But uh, yeah, more and more people are are realizing it did you teach the teacher?

no, not really, but we are working on it. That's. That's what we have Kevin for Kevin. Kevin is the head of education. He's good, yes, yeah, yeah uh you.

0:25:19 - Leo Laporte
She should have said my, this is my dad's company. What do you want?

0:25:22 - Paris Martineau
yeah, hello Alto, for god's sakes, I was gonna say those kids have all gonna have like seven parental like installed apps that their parents work at. Every kid in that school I'm sure has a different app so you also do.

0:25:36 - Leo Laporte
You have your own spider.

0:25:37 - Vlad Prelovac
Now are you able to do your own web searching yeah, we, we crawl the web but, as said, we focus on the non-commercial web, right, because this is what's being left out in all other search engines and you know, just anecdotally, from our users, this is, I think, what many of them find very valuable to actually read human blogs and things like that. I completely agree. Even when you do commercial queries like best you know, running shoes or headphones, you're going to see a lot of writing from people who are not trying to monetize their content, and that is extremely valuable. I want to read from the guy who tried five you know different running shoes and you know, picked one, and there are no affiliate links. Um, so that's what I'm looking for.

0:26:30 - Leo Laporte
So so, yeah, this is, this is the, the kind of the web that we are mostly interested in and have resources to crawl I noticed that the number one site people uh block, because with kagi you could say I don't want results from this site or that site is pinterest the number two site pinterestcouk, the number three site pinterestde, the number four site pinterestca.

0:26:58 - Paris Martineau
it goes on. I find this fascinating so I don't. I don't block anything I like pinterest, but I also do think this is definitely like a meme among techie people. They're like all of my search results have Pinterest in them, so it's very funny to see that.

0:27:12 - Leo Laporte
One of the other things you're doing is integrating with other third-party applications, which I think is very cool. If you're a Mac user and use Alfred, you can have that built into your Alfred search. Use alfred, you can have that built into your alfred uh, search, um I. I imagine there'll be more of these as people become more and more aware of of kagi. You also if and this is really important for people who are very privacy focused you have a very clever privacy protecting tool that makes it so that you can't even see what people are searching for. Tell us a little bit about that feature.

0:27:47 - Vlad Prelovac
Sure. So one of the main features of Kagi, I would say, is that privacy basically comes with a business model, because we are not interested in user data. We are interested in your $5 or $10 a month and that's it. That's where the transaction stops and which makes user data a liability for Kagi, actually. So we want the least amount of it as possible, but we also realized that saying that is not enough for some users, for hardcore privacy users, and so we developed this technology called Privacy Pass, which actually builds on some very interesting research that has been done recently in the field of cryptography.

And basically, when you use Privacy Pass in Kagi, we cannot tell who the user searching is on a technical level, like, even if we wanted to know who you are, we still cannot tell. But we can authenticate and know that somebody paid to be able to search, because ultimately, you need to be able to pay. This is why this was such a challenge because you need a payment, but we wanted not to have ability to know who the user is on a technical level. So this is what Privacy Pass does.

0:29:13 - Leo Laporte
And then, if you really want to add to privacy and Lee Hutchinson talks about this you also have a Tor address, so people can go to you without even using their IP address. So you can be completely anonymous on.

0:29:24 - Vlad Prelovac
Cockpit yes, you can be completely anonymous on Cockney. Yes, you can be completely anonymous and still have paid for your account. Really nice, do you work?

0:29:31 - Jeff Jarvis
with folks like Common Crawl. Are there others who want to use your crawl? Is there a crawling community?

0:29:40 - Vlad Prelovac
Yes, there are. Common Crawl is something we use. Actually, one of our employees comes from Common Crawl. We are also opening our search API very soon. Oh, nice demand for that, Especially now in the age of AI. I would say all these models are sort of converging to the same place and what makes the difference in accuracy is the quality of search you feed to them. So having something like Kagi search results instead of your Bing or Google results can tremendously increase the accuracy of AI answers, which is still their weakest sort of link. So very bullish on the search API that we are working on.

0:30:32 - Leo Laporte
We should somebody's asking in our club to Discord, we should mention that you are on all platforms Windows, mac, linux. It's easy to integrate into almost any browser. You offer your own browser, orion, which is for mac only, or ios only, but do you plan to? You also plan to do that in linux? I think yes, yeah yeah, we'll now.

0:30:51 - Vlad Prelovac
Yeah, you know, in kagi there's no lack of ambition, there's only lack of resources. So it took us a while to get to the stage that we announced we are now going to do orion on lin, linux. And then a lot of Windows users said, well, why not Windows? Like we are the biggest market, linux is the smallest. Like we are not optimizing for for market share or growth, we are doing things that we can do. Um, so yeah, linux is in the works. It's coming in about nine months. We we hope so, yeah, yeah.

0:31:24 - Leo Laporte
Well, it's based on WebKit, which is based on, of course, k back in the day, so Linux is kind of a natural place to go. Next, I think, from a development point of view, doggo is your little mascot, right Is? That your dog. Yes, do you actually have?

0:31:41 - Vlad Prelovac
a doggo. The one on the screen is user generated one. We had a contest oh, I'm not sure, actually, my eyesight is pretty bad but yeah, we had a contest with users who draw a dog. That's very cute.

0:31:56 - Leo Laporte
When you click on him, he looks down and says let's fetch. So, that's a little interaction there.

This may remind you of the old days when a search page didn't have a lot of crap on it, it just had what do you want to find? You still have you do a lot of the Boolean stuff that Google did, including the quoted searches. You have the lens, which is fantastic as a way of customizing the sources. I can go on and on. Read Lee Hutchinson's very good article on Oztechnica. He talks about a lot of features, some of which I didn't know. In fact, now I'm starting to think, instead of paying for Perplexity and ChatGPT and Anthropics Cloud and I have $20 subscriptions with so many different AIs they're all available in the Assistant I might just move it all to the Kagi assistant. Right, it's the same. What do I need all of them for? In fact, it's better because I can customize it and you have some really interesting ones, including this new Moonshot LLM, which is pretty interesting. I'm starting to think this is going to be. I'm going to spend more time than ever before in Kaki. I congratulate you.

53,000 members as of this recording. That's great. Try it for free. Folks Pay for it if you want it to continue. I really think this is so important, having been, you know, having seen Neva go down because they just couldn't make it as a paid search. I really want Kagi to make it. Um, and I think, vlad, you got the right idea. Um, so glad you're doing what you're doing. Thank you, vlad prelovac, it's really nice to talk to you. I didn't give you guys much chance to talk to Vlad because I'm kind of uh, oh you're, you're the super fan.

0:33:42 - Paris Martineau
I'm starstruck yeah, he's really been talking about Kagi for a long time on this podcast.

0:33:47 - Leo Laporte
Yeah, I don't know why everybody doesn't use it. This is fantastic.

0:33:52 - Vlad Prelovac
It takes time, and thanks for helping spread the word. Leo, it does take time.

0:33:58 - Leo Laporte
It does take time. I learned about it first, as many did on Hacker News, and so Hacker News probably spread the word first. But man, it's just, and you keep adding features, which makes it even more interesting. We love what we do. Yeah, lifetime member. Orion Plus member, by the way. Oh, thank you so much. You know, I want you to make it, I want you to succeed. It's so exciting and you know what You're getting there A million queries yesterday. That's good. Yeah, that's a lot of traffic going on.

0:34:32 - Vlad Prelovac
I like to say that's about nine seconds of Google searches in a day. So you know we are nine seconds of Google, that's not too bad.

0:34:43 - Jeff Jarvis
Are any VCs coming to you now?

0:34:46 - Vlad Prelovac
Yeah, we have VC interests now, but now we are not interested. I was going to say it's interesting.

0:34:53 - Paris Martineau
You exclusively raised, like the $2.5 million you guys have so far, outside of your contributions from user investors. So, do you have any further plan to, I guess, raise from either user investors or go the VC route?

0:35:09 - Vlad Prelovac
Yeah, we found out that the user route was really nice because it was super easy. We had a lot of interest. I think about 90 users ended up investing and these are, all you know, our biggest fans. We get a lot of help and support from them. I brainstorm with them all the time. So it's much better, I think, than having an investor-based board that tells us what to do.

0:35:38 - Paris Martineau
You don't have the pressures to scale and no pressure.

0:35:41 - Vlad Prelovac
Yeah, I like that for me. No pressure, yeah, I like that for me. Um, so so, yeah, well, we do have some pretty ambitious projects down the road that we are experimenting with now.

0:35:51 - Leo Laporte
So if that needs extra capital, they're probably going to go that route good, yeah, I I don't allow myself to invest in tech companies, but I think this would kind of this would be okay, it's a public benefit.

0:36:01 - Paris Martineau
After, let's get there a way to um a user martin sassenberg. If there's a way to like, do you guys have documentation on how to SEO websites for Kagi to try and put a website into getting scraped by you guys?

0:36:15 - Vlad Prelovac
It's super simple. We have something called Kagi Small Web. So if you have a personal website or a blog that's non-commercial in nature, you just add it there. We automatically crawl you. And not only that, we're going to push it to number two or three and results for relevant queries all automatically. So this is basically our way to surface all the high quality stuff people are writing out there. If it's a commercial website, then it. If it's a commercial website, then it's a different story, and one day we're going to launch Webmaster Tools and have more information on that. Nice, awesome.

0:36:53 - Paris Martineau
Thank you so much.

0:36:54 - Leo Laporte
I almost thought you'd say, no, we don't want any SEO. I'm going to add my little blog to the Kagi small web feed and also you're offering AI tools to small blogs, which I have also applied for, because I think that would be very interesting. Yeah, really cool You're doing it just shows that you could do good, important, innovative work, without being about monetizing your customers in any way. But give us five dollars so that we can, you know, support you. I think that's fantastic. Thank you a lot. Uh, kagi is also on Discord, uh, so you can talk back um and uh and learn more about kagi from the people who, uh, who are working on it right now. Vlad, thank you so much for your time. We really appreciate it. Thank you Been an honor. Kagicom.

Try it for free, this episode of Intelligent Machines brought to you by Monarch Money. Oh, I like Monarch Money. This is actually a good example of a company who says look, we don't want to take your information and sell it. You can subscribe and we will give you objective information about your money and where it stands. I, in fact, I like it so much I immediately signed up for a year as soon as I found it. Monarchmoneycom.

Finances can be messy and confusing, and especially nowadays. Messy and confusing, and especially nowadays. It's like what is going on. Monarch money basically acts like your personal CFO, your chief financial officer, giving you full visibility and control so you can stop, just you know, kind of grinding away and start growing your assets. It's a lot more than your average budgeting app, of course it does. It does budgeting, it's great, but Monarch Money is more like a complete financial command center for all your accounts credit card checking, savings, investment accounts, assets and your goals. You also put in your goals. You don't just manage your money, you're going to start building your wealth, and right now we've got a great deal. You could start managing your wealth by getting 50% off your first year if you go to monarchmoneycom and use the code I am.

It's really about building the life you actually want and, trust me as somebody on the other side of retirement, you got to start managing your finances now, you young folks, to build that life. Without a clear financial picture, financial dreams, that first home, the new baby retirement, can just feel impossible. But burying your head in the sand is not the way to get what you need. Monarch makes managing money simple, even if you're busy. All of your accounts are in there all your credit cards, all your investments, your investments everything in one place so you know exactly where you stand without the hassle. You could track your spending, your savings, your investments effortlessly, so you can focus on what matters most making your biggest life goals a reality. And I've set it up to get notifications about things that are important to me, but not be bugged about the stuff I don't care about. It really is the financial tool people actually love.

Join over 1 million households. Named Wall Street Journal's best budgeting app of 2025. Is the top recommended personal finance app by users and experts Over 30,000 five-star reviews. Get control of your overall finances with Monarch Money. Use the code I am at monarchmoneycom in your browser for half off your first year. That's 50% off your first year, monarchmoneycom the code. I am Highly recommend this. Ladies and gentlemen, let us continue on with the AI news. This was a week. This was a week. This was a week.

It almost makes me feel like uh, I'm not accelerationist enough I wouldn't have guessed that that was the sentence that was gonna no same here. Um, I don't, I don't even uh know, uh, where to begin. We're still waiting for ChatGPT5 from OpenAI.

0:41:09 - Paris Martineau
Did you see someone? Was it Sam Alvins? Somebody from OpenAI said ChatGPT4, it's represented as a tiny dot. Chatgpt5, a much larger dot. Get excited. So you know there could be a much bigger dot coming. Guys, everybody get pumped.

0:41:27 - Leo Laporte
He's really a good marketer. He also uh tweeted something like uh, imagine having the smartest person you ever met in your pocket.

0:41:37 - Jeff Jarvis
Well, that's chat that's a little weird, I wouldn't want.

0:41:41 - Paris Martineau
No, I'm not putting it out of their pockets but I think it also it's overselling the idea that this thing is smart.

0:41:48 - Leo Laporte
It's not smart, but it is very very useful so what they did release, uh, just yesterday, is their first open weight models in a long time. Uh, chat gpt, oss 120b and 20b. Did they ever release open weights before? It says in this wired article chat gpt2.

But nobody wanted now the able they'd be ability to run these models on your local system. Now, the 120 gigabyte model you probably can't run on a system unless you've got a lot of uh, of vram, a lot of GPU how does that compare to the, uh, the small llamas?

I think llama's 70. So it's it's big. It's 50 bigger, uh, more than 50 bigger. Um 20b. You could run on almost any machine. Um Sam said we're excited to make this model the result of billions of dollars of research. They have billions. In fact. They are now letting their employees sell their stock at a valuation of half a trillion dollars, which is pretty good for a company that I don't think has ever made any money, but I might be wrong.

You can download them both for free on hugging face as usual, okay, uh, yeah, the last open weight model released by open ai was chat gbt2 back in 2019 here, anybody could drive our edsel but these are, these are their state-of-the-art models. Well, they will be or they were, or they will be until five comes out right.

0:43:24 - Paris Martineau
And for clarification, models are just basically models that you're able to download locally and tweet and run locally without sending any information back to the home office.

0:43:35 - Leo Laporte
Um, and I think they're being pushed into this by the active chinese community, the qn from alibaba and deep seek, and, yes, french mistral also.

0:43:45 - Jeff Jarvis
It's interesting that that Zuckerberg has been backing off open, yeah, but at the same time, yann LeCun is staying, is waving the flag out loud of open. So I think there must be something going on in meta to decide what's going to be open and what's not.

0:44:00 - Leo Laporte
Yeah, I mean they never officially said we're not, we're gonna stop pull back on our source no, no, but it's been obviously some discussion yeah it was reported by uh, the new york times, that they were considering yeah, internally, but that was it was.

0:44:17 - Jeff Jarvis
It was an interpretation of what he was saying about the, about the difficulties of this, but yeah, I think the pressure to be open is remains huge.

0:44:21 - Leo Laporte
There's also I think, uh, it's fair to say a benefit to being open. Then you get a lot more people working with your tools and and there's some value to that hi, this is benito.

0:44:31 - Benito Gonzalez
I also think like people who are really really serious about the ai stuff.

0:44:35 - Leo Laporte
They they use local, oh yeah oh yeah, in fact I'm very, uh interested in running on my ai locally, um, or, you know, it's interesting. To do it through, for instance, kagi in a more privacy forward way would be very interesting. One of the things kagi does with its assistant is you can say preserve all my chats, which is the default in most ai's ai chat bots, or delete it every day, start fresh. I don't want you to remember anything I say to you, and now that we know that, which sam altman has been quick to say, everything you say to chat, gpt could be subpoenaed at some point by the New York Times. Yeah, the court is forcing them to preserve it in this New York Times lawsuit. Uh, so he wants to make sure you understand that you have no privacy when you're chatting on the server. That doesn't mean if you download the models and chat locally. It's just that it takes a lot of hardware to run.

0:45:27 - Jeff Jarvis
So I spoke to the AI librarian at Stony Brook yesterday.

0:45:31 - Paris Martineau
Is this a human librarian?

0:45:33 - Jeff Jarvis
that knows AI. Yes, yes, it is the librarian for AI, not the AI librarian. I misspoke. And he's running a local model. They've got a few GPUs and he's using Lama and it allows the university to take the collections that they don't necessarily want to make public or can't, and make them open to the students. The openness enables so much university innovation.

0:46:00 - Paris Martineau
So is this? I mean, you'd previously described him as the AI librarian. Is this this guy's full-time job? Yeah, Focusing on he was just hired Wow From.

0:46:08 - Jeff Jarvis
USAID, by the way. Wow, you might want to have him on the show at some point.

0:46:14 - Paris Martineau
I think that would be really interesting. I mean, I'm very interested in the niche applications of these tools and modeling. You know that you can get. Uh.

0:46:27 - Leo Laporte
I'd be curious, I guess, also to hear how students are using it or faculty yeah yeah um, but that's just the beginning of this wild it's a sandstorm coming it is a sandstorm. I'm telling you we're getting close coming.

0:46:46 - Jeff Jarvis
It is a sandstorm. I'm telling you, we're getting close, we're getting, we're getting close, I'm I still love it.

0:46:49 - Paris Martineau
We're not. We're not getting close until the technology to make an ai powered video graphic for this podcast, that is sand focused, exists well, I'm glad you mentioned that there is a new open world model that has just come out from Google.

0:47:10 - Leo Laporte
Let me show you the video that Google posted on the keyword blog. It's called Genie 3. Now, what's different about this? Now, of course, they have Veo, right, but Genie 3 creates open world models, models, models you can wander around in.

0:47:28 - Paris Martineau
They have some persistence as well are you saying the ai bubble has gotten so big that we've circled back around the metaverse bubble? I honestly think are there gonna be nfts in my ai?

0:47:43 - Leo Laporte
if you're a believer in the simulation hypothesis, this would be an important step into the simulation. So you type in a text prompt. It will generate a video, like Vio does, but a video you can wander around in.

0:47:59 - Jeff Jarvis
And that is live.

0:48:00 - Leo Laporte
It's live. So this is the video.

0:48:05 - Vlad Prelovac
You're not walking through a pre-built simulation. Everything you see here is being generated live as you explore it.

0:48:12 - Leo Laporte
Now it's not out yet and we know Google sometimes can oversell stuff Usually they do it on stage. Now watch this so he's painting a wall with a blue. Oh yeah, you can look at your feet, by the way. You can have legs in this open wall, oh anything. So he's painting a wall with a blue. Oh yeah, you can look at your feet, by the way.

0:48:25 - Paris Martineau
You can have legs in this open world, oh, anything.

0:48:28 - Leo Laporte
So watch, now he's painted this pattern. Now watch, he's going to look away. In the past, ais might have generated a whole new wall when you look back, but no, the pattern is persistent and that's one of the Watch this. Watch this new events on the fly something like another person a man in a chicken suit appears from the left.

0:48:49 - Jeff Jarvis
This is about a man from their office in london. So notice, the water comes up on the, but doesn't stay there yeah, it's still not perfect.

0:48:55 - Leo Laporte
If you look closely you probably could tell it was generated all kinds of unique environments, but you'd have to look pretty closely yeah it's pretty amazing, leo the.

0:49:03 - Jeff Jarvis
The video they put up comparing uh version two to version three is striking oh yeah, let's go back to the uh the blog?

0:49:10 - Leo Laporte
is it on that blog?

0:49:11 - Jeff Jarvis
post that blog post.

0:49:12 - Leo Laporte
Yeah, yeah, so, uh, they still have vo. There are limitations to what genie 3 could do. It's only 720p. Uh, it can only do a few minutes, which is actually more than vo can only do a few minutes, which is actually more than VO can only do eight seconds. So this is are these the two together?

0:49:31 - Jeff Jarvis
Yeah this is. It Is that it. Yes, it is. You've got to back up, though.

0:49:36 - Leo Laporte
How do I even control?

0:49:38 - Jeff Jarvis
this Reload the page or something.

0:49:40 - Leo Laporte
Maybe that's it. Yeah, there we go. So, if we go, if we look more closely, genie 2 is on the left. This is the other thing that I think encourages me that we are getting very close to some sort of massive breakthrough is the speed with which these improvements are happening. I mean, it's happening fast, so not only does it look better, you could see it lasts longer. The movements are smoother, yeah, and you're controlling this with your keyboard, with your arrow keys.

0:50:12 - Jeff Jarvis
Paris. As a gamer, are you not jazzed?

0:50:17 - Paris Martineau
I mean, I'm probably not the person to ask, because I like games that are like very human and I enjoy like the actual physical, like one of my favorite games is that gutenberg uh press era, one where everything is pentiment like hand hand pen petment, where everything is as soon as I can buy a switch, a new switch, I'm going to send jeff my old switch so can play it.

Jeff is going to literally love it. It is every aspect of it. Is Jeff, babe? But I mean I'm interested in stuff like this, but it kind of reminds me of some of the criticism around. Oh, probably Benito's going to have to help me. What is the open world game put out by Bethesda that had all the AI-generated worlds that everybody criticized for it feeling too sparse and just procedurally?

0:51:07 - Benito Gonzalez
generated Starfield or something like that. Starfield, yes, starfield, no, it wasn't Starfield, it is Starfield.

0:51:13 - Paris Martineau
It came out, I think around the time of Baldur's Gate 3, and I remember being like, oh, I should get into this. I'm also into Baldur's Gate 3. I love Bethesda games and I mean maybe these advents in kind of the processing power, that world, or think about it in a way that is like what about? This is going to be interesting to a player. It's not going to feel the same as something crafted to be fascinating and interesting.

0:51:55 - Leo Laporte
And so that's also what brought my head with no Man's Sky was kind of similar. I played both Starfield and no Man's Sky and they both got kind of tedious because procedurally generated worlds are not.

0:52:04 - Benito Gonzalez
Exactly Like this isn't new for games Like we've had procedurally generated worlds for a long time.

0:52:08 - Leo Laporte
Well, but what's interesting is, these might be. I mean, look, there's no way these aren't going to be in games over the next few years that these movies, games and other things aren't.

0:52:18 - Benito Gonzalez
Yes, but not wholesale. This way, it's still going to be heavily curated by somebody who's designing an experience for another person. It's not going to be just like a tool for them to do that with. Yeah, it's a tool for them to do that with. It's not going to be like me jumping into this and being like go, you know, maybe it will. That would empower you that's the problem.

It's not like I don't know what I want. I need a good game designer to give me a good game, because I don't know what I want.

0:52:41 - Leo Laporte
Well, then you shouldn't design games. Okay, let's just say that, uh, like you don't know what you want, either until you get what?

0:52:48 - Benito Gonzalez
until you get it, you know what I mean. Like well, yeah, but there's a reason why game design is a field exactly.

0:52:54 - Paris Martineau
Some people are very good at what he's doing, don't you think that?

0:52:57 - Leo Laporte
uh, john carmack?

0:52:59 - Benito Gonzalez
yeah sure. So I want john carmack to design something with this tool for me well, that's exactly what's going to happen.

0:53:06 - Leo Laporte
That's not Okay, but that's exactly what's going to happen. You're going to have the John Carmacks and there's going to be some kid now who doesn't have. John Carmack was an amazing coder, right, he's the guy who wrote Doom and was able to do stuff with very primitive hardware. Nobody thought possible. But there's going to be some kid out there who doesn't code but who has a vision for a game. Who's going to start using tools like this. Maybe not today, they may be a few years off. Here's another example of breakthrough technology. This is called kitten text to speech. It runs on a potato, by which they mean a really crappy computer. This is an open source voice model that can run on your computer CPU only 25 megabyte model. It's teensy, weensy and you might say, well, we've had text to speech before, but it never sounds very good at that. Listen to Kitten.

0:54:01 - AI
Kitten TTS is an open source series of tiny and expressive text-to-speech models for on-device applications. Our smallest model is less than 25 megabytes.

0:54:12 - Leo Laporte
The world needs frontier open source models that are tiny enough to run on edge devices. Well, they're using kind of Kitten.

0:54:19 - AI
That's what Kitten ML is here for I don't think you have to use QC voices.

0:54:23 - AI
It was started just two weeks ago and we're excited to launch a preview of our smallest model.

0:54:30 - AI
I don't think you have to use QC voices oh don't like that one that's grok-esque.

0:54:47 - Jeff Jarvis
Child labor. Leo, have you taken a PDF?

0:54:50 - Paris Martineau
of a book and put it in 11 labs and had it read to you model.

0:54:51 - Jeff Jarvis
Oh no, I haven't done that, yet how do you get it to do that? Trained on, I just saw that it was capable of that. You can also do video.

0:54:56 - Leo Laporte
No, they've added video. Now to notebook.

0:54:58 - Paris Martineau
wanted to, uh, do this with books, with audiobooks, while on my road, with books, while on oh yeah, you were asking about this. We were texting about this Is that books are, it's kind of the book equivalent of copyright, like they're locked down, the EPUB files.

0:55:14 - Leo Laporte
Oh, I showed her how to get around that.

0:55:16 - Paris Martineau
So I need to do that.

0:55:19 - Jeff Jarvis
Like I sent you a PDF of a book about the invention of the loudspeaker. Yeah, let me put that in here. Can you put that in 11 Labs?

0:55:25 - Leo Laporte
Yeah, let's try it right now. By the way, 11 Labs voices are much better than almost anybody else, but the fact that Kitten can do this in 25 megabytes without a GPU.

0:55:36 - Jeff Jarvis
Yeah, yeah, God bless them. No, no, no, no, no spoke on them.

0:55:41 - Leo Laporte
Yeah, and I don't think the voices have to sound that crappy. I have created a new folder just for that book called history. In my book I have a lot of you didn't? Have a history folder in your computer no, most of well, because my collection started mostly with programming manuals. But now I uh and, by the way, we should mention that, uh, uh, this is public domain now you write this book oh yeah, it's as old as hell it's old as hell.

0:56:14 - Jeff Jarvis
It doesn't mean it's an autobiography of a guy who invented the loudspeaker in napa it's called the great voice magnum box, in other words now, how do I get it to?

0:56:23 - Leo Laporte
can do the audio overview, but how do I get a book?

0:56:26 - Jeff Jarvis
Where did I read that?

0:56:27 - Leo Laporte
Damn.

0:56:27 - Paris Martineau
I thought it was on the page Well what I've done before in ye olde chat GPT is I upload a PDF and then I say read this to me out loud, verbatim, in chunks or whatever, and usually then it will break down like a chunk, but I don't know if that would work for a book Cause, then you have to like every time it's done you have to go prompt it again to do another one and Download its reader.

0:56:52 - Leo Laporte
Oh, there's a notebook. Lm reader.

0:56:56 - Jeff Jarvis
That's right. So I'm reading now 11 labsio slash blog slash.

0:56:59 - Leo Laporte
read PDF aloud I will I can do now Cause. So we always had the audio overview, which was a goofy podcast hosts. So we now know. By the way, 11 readerio. Ah, oh, this is oh, you're taking 11 reader. I thought you're talking about notebook lm. No, I was talking about the water. I'm trying to do this in notebook lm oh, I, I have the app 11 has.

0:57:20 - Jeff Jarvis
That's what I'm saying. That's why I was asking you how to do this, because, oh, paris wants to do this oh, I'm sorry.

0:57:25 - Leo Laporte
I thought you said notebook lm, didn't she say notebook online?

0:57:28 - Jeff Jarvis
well, she then went on to that.

0:57:30 - Paris Martineau
But we're talking about I don't even think I brought up notebook lm, to be honest, but oh no, you didn't.

0:57:35 - Leo Laporte
Okay, I said chat gpt all right, well, I pay for 11 reader and I like 11 reader a lot and, uh, yeah it. It's the one where they paid for some very famous voices. I have sir lawrence, olivier and I don't want a famous voice.

0:57:51 - Paris Martineau
For some reason I don't know that. Does that feel strange to you?

0:57:54 - AI
his eyes sparkled and he sent up a great blue triumphant cloud that's like olivier, the paper was made no, it's not, I said precisely, and the man who wrote the note is a german. Who do you know the peculiar construction?

0:58:10 - Leo Laporte
synthesized voice of lawrence olivier. Okay, it's not actually lawrence olivier, he's gone reading, uh, sherlock holmes. Ai is so big, they've resurrected lawrence olivier. Okay, it's not actually lawrence olivier, he's gone reading, uh, sherlock holmes ai is so big, they've resurrected lawrence olivier frenchman or russian could not have written that it is the german. Here's the thing that I find interesting. They are so good at this that it's how he's pausing. He's it sounds like he understands what he's reading, which he obviously ai does not, but it's it's.

0:58:35 - AI
It communicates I'm courteous to his verbs. It only remains, therefore, to discover what is wanted by this german who writes upon bohemian paper and prefers wearing a mask to showing his face yeah yeah, pretty good, all right, all right I will say I like this more than the average audio book narrator.

0:58:58 - Paris Martineau
I keep having this problem that I've brought up in the show probably a million times before which is whenever I'm full time US Employees to adopt some form of flexible scheduling. It's like we've accidentally created.

0:59:09 - Leo Laporte
That's not a famous person, but I like that. That's just an obnoxious person, plush scheduling.

0:59:15 - AI
Cinderella read by Judy Gar garland an old frock to put on and laughed at her and turned her into the kitchen. Then she was forced to do hard work to rise early before daylight to bring the water, to make the fire, to cook and to wash. She had no bed to lie down on these people like that stop making them work freaking.

0:59:32 - Jeff Jarvis
Amazing the phrase he does yeah, let her rest.

0:59:35 - Leo Laporte
Let's let these people rest oh, you don't want me to use dead people anymore, huh I don't really want you to use live people anymore.

0:59:42 - Paris Martineau
I kind of I kind of like I honestly enjoyed the judy garland one.

0:59:45 - Leo Laporte
I do like that these um I've had a annoying feeling lately about ai.

0:59:50 - Paris Martineau
Although the prophecies that chatbots, that's alex kantrowitz, that's the real person just testing that kind of sounded the fakest, but I think that there's something that you know, like something I've realized about some of these audiobook narrators or people reading scripts for podcasts, is when you're they're focusing on reading. Sometimes they don't modulate their voice because they are focusing on reading and that is really hard for my brain to grab on to.

At least the machine-based readers know like intrinsically in the programming you've got to kind of modulate your tone and kind of switch it up a little bit.

1:00:26 - Leo Laporte
It's hard work someone interesting. This is why actors are very good at this because they actually so. Here's a here's peter rabbit by beatrix potter burt reynolds reading the tale of peter once upon a time there were real bird rabbits and their names were he's dead too, he could have read it in the day that's all ai

1:00:46 - AI
they lived with their mother in a sand bank underneath the root of a very big fur tree I think that's pretty freaking amazing.

1:00:53 - Leo Laporte
Do you? You want Maya Angelou reading it, or Richard?

1:00:57 - Jeff Jarvis
Oh, here let's do.

1:01:05 - Paris Martineau
Do you have yourself reading the AI thing One? The end.

1:01:14 - Leo Laporte
Yeah, I think the AI would do a better job.

1:01:16 - Jeff Jarvis
I hate to tell you you what sets magazines apart from their printed brethren?

1:01:20 - Leo Laporte
no, that's good, see he, because he knows his course. Often I like the author reading their books. By the way, if you're playing this from audible, you're going to get us taken down, so tactile impression of a doubt.

1:01:32 - Jeff Jarvis
Have you ever heard me speak so slowly?

1:01:35 - Leo Laporte
That's why they have producers.

1:01:36 - Paris Martineau
How many times do they have to tell you to slow down?

1:01:41 - Jeff Jarvis
A lot. It's more muffed words. How many takes, do you think you take?

1:01:45 - Leo Laporte
When Audible was a sponsor, I went to their Newark headquarters and they said here you can read a story. And, man, they had to stop me every five words and slow down.

1:01:54 - Jeff Jarvis
It is so hard and you've got a producer who's reading the text and if you leave out a the, they see it and you've got to go back and redo it.

1:02:03 - Leo Laporte
You've got to go back, yeah.

1:02:04 - Jeff Jarvis
So Paris, other cheaper productions, they back the tape up to where you screwed up, and then you've got to pick up. Oh boy On the more expensive productions. You just go back and repeat it and they edit it wait, they still use tape, are they?

1:02:17 - Benito Gonzalez
they're using tape. What are you talking?

1:02:19 - Jeff Jarvis
about virtually. Yeah, there's a guy, they've got a whole they got to stick out scissors and tape on the tape and then they splice it sometimes they run out of tape here is the physicist, richard feinman, reading the little match girl girl by hands, christian anderson.

1:02:36 - AI
Most terribly cold it was, it snowed and was nearly quite dark and evening that's more christopher last evening of the year.

1:02:42 - Leo Laporte
This is no, that's. That's really what a fireman sounds like I have uh no, no, he doesn't find his lecture. That sounded like yeah, he sounds just like that so this is all 11 lives you've been playing. Yeah, it's an app. It's a great app. You can put in any pdf that's anything anything you own. Yeah, you can even train it with your voice if you really wanted to hear yourself. Oh, you want to hear deepak chopra.

1:03:06 - Benito Gonzalez
No, it's with love, compassion and gratitude you open yourself to the infinite potential of the universe.

1:03:13 - Leo Laporte
See, they licensed all this stuff. That's what's interesting. How about John Wayne reading Sherlock Holmes? That'd be pretty funny, wouldn't it?

1:03:30 - Paris Martineau
I just imagine that this is what you do in your free time, when you're not on the show, but just maybe to no one at all yeah, just by myself, lonely man, I just sit here and um has anybody used ai for anything interesting this week.

1:03:50 - Leo Laporte
Personally, or I use it every day.

1:03:54 - Paris Martineau
But I know, but is there anything you're like? Oh do it it's kind of useful yeah.

1:03:58 - Leo Laporte
So a family member has a? Um got an email from his doctor saying well, we gave you something interesting on your ct scan. We think you should come here. It's a. It was cause for concern. They wrote and then they described it and and lisa and I are looking at each other going what? What does that mean? The cause for concern, how serious is this? We put it into perplexity and got a lot of great information out of it. Um it. It actually eased our minds because of the variety of causes possible for the symptoms he was showing did you share those results immediately with this person?

Yeah, we said hey, here's the good news, because you know what everybody always goes to cancer. Here's the good news there are 20 other things that could cause this, not cancer.

1:04:51 - Paris Martineau
And what did the doctor say?

1:04:53 - Leo Laporte
Well, he's going to go in for the test. But this is the problem he got that email, didn't even talk to it. This is the modern medicine. This is a very good use for ai, because modern medicine now is just terrible. In the us, no doctor called, which I think on something like that, a person should drive me nuts, yes, and then said and we'll make an appointment for you in a week. So he's got to sit for a week in a mystery. So I think that's a very good use of AI. Of course you're not looking for well, that's cancer for sure. You're looking for options, information, and it was very good at doing that. Um, what else have I used it for lately? I use it all the time Did you say I use it for my diet and for exercise.

I will say so.

1:05:38 - Paris Martineau
The only uh the reason this is prompted by is, uh, I log uh all my meals through like one of those like fitness tracker, like like calorie breakdown, macro sort of thing.

1:05:51 - Leo Laporte
Which one do you use?

1:05:53 - Paris Martineau
I use, lose it, I don't have any particular chronometer listen. Yeah, I probably should change. It was just the first thing I decided to download. It works. Take a look at chronometer.

1:06:02 - Leo Laporte
They just added a new feature I do like you, take a picture of the meal and it tells you what you're eating and what the calories and macros are using what chronometer using what ai tool, because what I'm about to describe is how I use like chat, gp, chat gpt 4.5 for that and it usually ends up being fairly accurate.

1:06:25 - Paris Martineau
I don't use it. I use it from like eating out or something, or, if I like, get a biscuit from the bakery and then I weigh it at home and I want to get a rough estimate of what the calories are so I just took a photo of my lunch.

1:06:38 - Leo Laporte
This isn't a good test, because it says what's on it, yeah, but it's scanning it right now and it's going to tell me I have.

1:06:46 - Paris Martineau
No, I don't imagine that chad bt or whichever uh llm is powering. That is 100 accurate, but it's kind of fairly oh, does it go straight in there?

1:06:58 - Benito Gonzalez
it's also written on the side of the thing you just scanned, by the way. Well, yeah, he read that I mean, no, I know.

1:07:02 - Leo Laporte
In fact, mostly what you do is take a picture of a plate of food without a description, but you know it and it may be wrong. It says that was 2801 calories, but I'm not gonna eat the whole thing is that correct?

1:07:15 - Paris Martineau
no, but no, chronometer is fantastic.

1:07:18 - Leo Laporte
I have used, I use Lose it, I've used them all. This is the one I use now. It ties into Apple Health. It ties into my aura ring.

1:07:26 - Paris Martineau
I mean they all tie into that.

1:07:27 - Leo Laporte
It ties into everything.

1:07:29 - Paris Martineau
Yeah.

1:07:30 - Leo Laporte
My scale, so it knows everything about me.

1:07:33 - Paris Martineau
I don't know. I found it uh somewhat kind of useful daily thing that I like even I I've been using you're not on a diet, though I just track my what I like consume it's a great habit.

1:07:47 - Leo Laporte
You should get it.

1:07:47 - Paris Martineau
I think it is very good, yeah, you know, but I I recently like was um purchasing some of those like ready meal things from one of those companies, thistle. Honestly, it's been pretty good. I'm using.

1:07:57 - Leo Laporte
Oh, is this all good? You like that? Yeah?

1:07:59 - Paris Martineau
I, I have, I've had. I'm getting allegedly a delivery of it during this show, so I'll go and get the go and show us I will show you um.

1:08:09 - Leo Laporte
Their pitch is something right.

1:08:11 - Paris Martineau
They do health or something they do like a kind of vegan plant focus. They also have meat options, but I really just enjoy it like they've got like good salad options for lunch and, uh, interesting dinners. I'll show you, I guess, when I get it. But part of it is, um, they have the macros and the breakdown on the back of each package, but it's like in these like little small font, and I'm like I want to get that into my app, but I don't have to sit there and like oh, you could take a picture of it.

I just I used to say. I just took a photo of it in chad gpt and had it combine both the salad dressing and the other one and all of it into one thing, copied it over, and it was great, isn't that nice.

1:08:45 - Leo Laporte
I don't know. So here's another thing I do. I gave uh, I give um the ais in this case, I think I did all of them just to see who gave me better results a picture of all my supplements and medications, and then, and then it gives me information about them, whether I should take it with a meal or not. And then I, and then I said, okay, is this a good mix? What my? They gave me an overall assessment. What are the? What are the interactions? Uh, what? What should be monitored? What my next blood work should be based on, what I'm eating? Um, this was really useful. This was fantastic so what did?

you go back to the photo.

1:09:26 - Paris Martineau
What are all the supplements you're taking?

1:09:28 - Leo Laporte
oh, it's a little out of date. Now I don't do creatine anymore, but that's a picture and I and I.

1:09:33 - Paris Martineau
It's on the internet. No, okay, I thought you. It looks like you were scrolling through a web page. I was no, it's obsidian blog about your supplement I could.

1:09:44 - Leo Laporte
I could take it straight out of obsidian, but no, this is obsidian. And it recommended a couple of additional supplements it said you should given and I told you know the prompt was my health conditions and all my weight and all that stuff it said you might want to try ala, it helps with diabetic neuropathy and reduces fasting glucose. Also berberine vitamin k2. It recommended some stuff to try, uh to add to my uh.

1:10:08 - Jeff Jarvis
It's good it was really. What do you think of illinois's law just signed?

1:10:13 - Leo Laporte
I was okay let's talk about this. Illinois has been on the forefront of anti-ai legislation. Um, they uh have a very good, I think, anti-face recognition law.

1:10:24 - Jeff Jarvis
Tell us about the new one, jeff so they've outlawed the use of um ai for therapy for mind therapy, either marketing it as a therapist or allowing therapists themselves to use it as a part of their services.

1:10:42 - Leo Laporte
I think this is. I couldn't disagree with this more. I acknowledge, as we had an interview on this and we've talked about it, that there are problems with using it, but many therapists have said there it's a valuable tool and my biggest problem is it's premature for the for a state to just outlaw it.

1:11:02 - Jeff Jarvis
People in illinois will not have the option to try this but I think there is a need for some mechanism for standards, because I can go along tomorrow and I can just use chat, gpt and I can say this is going to save your life and I can market it and make money and, uh, be a charlatan of charlotte. This isn't a standard, it bans it. I, I know, but I'm saying there's something between this and that. Oh, yeah, yeah, yeah that that just to allow it and to not have a caveat, uh, mtor, uh, here is, uh is dangerous, I think all we should have is a caveat mtor, uh and information for people and let them choose, because the jury's still out.

1:11:40 - Leo Laporte
There are plenty of people, including therapists, who say it's a valuable tool I don't know.

1:11:46 - Paris Martineau
I think it's fairly reasonable. Hold on my headphones. Just died um gosh darn it okay, one of them is still in, so I'm going to keep talking. I think it's fairly reasonable that this law prohibits AI systems from delivering therapeutic treatment or making clinical decisions. I think that's pretty straightforward and perfectly reasonable. At this stage in the game, you should not have a LLM like chatbot lithium clinical decisions and giving therapeutic treatment well, I think that's probably true.

1:12:21 - Leo Laporte
I don't think an ai can prescribe anything in absent an actual physician.

1:12:25 - Paris Martineau
Those laws already exist, so a physician would still have to write the prescription yeah, but I mean, I think that this is kind of common sense like first step.

1:12:35 - Leo Laporte
But that's not what this does. This says you can't use it at all.

1:12:41 - Paris Martineau
Let's go to the text of the bill.

1:12:43 - Leo Laporte
It's an outright ban and by the way. Florida's Ron DeSantis says we're next.

1:12:50 - Paris Martineau
AI is very dangerous, it says this act provides that an individual, corporation or entity may not provide, advertise or otherwise offer therapy or psychotherapy services to the public in Illinois unless those services are conducted by an individual who is a licensed professional, and that licensed professional can only use an AI system to the extent that that AI system meets like the definitions of an AI system. They cannot use that AI system to make independent therapeutic decisions, directly interact with clients in any form of therapeutic communication none of their business therapeutic right.

1:13:30 - Leo Laporte
So, leo, do you think I should be free? I don't want them telling me what I can and cannot do well, that is. Bs so. You're not a licensed therapist, it's not telling you anything. No, it's saying that a licensed therapist cannot offer you licensed therapy services. Oh, so an individual in the state of Illinois can use this.

1:13:47 - Paris Martineau
I could go to.

1:13:47 - Leo Laporte
Abby, the website where they have therapy. That's legal in Illinois.

1:13:51 - Paris Martineau
Well, I don't know if it would be legal for Abbyby to offer that, but I think that you could use an llm to give yourself one on one.

1:14:00 - AI
They're not going to take you to jail they're going to say, that's the most dangerous use of it?

1:14:06 - Leo Laporte
is some individual saying I think I'm going to get some therapy from chat gpt?

1:14:10 - Paris Martineau
that's the most dangerous use of it, arguing on this show a million different times. That should be totally fine.

1:14:15 - Leo Laporte
Yeah, well, I don't want the government to tell me what I can do with my chat gpt. But if you wanted to say, well, here's the most harm. It's when individuals absent the help of a therapist.

1:14:26 - Paris Martineau
Use it well, that's what we're saying no, they didn't ban that it bans um using an AI system to make independent therapeutic decisions, directly interact with clients and generate therapeutic recommendations or treatment plans without the review and approval by a licensed professional. So a licensed professional could use these tools, but they have to have review and they have to be in the loop.

1:14:54 - Leo Laporte
Can I use it in my house?

1:15:01 - Jeff Jarvis
In the privacy of my own home. Can I use it? Say anything about that, leo? Put this way, can I just go to Illinois, I'm gonna move to Illinois and I'm gonna put a shingle outside my door and I'm gonna declare myself a therapist that's the way it is in in California.

1:15:10 - Leo Laporte
You don't. Anybody can be a counselor in California.

1:15:12 - Paris Martineau
That's California is different than a licensed therapist.

1:15:15 - Leo Laporte
Well, no, I think that's true. In most states there is a licensing program. I know my wife was a mft and before that an mfcc, and she went through a lot of stuff to become a licensed therapist my ex-wife uh. But that doesn't stop somebody from just putting out a shingle saying I am a counselor and if you've got ptsd I can help.

1:15:36 - Jeff Jarvis
That's not illegal that makes me uncomfortable I mean you have to get a license to cut hair.

1:15:44 - Leo Laporte
It makes me yeah, well, that's another problem. It is a lot harder to become a hair cutter in california than it is to become a lot of things, uh, and I don't think it's exclusive to california, I so, first of all, this is an issue because each state is going to have its different laws and it's going to be a variety of laws, and this was this was why they wanted to.

You know, marcia blackburn wanted to ban all state regulation state regulation, but the problem was, they didn't propose they didn't propose anything, any federal regulation either, so they were essentially saying we're going to ban regulation. I think, if I think, if there's going to be regulation a, it should be very judicious and careful, because we don't know a lot about what's going on right now but now that person who would do it federally is just killing mrna, which might cure cancer, and we're not going to get there.

So I don't really want federal regulation well, yeah, so maybe we shouldn't have any regulation, right? Yeah, that, by the way, is the most shocking thing that's happened.

1:16:45 - Jeff Jarvis
Yeah that's as anti-science as you can get basically, you want to talk, test real.

1:16:51 - Leo Laporte
It's clear that rfk jr is a eugenicist and he thinks that if you get sick, you should, you know, either get better on your own or die and and make the species more strong. And I, this guy is insane and he should not be in charge of our vaccine program. Yep, but he is uh. There is a lot of uh information. It look at the information is unclear at this point about how useful or unuseful, and that's why I think it's a premature for illinois to ban. It says bans ai from providing mental health services.

1:17:26 - Paris Martineau
I don't you're not looking at the text, the bill, you're looking at someone's write-up, the. The write-up says it uh prohibits a licensed therapist to use an AI system to make independent therapeutic decisions directly and direct with clients.

1:17:42 - Leo Laporte
Why shouldn't a therapist, a trained therapist, be able to do that? Can your physician without? Can't your? No, no, without Let me finish speaking.

1:17:49 - Paris Martineau
It says generate therapeutic recommendations or treatment plans without the review and approval by a licensed professional. I think that's really reasonable. I think that if you're trying to offer services that are being described as coming from a licensed therapy professional, you can't do that just by handing it off to an LLM and claiming that that's you.

1:18:14 - Leo Laporte
Okay. So it's a very narrow thing, which is I'm a, I'm a licensed professional. You come to me for therapy. I say, hey, here's your chat bot. Go have fun. I'm not going to pay any attention to what conversation you have. That's what it protects against.

1:18:30 - Jeff Jarvis
And it also says I can't start a company that's going to offer.

1:18:34 - Leo Laporte
Well, that's what that therapy? Yeah, so those websites might um might struggle in illinois.

1:18:44 - Paris Martineau
And I mean I, yeah, I'm not sure that's a bad conversation. I don't think, okay, all right, yeah.

1:18:49 - Leo Laporte
But I think it's a little moral. I think it's moral panic jeff because it comes from reading these stories about oh, this guy killed himself.

1:18:56 - Jeff Jarvis
He said it don't we get out, don't we get a bit. You know, it's one of my favorites.

1:19:05 - Paris Martineau
It really is great.

1:19:06 - Leo Laporte
I don't like laws that are written to respond to sensationalized stories.

1:19:12 - Paris Martineau
I mean, then you don't like any laws, I guess.

1:19:14 - Jeff Jarvis
That's pretty much all it is. Laws are ripped from the headlines, yeah.

1:19:20 - Leo Laporte
Did you see? Wow, and I put this in the show notes. First of all, palantir has just signed a $10 billion army contract. Palantir, which is Alex Carpentereter teal's ai for military and police and police. This is a 10-year deal. They are they are raking it in in dc, but I put a link in here to a video Palantir's CEO talking about what kind of his vision of what AI's role in and what really defense's role military defense's role is in the world we live in today. This is Alex Karp, the CEO of Palantir.

1:20:15 - YouTube
Americans are the most loving, God-fearing, fair, least discriminatory people on the planet. They want to know that if you're waking up and thinking about harming American citizens, or if American citizens are taken hostage and kept in Dungeons, or if you're a foreign power sending fentanyl to poison our people, something really bad is going to happen to you and your friends and your cousins and your bank account and your mistress and whoever was involved. And you know when Americans are spending a trillion dollars on defense. What I know, what I want and what I think my peers want is why are these people keeping our citizens hostage, torturing our people, attacking our allies?

maligning us in what was once called the United Nations, basically a discriminatory institution against anything good. We need to stand up and those people need to be scared. And that's why this conference is so important, because we have the best products in the world and we cannot have parity. Our adversaries do not have our moral compunction. If it's even, they will take advantage of our niceness, kindness, our desire to be at home in Nebraska or New Hampshire or wherever we live in our peaceful environments, and they need to wake up scared and go to bed scared. And if you give that to the American people, the American people will go back and say and honestly probably shouldn't say this is why I thought the Democrats are going to lose the election and why they did because people want to live in peace, they want to go home, they do not want to hear your woke pagan ideology. They want to know they're safe, and safe means that the other person is scared. That's how you make someone safe.

1:21:52 - Leo Laporte
Safe means the other person is scared. This is the nut job. That's the telling make someone safe. Safe means the other person is scared. This is the nut job.

1:21:58 - Benito Gonzalez
That's the telling part, yeah.

1:22:00 - Leo Laporte
He's a nut job. Oh yeah, oh yeah. So I was trying to give him the benefit of the doubt. You may remember, I read his book the Technological Republic. I thought it seemed fairly reasonable in the beginning, it goes off the rails. Much like this as it continues. This guy just got $ billion dollars from the united states army to put ai into our military and other governments.

1:22:22 - Jeff Jarvis
Germany is having huge discussions right now because they plan to use it in policing their germany plus israel, plus other nations. He's a nut.

1:22:32 - Leo Laporte
Uh-huh, it goes on.

1:22:34 - Paris Martineau
By the way, I cut it off because I couldn't take any more, but I think you got the the flavor of it, yeah my, my first thought is if this is the sort of thing he's saying on stage at a conference, what are the things he's saying?

1:22:48 - Vlad Prelovac
you know, this is in public in front of cameras behind closed doors.

1:22:52 - Jeff Jarvis
Yeah, I met him in uh vienna. We were spoken the same at the same, he was a student of Habermas.

1:22:59 - Leo Laporte
Well, I don't know what that proves he studied, uh, he studied Goethe, I mean he's, he seems like an intellectual, always the IQ's there, same with teal but how is it used?

1:23:15 - Benito Gonzalez
That doesn't mean anything about morality.

1:23:17 - Leo Laporte
Oh no, I feel like it should. Maybe that's my mistake. All right, let's take a break.

1:23:27 - Jeff Jarvis
But he so just one more thing about. So his view on regulation was yes, regulate us, and then anything you don't regulate we are free to do yeah see, that's an issue yeah that.

1:23:43 - Leo Laporte
No, we're expecting you to use your good judgment.

1:23:46 - Jeff Jarvis
No, he says that's your job as a regulator you tell us what we can't do if you don't tell us everything else is fine and dandy, and we have no responsibility because you didn't tell us not to exactly.

1:23:55 - Leo Laporte
You didn't tell us not to exactly, didn't tell me not to.

1:23:57 - Jeff Jarvis
That's a teenager's point of view that's what he said on stage in vienna.

1:24:01 - Benito Gonzalez
That's a teenager hey, you didn't tell me not to right right well, the whole, the whole my enemies need to fear me thing is teenager thinking too, like it's all teenager thinking all the way down yeah, yeah, yeah, if you're not scared of me, I'm not safe if you don't go to bed scared, if you, your children, your mistresses, your family don't go to bed scared, afraid that I'm gonna come and kill you.

1:24:23 - Paris Martineau
Then I'm not safe I don't think we should bring my mistresses into this what did he bring that up?

1:24:28 - Leo Laporte
that was telling all right, we're gonna take a break. I just when I saw that, I was like, oh my god, what brought you to watch it? Uh, it was on reddit. Uh somebody it's it's, by the way, it's fairly old. He did it it's a few months ago from the reagan national defense conference. He was talking in front of uh like-minded individuals. That's why I think he felt free to to be so open about the reagan defense conference yeah, you know, someday we're gonna have a golden dome to match the golden ballroom.

It's hard not to cry. All right, I want, I'm gonna take a break. I don't want to get political. I know people don't like it when we get political, but I think it's important on an ai show to talk about the people who are in charge of the AI that our government is now buying. They're also being used to decide which regulations to cut right. They brought in Palantir to go through all the federal regulations. They want to cut half of them and the AI will pick which half. I guess if it doesn't make people scared, we don't need it. Yeah, he is a very rich nut job.

1:25:44 - Paris Martineau
That's the problem with a lot of these guys they're so rich they're untouchable kind of like the old internet adage that uh, the age in which you first find large commercial success is the age at which you stop maturing, uh, socially, yeah personally yeah, that might be true.

1:26:02 - Leo Laporte
I think these guys stopped maturing a little earlier than that even well, they all just started, they started with money, so you know they went to a freshman seminar and they talk at that level.

1:26:12 - Jeff Jarvis
Did they all started with kind of feels that?

1:26:14 - Benito Gonzalez
no doubt that they all started with money. Of feels that. No doubt that they all started with money.

1:26:18 - Leo Laporte
I don't know if carp uh started with money. I don't think so. Let's look him up. I don't think he did. I don't know what his history is. Uh, martin from Germany says, and he has some experience with this if we don't get political, the wrong ones make the politics. Thank you, martin. Martin, I learned a new german phrase. I don't know how how common it is. Was it martin that taught me this? Oh, what was it? It was a schnachmor, which means, uh, snooze mouse or something like that. Which means snooze mouse or something like that. Something's really boring, it's a schnachmörer. Have you ever heard that phrase before? I have not. Okay, jason Calacanis, weighing in in our Discord, says he's glad Alex Karp is on our side. Yeah, that's probably true. That's one way to look at it. At least he's not making us scared, or is he? It's not Jason Cal scared, or is he? It's not jason calacanis? Oh, yeah, that is, it is. Yeah, absolutely. Oh, hi, jason drops in all the time.

Uh, let's take a break, because jason had have you had carp on your show, jason, oh, I bet he has have you have you had him on all in, got it schnachmer, you know it.

1:27:41 - Paris Martineau
Martin, right, what does? What does that mean? Snooze, snooze mouse, snooze moose. If you give a snooze a moose.

1:27:45 - Leo Laporte
Give a moose a snooze Schnachmör. You can add that, Jeff, to your.

1:27:50 - Jeff Jarvis
German vocabulary yes, my poor high vocabulary.

1:27:52 - Leo Laporte
Snore. Oh, I'm sorry, a mör is a carrot. It's a snore carrot. That's better If you give a snore. Oh, I'm sorry, a murderer is a carrot. It's a snore carrot. That's better. If you give a snore a carrot. I can't pronounce it right, but uh, god, now we have a second title after nine seconds of google. I don't know. We're gonna.

1:28:08 - Jeff Jarvis
This is gonna be so jason just said that he hasn't had him on, but he will be at the all in summit. Ah, he's fascinating, by the way. By the way, jason, when you say that he's on our side who is our?

1:28:21 - Leo Laporte
who's on our side?

1:28:22 - Jeff Jarvis
he's on their side, he's on my side anyway let's take a break, but I'm scared so it's working.

1:28:30 - Leo Laporte
It's working. You know, I, I I have to say, sometimes this stuff keeps me up at night, but not last night, last night the best ad segue ever and I just ruined it last night I slept on our sponsor, helix sleep, and I slept the sleep of dreams.

I am telling you this is the. So I did this. Actually, perplexity told this. I said I asked the ai. I said, uh, when is it time to get rid of our mattress? And it said well, it depends, but six to ten years is sometime in that time frame. I said, oh my gosh, we have an eight-year-old mattress. So I started looking and I found helix sleepcom, and I am so glad I did.

Your mattress is more than just something you sleep on. By the way, I think you know this, we spend more than a third of our life on our mattresses. I mean, my favorite thing is to curl up with a good book, a nice comfy mattress. We got the adjustable one so I can sit up a little bit. So nice Movie nights with your partner he's lying back watching a great tv show or morning cuddles with our kitty rosie, or just, you know, winding down after a long day. Light a candle, put some nice music on your mattress is at the center of it all, really. It's, it's underneath everything, right, but if you have a bad mattress, a saggy mattress, an old mattress, it's not helping you, it's not helping your life. It's not helping your health. Maybe you're waking bad mattress, a saggy mattress, an old mattress it's not helping you. It's not helping your life. It's not helping your health. Maybe you're waking up in a puddle of sweat or you've got a achy back from a sag or you feel every toss and turn. Your partner makes a classic mattress nightmare. But that changed with my helix sleep. We just got it a couple of weeks ago no more night sweats, no back pain, no motion transfer. You can get the deep sleep you deserve.

One buyer recently reviewed the helix and this is, by the way, what sold me. I saw the reviews with five stars saying I love my helix mattress. I will never sleep on anything else. I second that emotion, by the way. I recommend it. We got the mattress and then we got this lovely topper and it cools you off. It's time and time again.

Helix Sleep remains the most awarded mattress brand. Just this year alone Wired's best mattress of 2025. Best Good Housekeeping's betting awards 2025 gave it the award for premium plus size support. That's nice, because I'm not a light guy, so I need a little bit of support there. Gq Sleep Awards 2025, best hybrid mattress. New York Times Wirecutter Award 2025, featured for plus size. Oprah's Daily Sleep Awards for 2025. Again, best hotel-like feel and actually as silly as that sounds I can think of. There have been a few hotels like the, the raffles in singapore or the mandarin in dubai, where I just go. This, this is the best bed I've ever slept in. Now I've got that in my home with my helix sleep.

Go to helixsleepcom slash twit 20 off site wide during the labor day sale early access. Helixsleepcom slash twit 20% off site-wide during the Labor Day sale early access. Helixsleepcom slash twit 20% off, and that's site-wide. The offer ends August 14th, so don't wait. Make sure you enter our show name after checkout too, because that's important. That way they know we sent you and if you're listening, after the sale ends august 14th 2025, be sure to check them out. They've always got something going at helixsleepcom slash twit. Do what I did read the reviews, then replace your old mattress with something that's guaranteed going to give you the best night's sleep you ever had helixsleepcom slash twit I just started shopping for a mattress, and it is hell to shop for a mattress, oh, trust me, I wouldn't tell you, I wouldn't steer you wrong.

Well, you know what? Okay, there's an. There is a bad thing about the mattress business. First of all, most of the mattress stores in your town are owned by the same company. Yeah, there's like all different names, but they're all the same company and then across across stores.

1:32:35 - Jeff Jarvis
It's the same mattress but they change it so you can't compare.

1:32:38 - Leo Laporte
Yes, you can't go to consumer reports and find out because that's a local san francisco thing.

1:32:43 - Benito Gonzalez
Right, which one mancini. That's all this. No, is it?

1:32:48 - Leo Laporte
all the same that's not.

Yes, okay, I don't know so it turns out it the mattress business is a very well read. I read. I read a good article about this somewhere, I can't remember. There's a fascinating business the, the stores, the storefronts, the brick and mortar stuff not where you should go to get a mattress. I know people think, well, I should lie on it, but that isn't gonna. I we've bought the worst mattresses ever because lisa and I we went to the store. I thought she wanted a really soft mattress. She thought I wanted a really firm mattress. So I got on a really soft mattress, which I'm not a fan of, to be honest, and I thought this is what she wanted. So I said, yeah, this is great. We got it home. Four years later. Long leo, I'm mr get along. Four years later, she says god I hate this matter.

Says wait a minute. Her. She says god, I hate this matter. Says wait a minute. You said you wanted us off. She said this is a terrible mattress. Said I'm sorry, I thought this is what you wanted. Don't. Don't go to the store with your spouse.

1:33:46 - Jeff Jarvis
You go together to chat gpt for marital counseling about mattresses we should fix.

1:33:51 - Leo Laporte
Well, what we did with helix sleep is they ask you like are your side sleeper, front sleeper? They ask you your style and what you're looking for and we really went through the whole thing and it really recommended the right one.

1:34:00 - Jeff Jarvis
It was really good, I've got it open right now.

1:34:02 - Leo Laporte
Yeah, take a look. Okay, so there is and I think we're going to cover this Thursday on Tech News Weekly. Micah says he wants to cover it. It's a tough one.

Uh, cloudflare published a blog post about perplexity. Now, remember when we had a cloud flare on, we had John Graham coming on. He's no longer their CTO, but he was. He's on the board. He talked about this new tool they had just added which allows sites to block. See, there is, in theory, this norm in the web called robots dot text, where all crawlers Google's and everybody else is supposed to look at it and see if they're allowed and if they're not, say, okay, I won't crawl your site. Uh, and the ai bots are supposed to honor it too, but they don't. So perplexity stepped up the war and has a much more aggressive uh, no crawl tool a blog post at cloudflare saying perplexity is using stealth, undeclared crawlers to evade our no crawl directives.

When I read this, I was very disappointed. As you know, I'm a perplexity user and fan and I thought geez, that's, they might have to give up on the perplexity. Not so fast, not so fast. Perplexity fired back and said what are they? This is a complete misunderstanding of how things work. But what? What cloudflare did is they set up some dummy sites, set the strongest blocks they could and then went to perplexity and said tell me about this site. And then went to perplexity and said tell me about this site, which perplexity did just as your browser would do if it. If you asked it to show the site, cloudflare took this as willful, ignoring willfully ignoring their directive not to scrawl the site.

Perplexity said we aren't crawling the site. We are not taking information from the site and training on it. We went there because the user said tell me about this site. They say and this is their blog. This is fundamentally different from traditional web crawling. We're fetching content when a real person requests something specific and they use that content immediately to answer the user's question. Perplexity's user-driven agents do not store the information or train with it. So I don't know who's right here. I don't have enough technical information, so I have things to say. Yes.

1:36:43 - Jeff Jarvis
One, that there's nothing to say that perplexity has to follow someone's robotstxt. It's a courtesy, it's a norm. There's no law, but it is a norm. It's a norm, I'll get there. I'll get there. Two, my browser going to a site is my agent going to a site, if you're talking about them training on it. But if I, if I ask my, if I ask perplexity to do something, it is a newfangled browser doing that and, um, I think that, uh, they're trying to cut off three. Three who died and made whatchamacallit? Who's the company? Again, cloudflare, cloudflare cop to the world. Now we trust Cloudflare to say well, we're going to get rid of spammers, we want to get rid of Macedonian campaigns, that kind of stuff. But Cloudflare, on its own, unilaterally, just came along and basically declared AI companies evil furthermore, they said that the Cloudflare was obfuscating their searches by using something called browser-based.

1:37:54 - Leo Laporte
Cloudflare say yeah, we do use browser-based very occasionally, but Cloudflare's misattributed three to six million daily requests from browser-based to us. That's not us, we only use it a little bit. Uh, for specific reasons. They said this is really cloudflare going for publicity. Uh, they also say even more embarrassing cloudflare published a technical diagram supposedly showing quote perplexity's crawling workflow. That bears no resemblance to how perplexity actually works. If Cloudflare were truly interested in understanding the data they were seeing, how our systems work, or these fundamental concepts outlined above, they could have done what we encourage all perplexity users to do just ask. So I don't know. This is kind of beyond my pay grade. Is Cloudflare?

using a power that it has well unilaterally, which is what bothered me about it we've all stood for the right I think we have of a, of the right to read right that, especially if it's just paris is making faces, and everything I said and everything you're saying, okay paris. Tell us what you think.

1:39:05 - Paris Martineau
I mean I don't disagree with what you guys are saying. I disagree that we all agree on the right to read as it relates to corporate products, but I mean it seems like perplexity's argument here. I think obviously the argument they're making with saying, oh, a lot of the traffic you're noticing is user agents, yes, that's an important caveat and should be a part of the story. But Perplexity still said, yes, we are like scraping some websites that aren't supposed to be scraped, using, like by ignoring kind of this robottxt, but it's only in limited capacity, for limited things, and that means it's fine, and I for limited things, and that means it's fine, and I don't know that means it's fine. I think that if people want to complain that perplexity is violating norms and perplexity's argument is, oh, we're only violating the norms occasionally, in limited ways, I mean you're still violating norms.

1:40:01 - Leo Laporte
Well, yeah, I mean. So somebody made the point. I think it's a good point. This is from from a blog by Rob Knight in which he says the real issue is that perplexity should respect consent. I don't want Google's Chrome to visit my site and I had some technical means to say Chrome, you can't. If you're on Chrome, if you're using the Chrome browser, you can't look at my site. Yeah, I guess I should have the right to say that it's kind of antithetical to the idea of an open web.

1:40:38 - Jeff Jarvis
Exactly, it becomes discriminatory. And what if you said I want no EDUs to read my site?

1:40:44 - Paris Martineau
Right, you could do that, it's your right, but I would think you're a schmuck for doing it yes, but I think like you can make a decision that makes you look like a schmuck and that's okay, I think jeff and I are kind of old school because we're old pitching the idea of the open web like when you put something on the web.

1:41:04 - Benito Gonzalez
Okay, wait, the truly open web, though you're gonna this let's go ahead.

1:41:14 - Paris Martineau
I was saying I think it's incorrect to conflate someone being like I want you to respect the norm of if I put a robotstxt thing there, to respect the norm of not scraping my website, it's. It's wrong to conflate that sentiment with yeah, screw the open web, the open web shouldn't exist. I, I think. I think those are two totally different.

1:41:31 - Jeff Jarvis
Leo uses perplexity, asks a question that goes to a website. Should it then come back and say leo, I can't do that because this website won't let me?

1:41:38 - Paris Martineau
yes, that's what chat gbt does all the time with me, and so that would very much that would very much make perplexity less useful.

1:41:45 - Leo Laporte
Because let's say, I did that search about my supplements and it and the that uh, one of the best supplement sites let's say it's webmd said, yeah, I don't, I don't want per a.

I take them to the worst crack too then all I'm going to get is stuff that that isn't good, right. Right, because the good site said, well, yeah, but I don't want anybody with an ai. And this is going to become more of an issue, because it's not just perplexity, it's going to be agents, it's going to be mcp, it's going to be.

1:42:09 - Paris Martineau
Perplexity does not have a right to everyone else's websites for free, ad nauseum. I think that perplexity has never right to be as useful as firefox firefox is being operated by people in most cases and I think, as that changes, people are going to try and uh set different norms when leo asks for something to be asked he's a person asking for something, he has perplexity for it and it goes and gets it.

It's not the computer it's a different kind of browser I think website operators disagree and someone who's decided to create a website, I think can uh make decisions about how they want it.

1:42:51 - Jeff Jarvis
They can, but I.

1:42:52 - Paris Martineau
I dislike them for their discriminatory nature, but I don't think that means it's like morally wrong or should be made like illegal it's wrong to the, to the.

1:43:03 - Leo Laporte
It's wrong to the ethic of the web yeah, I don't think it's made illegal either, I agree with you. I'm not saying it's illegal, but I don't think there should be a law about robotstxt, but either way I do think that there is either there is an unwritten norm that says the web is open and if you put something on the web, people should be able to use whatever browser they want. And that's what I'm doing with perplexity. That's the browser I choose to use I don't know, though.

1:43:31 - Paris Martineau
I mean, we don't do that, if you put like just because you're putting copyrighted material on a website does not mean that it's suddenly the property of everyone in the world it does not make it my property.

1:43:42 - Leo Laporte
I'm reading it as I would with firefox or chrome, or I'm reading it does make my property. How is it my property if? If perplexity was training on it, I would agree that's a little different, because that's actually the opposite there.

1:43:56 - Jeff Jarvis
I think training is what you have the most right to read and learn. Right training is transformative and fair.

1:44:02 - Leo Laporte
Use right rag and a direct response is um a different question I mean, google would also not be able to do snippets, and so, uh, every newspaper in the world would then say using the same technology, no googling I think we went through this.

1:44:22 - Paris Martineau
having no google snippets would probably be a positive for the world personally, but Well what this is.

1:44:29 - Jeff Jarvis
We went through this with Google and thumbnails, and the thumbnails were wrong and then we had court cases and said no, it's, it's for the purpose of search, it's okay. Uh, and and publishers tried to stop it and it was good for the. The publishers realized in the end it was good for them when the get ready Leister was short, was passed in germany. Uh, the publishers all agreed and said, okay, google, you can go ahead and scrape us. Uh, actual springer said no. And after 15 days they cried uncle and said no, we want the readers sorry that one was cut off.

1:45:03 - Leo Laporte
That was turned into a robot. Briefly, so honestly, paris, there's absolute merit in your uh this is a complicated issue and, uh, I I went back and forth. I read, you know, cloudflare's blog post. I read perplexity's blog post. I've read a lot of responses to it back and forth. I'm not sure what the right answer is. It's a very interesting question, but it's more complicated than what cloudflare said. Is, I guess, the point.

1:45:29 - Paris Martineau
Yeah, I think it's more complicated than what perplexity is saying too and it may well be Clint of uh DCN had kind of an interesting take on this in blue sky, which is he uh works for uh like a group that represents a lot of media publishers.

So of course, like he's got his own kind of perspective, uh, on all of this. That's worth considering. But I did think, like he commented on the PR aspect of this that he's like it's a bad PR move for perplexity to get all bent out of shape about this and it especially kind of reflects bad from a public relations perspective on the broader category of agentic AI, because there are the sort of companies that are going to be impacted by perplexity, making kind of big swing moves like this to say no, no, no, it's totally fine if we do this, perplexity is going to be fine either way. It's going to be the small, like agentic AI startups that are just trying to figure out how to do this themselves and actually trying to pay attention to all the complicated morality around this that I think are going to be hit by kind of the downstream effects of-.

1:46:36 - Jeff Jarvis
You see, paris. I think it's a terrible PR move by Cloudflare. I think Cloudflare is coming along and saying we're the cop of the web, we're going to decide what can go through and what can't when people trust them to be I mean mean that is cloudflare's business model no, well, yes but they literally now. They've now moved the bar a lot.

1:46:53 - Leo Laporte
They've said macedonians and ai are the same. In their defense, they would say we're just giving websites tools to express their consent or not, and uh, and we don't make the decision.

1:47:08 - Jeff Jarvis
And it's true, they're not making the decision uh, because the default is screw you ai right that's a decision.

1:47:17 - Leo Laporte
It's certainly a good conversation. We're going to take a little break. Uh, I don't know what the answer is. Uh, micah will be delving into this and I he has my sympathy. On Thursday, on tech news weekly uh, this is one of the big topics of the week. On our other weekly news show that we do every Thursday with Micah Sargent and a rotating panel of co-hosts, you're watching intelligent machines with a very intelligent pair of people. Uh, paris Martineau, now investigative reporter at consumer reports.

1:47:48 - Paris Martineau
I guess you're a journalist, because if it were a reporter, it'd be too redundant to be well, I think technically my title is investigative, senior investigative reporter at consumer reports, but I agree, repeating the word report and reporter within two words of it sounds ridiculous so I'd change it to journalist for this purpose, you've joined the department of redundancy department.

1:48:10 - Leo Laporte
Uh, also mr jeff jarvis, who is out of focus. I can get this out of focus. No, he is now uh running a program at montclair state university no, I'm actually not running anything there.

1:48:22 - Jeff Jarvis
I'm helping them out, but I'm consulting yeah are you?

1:48:25 - Leo Laporte
running a program.

1:48:26 - Jeff Jarvis
It's so helping hard programs at at at Stony Street, Nice, Very nice.

1:48:31 - Leo Laporte
This is intelligent machines. We do the show every Wednesday right after windows, weekly 2 PM Pacific, 5 PM Eastern, 2100 UTC. You can watch the show live on eight different streams. If you're a club member in the club, to a discord there youtube, twitch, tiktok, facebook, linkedin, xcom and kick wow. Uh, every week we start the show uh, this was kind of part of the new format with an interview. We're not sure who we've got next week, but we're working on it. We may have actually, we do have a pre-recorded interview. Maybe we should use that for next week.

1:49:09 - Jeff Jarvis
I think it's time.

1:49:10 - Leo Laporte
Tulsi Doshi is in charge of Google's models and she will be our guest. Let's do it next week, unless somebody else shows up. We recorded it already, so I think we need to use it because it was recorded last week. So, tulsi, let's say that. Let's say that, tulsi, dolce dolce doshi next week from google, talking about google models on intelligent machines. We're really doing, uh, our best to get interesting, smart people on the show to talk about ai, because you know what, as you can see from this conversation, it it's it's not clear, it's not clear. This is a brave new world we're part of on with the show we go, uh lava lamp owner.

1:49:55 - Paris Martineau
I will always come to cloud flares defense they use lava lamps to generate random numbers.

1:50:01 - Leo Laporte
By the way, uh uh, jason still had a very good tip for lava lamp owners yesterday on mac break weekly if the juice don't drink the juice, that's tip number one that open do not, although does have a bottle cap on it, it does kind of tempting in case of emergency what's in there? Is it water or oil?

1:50:23 - Paris Martineau
or delicious goodness. I think it's like a wax and some sometimes the wax balls.

1:50:29 - Leo Laporte
His wax balls got all the blobs, got all uh messed up into little balls and he thought it's ruined. He found that if you put it in a sous vide at a low temperature for several hours I don't know how it works they coalesce, they congeal back into the large wax balls that you love so much.

1:50:50 - Paris Martineau
There is also a lava lamp fan and appreciator web old school style web forum that I've previously plugged on this very podcast in my pick of the week. You can check the archives and in that there is one dude who's like the number one guy in the US that makes new lava lamp fluid, basically for anyone who wants it, and he specifically sells these little packs that if your lava lamp fluid is really messed up and the sous vide trick doesn't work, you can like, essentially in the same way that if your hot tub or pool is messed, is messed up, you can kind of shock it. You can kind of shock the lava lamp fluid with whatever he's got and it it sets it right wow so therapy nice.

1:51:37 - Leo Laporte
Uh, big story in the tesla world. Uh, there was a trial, an accident in 2019, a fatal crash in Florida. Guy driving Tesla under full self-driving he was using autopilot dropped his cell phone, assuming, I guess, that the autopilot would continue to navigate safely. He reached down to get the cell phone, but unfortunately the car continued through an intersection, plowed into a parked vehicle and sadly spun the parked vehicle and killed a pedestrian, seriously injured the pedestrian's companion. The pedestrian, the dead pedestrian's family, sued tesla and a jury verdict last week awarded them. I've seen different figures at least 200200 million. What's important about this verdict? First of all, almost always in cases like this, the lawsuit comes from the driver or the driver's family, and Tesla usually is able to settle before it goes to court. In this case, it wasn't the driver that sued, it was the dead pedestrian's family family that sued, so that was a big difference Tesla was unable to settle out of court they settled with the driver but not with the pedestrian right.

And the jury's verdict was interesting because it did assign two-thirds of the fault to the driver, a third of the fault to tesla, because tesla encouraged the driver to trust full self-driving and made no effort to make sure the driver was in fact paying attention, keeping the hands on the wheel which would tell me that the driver should be able to sue tesla as well but well, I don't know what will happen and there may well be an appeal, so it may not be over.

However, you know the driver said I trusted the technology too much. I believed that if the car saw something in front of it it would provide a warning and apply the brakes. This didn't happen. He hit the parked car at more than 60 miles an hour. It's a terrible, terrible, uh accident and you know Elon will probably say yeah, but look the driver's fine, you know it's a very safe car, uh. The bad news for Tesla, besides the lawsuit, is it did assign. It's the first time, by the way, as far as I know, in a court case that the responsibility was assigned to Tesla for this accident.

Furthermore, and maybe more shamefully, tesla, now that we've seen the transcript, did a lot to deny, obfuscate and even cover up what really happened. For instance, all Teslas are recording at all times. They have, in effect, a black box. Tesla immediately downloaded the recorded information from that car. The recorded information from that car, uh. However, when asked by the plaintiff's attorneys for that information said no, it's gone, we erased it, we don't have it. They did have it, wow. So they really did everything they could to cover up what really happened. Um, the plaintiff's attorney did learn, in fact, that that information still existed. Tesla said that today's verdict is wrong and it only works to set back automotive safety, jeopardize Tesla's and the entire industry's efforts to develop and implement life-saving technology. The plaintiffs concocted a story blaming the car when the driver from day one admitted and accepted responsibility. The car when the driver from day one admitted and accepted responsibility.

1:55:14 - Jeff Jarvis
Um 243 million dollars tesla owes the family, it will appeal. Tesla gave elon big bucks to stick around they have the money they just gave him.

1:55:22 - Leo Laporte
What was it? 29 million billion?

1:55:24 - Benito Gonzalez
I'm sorry yeah, billion dollars um and still every day I see, see Elon on Twitter being like hey, it works, full self-driving is here. Hey, full self-driving. Just today, yeah.

1:55:33 - Leo Laporte
He said these cars drive themselves.

1:55:35 - Benito Gonzalez
How is that okay?

1:55:38 - Jeff Jarvis
It should be. If anything should be regulated in this world, for God's sakes, it should be that.

1:55:43 - Leo Laporte
Yeah, yeah, so, um you know, I think really more um, what did you say, paris?

1:55:51 - Paris Martineau
your mic was low oh, I was just saying it's a very sad story yes, yeah, it's very sad for the family, obviously.

1:55:59 - Leo Laporte
Um, it's uh, I think, not acceptable that tesla uh actively obstructed the case. There are now at least 15 other active cases focused on similar claims involving Tesla accidents with autopilot. Three of those have been moved to federal court. I think this judgment perhaps, and I asked our car guy, sam Abul-Samad, on Friday about this, or, or Saturday we went to dinner and he said, yeah, this is, this is the, a break in the dike. This is just get ready. There will be a lot more uh because of this, um, but I have to say there is a Tesla makes a point. We do think, don't we, that it is safer to have these cars drive than humans humans technology.

1:56:51 - Jeff Jarvis
Not with, not with, just not with only what tesla has no humans cause uh massive deaths and accidents every day 50 000 people americans die every year

1:57:04 - Paris Martineau
this is my same argument that I make with the right to read. Yes, humans cause a lot of accidents, and that's terrible, but humans have rights, and just because something's terrible, we can't be like. All humans cannot drive. I mean, perhaps that would be an interesting policy solution to address the problem of fatal accidents on the road, but it's completely non-tenable given the way our society is organized. What does not have those rights are Tesla vehicles and cars generally, and so it's going to be a bit more of a complicated discussion to get those vehicles to a place where they have rights on the same level as human, as it should be, and I think the same goes for these ai systems.

1:57:48 - Leo Laporte
Well, here's one area that we can agree. Okay, airlines, including delta, are testing ai pricing in a way that no human can understand. This is a bloomberg story. That get ready for what one startup calls the exploitation phase.

1:58:10 - Paris Martineau
I'm shocked that this is a story, because I feel like they've already been doing this for a while and it's only going to be getting worse.

1:58:15 - Leo Laporte
They've been using very sophisticated pricing algorithms. For instance, frequent flyers this is from the Bloomberg story. For instance, frequent flyer this is from the Bloomberg story. Frequent flyers and consumer rights advocates often complain about the convoluted nature of airline ticket prices. Discount tickets can quickly become expensive through the use of drip pricing, extra fees for once, free amenities and a dizzying array of fare classes with dynamic prices that shift over time. Or even to know whether the price you're offered today and everybody who's booked a plane plane knows this is better or worse than the one you'll be offered tomorrow, who knows, it's a prison business model.

1:58:53 - Jeff Jarvis
It's the same as your phone, same as your cable company and the airline you got no choice, so many ways to get there and um. This should be regulated so, uh, it's.

1:59:03 - Leo Laporte
This is a paper authored by the co-founder and chief ai officer of a company called Fletcher, with two R's. It's an Israeli-based software startup that works with Delta and several other carriers. In the paper, the founder describes a pilot artificial intelligence program that Fletcher created for an unspecified partner airline. He describes taking a relatively simple pricing structure and replacing it with a head-spinningly complex one featuring many more fair classes, with prices that swing wildly from one moment to the next. So he's bragging about this. Yes, such pricing structures, he says, are so complex they quote, go beyond human cognitive limits. Great, you might say that this is how Bitcoin is priced right.

1:59:51 - Paris Martineau
I mean it's not with AI, but we're all talking about how Bitcoin should be the model for everything.

1:59:58 - Leo Laporte
The stock market kind of works like this right, it's up and down and up and down.

2:00:01 - Benito Gonzalez
Hey, Amazon is like this already.

2:00:03 - Leo Laporte
Amazon does all kinds of interesting pricing already. Amazon does all kinds of interesting uh pricing I. I remember years ago a friend of mine who sold um, uh, I think it was an air purifier on amazon that he branded. He was a radio talk show host and he would put it on amazon and then the bots would come see what his price was, buy it. I guess he might have been white labeling something from china, so get something this.

2:00:29 - Jeff Jarvis
So a dj had a branded air purifier. Yeah, it's a long story. Oh, you're missing things here. You should be selling.

2:00:38 - Leo Laporte
Oh man, I am missing out on all this oh god, you gotta put your face on like a dishwasher I should be making sandwiches in new york city for crying out yeah, by the way, that's a terrible business, poor hank. I will say.

2:00:54 - Paris Martineau
I took a brief dive into the reddit forum searching salt hank's name and deep, deep in the net. Don't never tell hank this. Deep, deep in the comments a couple people are being like well, you know, his dad is a famous tech podcaster, leo laporte, so he's basically like a sandwich nepo baby oh, no, no it's not true. I was like how do you think that your father, being a tech podcaster, helps you in any way sell sandwiches in new york city?

2:01:21 - Leo Laporte
in fact, he never even used my name he never called himself.

2:01:26 - Paris Martineau
I only bring this up because it was like two downvoted comments in the tents of Reddit.

2:01:31 - Jeff Jarvis
I just thought it was so silly you also saw a young wonderful a young woman in the teenager subreddit saying is he married?

2:01:39 - Paris Martineau
does he have a girlfriend? Is he single?

2:01:42 - Jeff Jarvis
he's single. Ladies, you also found one. Tell about the other one you found oh I mean.

2:01:47 - Paris Martineau
So this all came up because, uh, I texted the boys group chat being like, hey, like this came across the I'm in the food nyc subreddit and they're all kind of persnickety about stuff. And one thing they've been a little persnickety about since launch is salt hanks sandwich thing. Initially because they were like, oh, 28 sandwich, no way it could be good. But little by little, slowly surely, I keep seeing posts crop up that are like I went to salt hanks expecting to hate it because the sandwich is so expensive and it's the best thing I've ever eaten and I think it's actually really worth the money and I take back all of my rights and I was like fill me up for the day they filled me up for the day.

I had half the sandwich and it was so filling and then I'm gonna have the rest tomorrow and it's the best.

2:02:28 - Leo Laporte
I drank the juice hank said he says he really bugs him because every review says the same thing. God, this is the best sandwich I ever had. It seems to be universal best sandwich I ever had. It's so expensive. He says they're always cast in shade. He said let's look at if I could make it cheaper. I would. If we made it one1 cheaper, we'd lose money. Yeah, it is as inexpensive as it can be. He was so happy one day he made $300. I told you he sold a couple of days ago $11,000 worth of sandwiches in five hours, but they're practically a cost the roast beef is expensive, it's the bread.

2:03:09 - Jeff Jarvis
the bread is the special bread.

2:03:11 - Leo Laporte
He's using the best roast beef from it.

2:03:13 - Benito Gonzalez
He's learning firsthand, like why in shitification happens.

2:03:17 - Leo Laporte
Well, he told me this. He said no restaurant makes money on the food. They make money on the bar. I said yeah, but you're just selling soda and limeade. He said we're going for a beer and wine license. As soon as we get that, we'll be fine. That's where the profit is. Uh, so he's never going to make it on the sandwich because he knows he can't make it.

2:03:40 - Paris Martineau
A 30 sandwich, I mean, he's as inexpensive as he can and also it's's very common in New York for sandwiches to be $20.

2:03:46 - Leo Laporte
I know $28 is not that different. He's part of Big Sandwich, it's also a giant.

2:03:52 - Paris Martineau
It's a big sandwich.

2:03:55 - Jeff Jarvis
I'm dying to have one.

2:03:56 - Paris Martineau
I'm dying to have one.

2:03:57 - Leo Laporte
You could feed a family of four on that sandwich.

2:03:59 - Paris Martineau
We've got to go at some point, jeff, you've got to just go wait in the line. I've never had it.

2:04:04 - Jeff Jarvis
I'm going out to lunch with Craig tomorrow. Craig Craig, craig Newark, and I would take him to Salt Hank, but there's no way we're gonna get in. Can't get in? No, no um.

2:04:16 - Paris Martineau
So you're gonna go to Hamburger America which, frankly, I need to go to Hamburger America too yeah, if I come visit you guys I'm gonna have to.

2:04:22 - Leo Laporte
We're doing both for two days, so I can yes, exactly twice and I don't think you could have lunch and dinner.

2:04:28 - Jeff Jarvis
I don't think I can do both. You'll have beef sweats like you've never had.

2:04:32 - Paris Martineau
Yeah, in fact we should add a third day and then go to coat the korean steakhouse, just so we have three days of meat sweats well, I want to go to what's this day.

2:04:40 - Jeff Jarvis
The famous guys, uh, who is it? The cheesesteak, because I'm a major cheesesteak fan oh yeah, that was in the new york same rundown, yeah, all right do

2:04:49 - Paris Martineau
a week and we'll do meat week leo, where we have meat and meat week. All right, I just want everybody to know we're pivoting this show no more ai, no more intelligent machines.

2:05:01 - Leo Laporte
it's me, it's just meat, all meat, all the time we're gonna call it, call it this Week in Meat, and it's going to be amazing.

2:05:08 - Jeff Jarvis
There was until recently a German magazine that was just called Beef.

2:05:13 - Leo Laporte
I thought you were going to say Schwein Sausage.

2:05:16 - Paris Martineau
Sausage.

2:05:18 - Leo Laporte
A little different, diverse. Okay, I would hate to change the subject subject, but this is an appalling story. Cnn interviewed an ai version he's not on cnn anymore. Oh no, where is jim?

2:05:36 - Paris Martineau
acosta, wolf jim acosta oh publishing to substackcom, I believe oh, is he a podcaster they forced him out because guess who doesn't like him?

2:05:49 - Leo Laporte
Yeah, no, that's right, I forgot that's right. So on the Jim Acosta show, jim interviewed an AI that was created by the parents of a son by the parents, though that's an important factor here.

2:06:01 - Paris Martineau
I know, but that doesn't.

2:06:03 - Jeff Jarvis
I know.

2:06:03 - Leo Laporte
I know, but it's kind of creepy. Did you watch it?

2:06:08 - Paris Martineau
I haven't seen it, so we need to tell them what we're talking about. It is a AI version of a kid that was killed in an act of gun violence. I'm sorry, this is just a Texas representative.

2:06:21 - Leo Laporte
It looks like he's AI but it's not him.

2:06:24 - Paris Martineau
Oh, I was talking about a different thing. No, no, no, no. I know this is it.

2:06:28 - Leo Laporte
It was the same story. I mean the same show. So the parents created an AI of their son who was killed in Parkland.

2:06:38 - Paris Martineau
Yes, so it was a AI-generated avatar of their son, of their deceased son. That was made clearly from kind of like one static image.

2:06:49 - Leo Laporte
It wasn't, you know, uh this is the father uh talking right now. Uh, did he actually interview this the son?

2:06:56 - Paris Martineau
he did interview the AI copy of the sun. It's kind of creepy.

2:07:03 - Leo Laporte
What do you think I mean? I understand why the parents might do it if it had been done without the parents.

2:07:10 - Jeff Jarvis
I would hate it, but the fact that the parents did it makes me say they're trying to keep this story alive the.

2:07:19 - Leo Laporte
The comments in on his sub stack are pretty positive, you know like um posting the uh. The link to part of this was a hard watch and I admit to being skeptical about using AI in this way. This is Jenny commenting, but this is clearly expression of love from the father towards his beautiful son is the first time I heard walking and I could feel the vibrancy and kindness in his voice. Ah, such a tragedy it's heartbreaking, no matter what it's heartbreaking.

2:07:47 - Jeff Jarvis
But I think the response here was reflexive. It was the same one I think we all had, but upon reflection, yeah, maybe it's not so bad.

2:07:55 - Leo Laporte
I see uh often in uh on this on the ai subreddits people talking about. I saw one guy said you know, I lost my dad a few years ago. Uh, I took all of the emails that he sent to me and had AI generate a new email to me and it reconnected me. It was very moving. There are interesting uses of this. I know it sounds like Black Mirror.

2:08:23 - Paris Martineau
I don't want to knock parents who are going to be in the process of grieving for the rest of their lives for doing anything. I think that obviously you can use an AI to look at this in any way you want personally him want to draw. The purpose of this interview was to draw attention to the problem of gun, school shootings and the issue of gun control and I think, if you want to do that, there are a lot of living survivors of school shootings or living people who are experts on this topic, either through personal experience, academic experience or otherwise. That could have been a better subject for an interview, but they obviously, I guess, wouldn't have been as sensational as. Look at the uh ai deceased uh teenager which I don't know you were gonna say a little.

2:09:26 - Benito Gonzalez
Know, I was just saying like this is a fundamental part of the human experience, is it not? I mean, come on like grieving for a loved one grieving for a loved one who's? Died like this is just like a fundamental thing that every human is going to go through.

2:09:40 - Leo Laporte
I always hate it when they put these on the news channels, though I just CNN is really does an awful lot where they, you know, really kind of tearjerker, they show pictures and they talk to the parents media man yeah, I understand the need for ratings, but it uh, it feels a little bit so line 113. I was going to end the show, but if you want to keep going, what do you think?

2:10:03 - Jeff Jarvis
about. Let's end it on on a morbid topic.

2:10:06 - Leo Laporte
Yeah, let's keep going death okay, the rise of ai, tools that write about you when you die. Um well, you know. Is this any worse than the reverend who talks to the family 10 minutes before he has to eulogize somebody that he never met, gets a couple of notes and and does a eulogy. Is this any worse than that?

2:10:29 - Paris Martineau
That happens all the time, right? No, because we're talking about obituaries like written, published obituaries that, at least in the case of someone like perhaps a reverend, someone involved with the funeral poem, perhaps someone who knew the person who died, if those people are the one writing the obituary, they're probably at least going to be accurate, or, if not, they'll have some sort of kind of human uh reason as to why, or a person that you can come back to and be like. Hey, actually my dead wife didn't like dogs and spend her time volunteering at the uh like animal shelter.

2:11:06 - Leo Laporte
But in this case, you're having basically AI tools that are generating thousands of obituaries every single year and they're obviously going to get stuff wrong the example, though, that they start with which is, I think, a telling example is somebody who lost somebody and who wrote a prompt, just kind of emptied their heart into the prompt, just kept writing, writing, writing. It's not a coherent obituary, but it's all our feelings and all that, and then the ai generated a coherent obituary out of that. That's a good. I think that's a good use. We're seeing this, incidentally, more and more. I see it on reddit all the time, where somebody posts my mom sent me this. My mom sent me this uh message about my divorce. Did an ai write this? And almost always you can say yeah, I think that's. Your mom used an ai. Is that a? Is that a bad thing?

2:11:58 - Benito Gonzalez
for the obituaries the obituaries thing, I don't think isn't that bad, because there are a lot of people out there who die, who no one will write an obituary for exactly, yeah, exactly, and not everybody's a writer, like you are a paris, or like you are jeff, and so maybe they have a hard time expressing themselves would you want to write you if you knew the date?

2:12:16 - Jeff Jarvis
would you want to write your own obit?

2:12:20 - Leo Laporte
no, no it's not for me to say yeah yeah, this is also also others to say how wonderful I was, what a charming and intelligent soul. It's for everybody else to say they're not for me this is also I feel about.

2:12:34 - Paris Martineau
I have no preference as to what happens to me after I die, because that's not for me, that's for other people yeah, I agree, I'm very wise.

2:12:41 - Leo Laporte
You're wise for your years.

She is, she is yeah, that's true, she just wants to get out of, you know, putting together the funeral program. That's on you. I don't care I'll be dead. Uh, we have a relative who is on death's door and has been for years and, uh, she, I have, she has, she has very elaborate, very elaborate plans and even gave me pictures and a song so I could make a moving video tribute to her for her funeral, which I did. It was nice, it was very sweet, I think it's. You know, for some people it's kind of a comfort to know that you know.

2:13:24 - Benito Gonzalez
Well, some people think they're going to be able to watch their own funeral.

2:13:27 - Leo Laporte
They ain't going to be able to watch it.

2:13:28 - Benito Gonzalez
Yeah, but that's what some people think.

2:13:30 - Leo Laporte
Yeah, I don't think she thinks that. Maybe she does, I don't know. There is a new Catholic AI app that promises answers for the faithful. It's actually an LLM based on 27,000 church documents. I wonder if father robert had anything to do with this. I was just thinking that, yeah, magisteriumai chat gpt for catholicism. The company behind it, long beard, claims up to,000 monthly users and proclaims its mission on its homepage in a huge font we're building Catholic AI. I got to ask Robert about this.

2:14:20 - Benito Gonzalez
I think the Pope would have something to say about that right. But what Is?

2:14:24 - Jeff Jarvis
it official.

2:14:25 - Leo Laporte
But what Is it official Today is? No, I'm sure it's not. It looks, by the way, just like chat gpt. Uh, here it is. I'm showing it on the screen. You wouldn't know. So today is the feast of the transfiguration of the lord. How can I help you today? Let me a reading for today's mass. So this is obviously for priests who don't want to have to come up with their homily. Here you go. Here you go. Here's a reading. Here's a responsorial psalm. Here's the second reading. Look at this, look at that.

2:14:58 - Jeff Jarvis
I could be a minister.

2:15:03 - Leo Laporte
Well, the readings are scheduled, aren't they? I think so. Oh, your sister's a minister. She would know.

2:15:07 - Jeff Jarvis
Yeah, so there's. I forget what you call it, but yeah there's the lectionary. Yeah, somebody in the chat room is holier than I am. They'll know.

2:15:17 - Leo Laporte
Yeah, yeah, that's right. That makes it easy, isn't it the world's number one answer engine for the Catholic Church? That's how they I had assume the world's number one answer engine for the Catholic Church is God, but maybe this is number two. Well, well, you can ask in prayer, but if you didn't have that option, you can go to magisterium.

2:15:37 - Jeff Jarvis
The AI voice by the way, building God Paris here's a holy widget, one of the holy widgets really good contenders for show titles.

2:15:48 - Leo Laporte
This holy widget is a confession guide, so here's what you should think about. Try to recall all your sins. You can use the below examination of conscience to help with this. Did you give God a time every day to pray? Did you seek to love him with your whole heart? Here's the prayer, then inside the confessional. Here's what you should say. Here's what you should do. Here's the act of contrition Nice. Who needs to go to the confessional.

2:16:16 - Jeff Jarvis
It's done.

2:16:17 - Leo Laporte
You got a holy widget. Does the whole thing for you.

2:16:21 - Benito Gonzalez
Can I send my agent to confession for me?

2:16:23 - Leo Laporte
Yeah, send.

2:16:24 - Jeff Jarvis
Chad GBT in my agent's been very bad actually you know what my agent is perplexity.

2:16:33 - Leo Laporte
It has sins to confess. This is actually pretty good, because there is a lot of ritual, there's a lot of text, there's a lot of information and and you might have forgotten you know the apostles creed, but if you needed it, you could go here and it's it's the apostle liturgical faith.

2:16:45 - Jeff Jarvis
But this is an oral.

2:16:46 - Benito Gonzalez
This is an oral tradition, though you should be getting that information from your priest or from another person oh, but who can remember, you know, do you?

2:16:53 - Leo Laporte
do you remember who the saints for today are? Well, if you're a real catholic.

2:16:57 - Benito Gonzalez
Your magisterium knows on sunday, so yeah saint sixtus, saints justice and pastor.

2:17:02 - Leo Laporte
Saints ormistas. Blessed octavian.

2:17:07 - Jeff Jarvis
Blessed shekelin was a hermit sounds like shecky green feast of the transfiguration you could ask my mom?

2:17:16 - Leo Laporte
I don't want to be religious but this is, I think, in its nature sacrilegious. So, uh, there you go, ai is. Is there anything it can't do?

2:17:27 - Benito Gonzalez
Taste, taste food, taste anything.

2:17:32 - Paris Martineau
Taste anything, yeah, taste. It can't taste non-food either.

2:17:35 - Leo Laporte
But you could build an AI that could be attached to a chemical, a very highly sensitive chemical analysis tool.

2:17:43 - Benito Gonzalez
This goes back to the perception question. Is green to me the same as green to you? To me, the same as green to you is the same as green to an ai.

2:17:49 - Paris Martineau
I can only tell you, you know what photons are hitting my retinas can ai taste this spindrift, and if it does, will it taste the same to the ai as it does to me?

2:18:02 - Leo Laporte
but I can tell you something ai can do it will say that's pretty bland.

2:18:06 - Jeff Jarvis
Can't you just get something with a little more flavor?

2:18:09 - Paris Martineau
they'd say this is the perfect uh seltzer it can create bunnies on a trampoline oh, did you guys get got with this? You thought this was real okay, there was a brief moment where I did think there were bunnies on a, on a trampoline, yes, and then I looked at the comments if you can't trust this, what can you trust?

2:18:33 - Jeff Jarvis
it's a whole genre of animals bouncing on trampolines and it's great. I think this is so good like they were having a fun time.

2:18:41 - Leo Laporte
I'm just sad that it's so short it should go on there's also kangaroos on traveling

no, really yep I, I can't imagine. So. The person who posted this who's? Rachel the cat lovers, which is a little weird, uh, because it should be the cat lover. Guys, the rabbits are real, real taken down. They are not real, rachel. She says just check the home security cam. You won't believe what I saw. It is kind of believable, isn't it? It is the only thing that's weird. They're around the rim of the trampoline, they get on it and they start to bounce. But you know something's weird when they stay there and you keep bouncing like, oh hey, this is fun. She was smart to cut it off because that makes you feel like it's more real I just put up deer on the on the trampoline.

2:19:36 - Jeff Jarvis
Is it real in the? I don't know it's a I can't what.

2:19:40 - Paris Martineau
What's real now it has to be in the.

2:19:42 - Jeff Jarvis
In the chat I put it we don't know anymore, do we?

2:19:48 - Leo Laporte
in the chat I put it we don't know anymore, do we? We no longer know what a world we live in. Aren't we excited, aren't we happy? Ladies and gentlemen, we're going to take a break right now to think about our blessings this whole uh delivery, and then we're doing unboxing okay, folks, I got the ultimate here.

2:20:08 - Jeff Jarvis
Wait, wait, wait.

2:20:09 - Paris Martineau
We have elephant on trampoline, that's a deer, but oh, that's a really good one the deer starts bouncing and then wipes out so.

2:20:22 - Jeff Jarvis
So go to the. Go to the one.

2:20:23 - Leo Laporte
I just all right this is from the real. The AI mystery Elephant on a trampoline Collapses.

2:20:37 - Jeff Jarvis
I do really like that.

2:20:39 - Leo Laporte
I'm just saying I don't care if it's AI or not. It gave me joy. It gave me joy. And, incidentally, that is a perfect example of a human being who could never create that video, writing a text prompt and using ai to create a hysterical video. He had the creative idea, the impulse and made it happen with ai. It's just a tool, that's a great use of it.

2:21:03 - Benito Gonzalez
Sure it made you chuckle for two seconds and you'll forget about it for the rest of your life which is okay.

2:21:07 - Jeff Jarvis
What's wrong with that, you think, henny?

2:21:09 - Leo Laporte
youngman may be left, for I'm not saying anything's wrong with that?

2:21:12 - Benito Gonzalez
I'm just not going to ungrandize it as much as you are yeah, it made a funny video.

2:21:17 - Jeff Jarvis
It's a creative, it's cool, but it's not the greatest thing ever no, I'm just saying it's a creative tool neither is saturday night live, but it should have all the rights that people have, and you're sounding grumpier than ever?

2:21:30 - Leo Laporte
yeah, he is. You're turning into a curmudgeon way before you.

2:21:33 - Paris Martineau
He's got a balance, he's been one. He's got to balance out this podcast.

2:21:38 - Benito Gonzalez
Hey, I actually used ai this week, so you know oh, oh, what did you tell? I just used notebook lm to learn a new language, a new, a new, uh, like audio. There's like this deep audio program that's like all it's called pure data and it's really hard to learn and I'm using notebook to learn it.

2:21:57 - Leo Laporte
Yeah, to teach you. Yeah, are you not impressed and blown away with? What a great tool we didn't have that.

2:22:04 - Benito Gonzalez
It's cool. Yeah, it's cool.

2:22:06 - Leo Laporte
But it's not like tools, mind-blowing, it's like, yeah, that's cool well, we can also make you pay a lot more for that trip to the philippines you're going on. Just go to delta airlines and see what price you're gonna get delta. Yikes, I don't think delta goes to the philippines. All right, we are going to now take a break. Uh, and thank you all for being here and invite you to join Club Twit for ad-free versions of everything we do. Special events coming up this week Stacy's Book Club Friday, 1 pm Pacific. Right after that, chris Marquardt's monthly photo show. It's going to be a lot of fun. We answer your questions, we take a look at your photos and we give you a new assignment, as we do every month. We've got other things coming up, including the Google Pixel 10 announcement. We're going to do that in the club.

All the keynotes now are club only to avoid being taken down. I think for 10 bucks a month you're getting some real value. But the most important thing is you're supporting what we do If you enjoy the content we create. Content can't be, isn't't free. We've been ad supported for a long time and that was fine, but ads no longer cover the entire cost and we have cut way back. We shut down the studio, we canceled shows. We had to lay people off and still we operate at a deficit. Fortunately, the club covers that 25, that the advertising doesn't cover 25 of our operating, our operating costs. Thank you, club members. And if you're not a member and you'd like to join, or at least take the two-week trial and see if you like it. Twittv slash, club twit. We really appreciate our club members. Thank you. You make a huge, huge difference. Now it's time for the picks of the week and we always start with Paris, martin, no, paris.

2:23:51 - Jeff Jarvis
She always said start with paris martin. No, paris, she is away. Oh, maybe jeff jarvis, while we're waiting, I I made a pick for you, leo line 129 leo bait. I love it when you drive you crazy.

2:23:57 - Leo Laporte
I love it when you do that. Uh, my problem is the numbers are so tiny. Ai, not ai vegans, ai vegans. This will drive you not oh, these are people that and we know some of them who abstain from ai for environmental, ethical and personal reasons. She doesn't really fine.

2:24:19 - Jeff Jarvis
She doesn't look happy, yeah if you don't want to use it, please. Shouldn't she be happy now because she's off of ai? She's, her brow is furrowed, yeah she looks like she was made by AI, by the way it does.

2:24:30 - Leo Laporte
But it says it's a photograph, so it's from Getty, so it's a stock photo if you can't trust the rabbits on the trampoline, what can you trust? Ai vegans. See, this is the thing.

2:24:40 - Benito Gonzalez
This is the thing that I am actually against is being extreme on either side of this thing yes, being like AI is the best thing in the world, or AI is so useless you should never use it. Both of those positions are completely wrong it's fair.

2:24:51 - Leo Laporte
You know what I could see. I look there's people who say you shouldn't use a computer, you shouldn't use the internet, you shouldn't use smartphones. That's fine. If that's what your choice is, you probably have a happier life. To be honest, no, 1920, einstein, time is relative. 2025, delta airlines prices are relative. Thank you, martin sassenberg. In our youtube chat, paris your pick of the week and it better not be that soda you're popping my pick of the week.

2:25:19 - Paris Martineau
I just got from outside, oh I've got a couple things.

2:25:21 - Leo Laporte
It came the thistles here.

2:25:24 - Paris Martineau
Uh, this is in. I I'm just a lowly thistle user who has used this product exactly for one week, maybe this is not a thistle ad it's not a thistle ad, in no way. It's sponsored.

I literally just we have had food, a number of food box so it's like a um prepared meal service so you don't prepare it no, you don't prepare it, it's so these look good it's, it's. They're really good so far. This one's a thistle pad thai with sesame ground. Um you uh prepare it by heating it in a skillet with the sauce for three to five minutes. They're also their main thing is they do like little dinners and they do like kind of fun salads for lunch, and they all come like freshly made.

2:26:09 - Jeff Jarvis
Did you get the peanut butter bar? That looked really good today.

2:26:13 - Paris Martineau
I did not, but I got a affogato smoothie with biscotti crumble.

2:26:19 - Leo Laporte
Now do you have to go every once in a while and say this is what I want next week, like, do you have to?

2:26:25 - Paris Martineau
You can just have it set up, put like what you want, that you want like one or two breakfasts, two lunches, a dinner, a snack or whatever and set it and forget it and they'll pick stuff automatically from their menu. But I am kind of a picky person so plant-based only you can have plant-based. Only you can do meat, so I um meat only twice a mix of plant.

As far as, the protein goes you can do plant-based protein or meat oh I see um and uh, so I just go through like I just did it before the show for my sunday, I want to do this and well, if you're going to do it, use my referral code, oh yeah, it's not an ad, but as a lowly user I do get a referral code and you get 120 off your first week or something with my discount code, which is at line 163 in the chat and you know I used a friend of mine like a skeeball friend, uh, for the um for mine and so far it's been pretty good. I mean, I was literally talking to some friends today I was like this is really delicious, I enjoy it. Maybe the unit economics won't work out for me in the long term how much is it?

they. It depends, like if you get more it's uh stuff it's less, but it's like 15 uh thing, or like it could be like 13 or 10 or 18.

2:27:45 - Leo Laporte
So I just picked the default and it was $257 for six days, so 18 meals total, including snacks. I mean I think I could get by with less if I shopped and cooked yeah.

2:28:01 - Paris Martineau
I mean I think you also don't like I'm not getting it for every single meal, I'm getting like a couple a week. But I'm kind of busy right now. I just started a new job that I'm really enjoying, but it's just nice to not have to deal with cooking and I find these really delicious.

2:28:17 - Leo Laporte
Does it taste good? It does taste good.

2:28:18 - Paris Martineau
It tastes really good. Like I've done Cook Unity, I thought about trying Factor and things like that and I just felt like some of the frozen ones just feel too much like I'm eating a lean cuisine and I just really like that. This is I don't know consumer reports, uh rate these various services I have no idea and this is in no way related to my job or is it a? Reflection of anything relating.

2:28:41 - Leo Laporte
Yes, to be clear, to be very specific, um well, I have to ask the boss if I want to be part of that team but if you want to be part of that team.

2:28:49 - Paris Martineau
But I want to be part of the team that would rate these things. Oh my gosh, the testing team has got to have the best, like most interesting life ever, like we, uh well, sometimes do. I've learned the company sometimes does like auctions within the company itself of like the tested products we don't need this, uh Chevy Malibu anymore. They actually apparently do auction cars quite a few. I mean yeah, because consumer reports is really ethical about everything.

2:29:16 - Leo Laporte
They buy everything full price with. I did the same thing. We called it leo's garage sale. I take all the crap that I bought and put it in the conference room and people would fight over it. Actually, um, fight over it actually, um, oh yeah I don't know, but I've really might do this.

It looks good yeah if uh do it, I think they're available west coast, east coast, atlanta, chicago, maybe one other area, I'm forgetting thistleco. They do do it, uh, in my town apparently. Oh, I gave them my zip and they said sure yeah, I feel like they probably did over there.

2:29:49 - Paris Martineau
It's um. So far I've been really good. I would recommend, though, if you're kind of picky about food to like, actually go through the menu and you can select what you like and what you don't like. I've realized I don't really like the smoothies that much, or nor do I need the breakfast, but, um, yeah, it's been good.

My non-SpawnCon seeming though it isn't actually SpawnCon pick is two things related to Blue Sky, though. One is I've been kind of annoyed that Blue Sky doesn't have bookmarks, and some of the folks from Blue Sky posted that. There's like a featured feed where you can comment a pin emoji on someone's post and then it puts it in the thing. I hate that because I don't want anybody to know what I'm bookmarking, but I found in the replies for blue skies post about this that there is now um, a labeler run by some blue sky user called bookmark. Like it's bookmarks dot, blue canary dot dev.

It uses the modern, like the custom moderation feature of blue sky, and so I set up and it's perfect. I mean it would be better if bookmarks actually worked on the site, but it's a perfect kind of band-aid, so all you do is, if there's a post you want to bookmark, you subscribe to this labeler and then you just highlight a post, you go and you click like you're going to report it to Blue Sky moderation services. But Blue Sky moderation services is just one of many moderation services you can use, and so one of them is this bookmarks app. So you just report it for other or whatever and two bookmarks and it goes in your bookmarks feed I hope the blue sky doesn't doesn't judge that as if you are reported no, no, I mean because moderation feeds in blue sky, uh, and like labelers can exist for a bunch of different reasons.

Okay, good.

2:31:30 - Leo Laporte
So let me bookmark your post about your job, so I'm going to go here. Oh, I have to subscribe. I'm going to report the post.

2:31:38 - Paris Martineau
No, no, no. So you need to go back to the original link I sent. You have to subscribe to yeah don't report her.

2:31:43 - Leo Laporte
Oh, okay.

2:31:44 - Paris Martineau
So go to that post.

2:31:50 - Leo Laporte
And order. Oh, okay, go to that post and I gotta go to go to this bookmarks, so click up the top subscribe. Oh, I see subscribe to label.

2:31:53 - Paris Martineau
It's like you're just following something yeah, that was easy so then go to my post again, okay, before okay, how do I do that?

2:32:00 - Leo Laporte
let's see paris, and then I'm gonna go here now. I'm gonna report you now lady report me.

2:32:08 - Paris Martineau
I just click other, usually click other other oh, marks, so not moderation services. That's the one you don't want yeah and marks, and so you can just click submit report.

2:32:20 - Leo Laporte
That's all you don't have to do any surrender.

2:32:27 - Paris Martineau
Always fight.

2:32:27 - Leo Laporte
Never forget.

2:32:30 - Paris Martineau
That's great.

2:32:32 - Leo Laporte
Submit report.

2:32:33 - Paris Martineau
And so then, if you go to your feeds on Blue Sky, there'll be a.

2:32:39 - Leo Laporte
There's a new one called Bookmarks.

2:32:41 - Paris Martineau
Yes.

2:32:42 - Leo Laporte
Alright, where is it?

2:32:43 - Paris Martineau
Where is your?

2:32:45 - Leo Laporte
These are my feeds, so go to Discover Scroll down.

2:32:48 - Paris Martineau
Where is it? Where are your? These are my feeds.

2:32:51 - Leo Laporte
Go to Discover. Scroll down. Discover new feeds Just type in bookmarks.

2:32:55 - Paris Martineau
Maybe that's what you have to do. Maybe there's two steps.

2:32:57 - Jeff Jarvis
They cache a lot of blue skies.

2:33:02 - Paris Martineau
You might have to which one is this one. Bookmarksbluecanarydev.

2:33:09 - Leo Laporte
I don't think I had to do this, so maybe so this one you could dm to somebody, this one uh I would just refresh the page because, yeah, just refresh the page.

2:33:16 - Jeff Jarvis
It might be that I have to do that fairly often now you're seeing all my book, I guess just type bookmarks dot blue canary.

2:33:35 - Paris Martineau
This is odd. Well, hopefully this doesn't happen to you, dear listener, because this didn't happen to me, honestly well, it's probably just thinking it might be different on phone, on the app, than on the web.

2:33:47 - Benito Gonzalez
It might take a while for the web to populate.

2:33:49 - Leo Laporte
It's just thinking. I'm sure it's just thinking. That's a good tip.

2:33:53 - Paris Martineau
Yeah, I really like it um so now you have a new feed that is just stuff you've bookmarked yeah and it's great because otherwise, I've just been sending dms to my other Blue Sky account, which is just ParisMartineau Blue Sky dot, whatever, so that no one else can squat in my name, since I have a custom domain.

2:34:17 - Benito Gonzalez
Wait, isn't it one of the tabs at the top? Is there a new one up there? Those don't scroll no, it's.

2:34:29 - Leo Laporte
It hasn't showed up yet.

2:34:30 - Paris Martineau
It will, it will you'll get there, you'll get there uh, jeff jarvis, your pick of the week.

2:34:37 - Jeff Jarvis
All right, since we can't trust uh bouncing bunnies, I'm gonna do another animal story. Good guy, humpback whales. This comes from science, okay.

2:34:48 - Leo Laporte
Good guy humps back whales.

2:34:50 - Jeff Jarvis
So humpback, no Humpback whales. So humpback whales crash orca hunts as an instinct to rescue other animals and they don't eat the animals.

They don't do it. When they come in they say what did the Lone Ranger say? My work here is done and then they leave. Do you think it's true? I think so. It's scientists saying so, of course. Who knows? A study of 115 documented cases Of course I'm reading this from social media Says that nearly 90% of cases the humpbacks disrupted the attack. So the question here is whether animals have altruistic actions, whether these animals have it. Yes, these animals.

2:35:43 - Leo Laporte
You don't think this is AI, generated of the humpback.

2:35:46 - Jeff Jarvis
I don't know. Ferrying the seal to safety safety. The more I look at it actually, the more I think it's it really happened.

2:35:56 - Leo Laporte
If it did, that photographer really was in the right place at the right time.

2:35:59 - Benito Gonzalez
I think the scaling is all wrong. That humpback is too small alright, so I'm gonna.

2:36:03 - Jeff Jarvis
I'm sorry about that this is your bunnies yeah, it's the bunnies. It could be bunnies.

2:36:08 - Leo Laporte
It could be the story may well be true, I don't know. This is the problem.

2:36:12 - YouTube
You saw that there are now paper mills creating fake scientific studies.

2:36:20 - Leo Laporte
We're living in a world where I don't know if you could trust anything you read anymore. Just like the first days of print.

2:36:25 - Jeff Jarvis
Where I don't know if you could trust anything you read anymore.

2:36:28 - Benito Gonzalez
Just like the first days of print. And see, this is the problem with having all of our AIs sound like humans and making videos and stuff like that. This is the world we're making.

2:36:41 - Leo Laporte
Oh, I don't know. I'm just glad I have a good mattress because I'm going to go lie down. Thank you, jeff Jarvisvis. Professor emeritus of journalistic innovation at the craig newmark graduate school of journalism at the city university of new york, author of the gutenberg parenthesis, now in paperback. The web we weave and madagascar, now an audiobook. Great to have you. As always, jeff, as always boss, have a wonderful. Give craig newmark our regards tomorrow have a wonderful hamburger at Hamburger America.

Paris Martineau is an investigative reporter at Consumer Reports, where she reports on investigative reporting for Consumer Reports, so it's nice to know that you're a reporter.

2:37:19 - Paris Martineau
Are you a consumer? I want to report you.

2:37:21 - Leo Laporte
I'm a consumer, I'd like to report. She's at Parisnyc. She is awesome. You're both awesome. I love doing this show. I hate ending this show, which is why it goes on and on and on. Sometimes we do the show every, as I mentioned, every Wednesday. I hope you will watch live, I hope you will download it, I hope you will tell your friends and I hope you will join Club Twit. But I have many hopes and they are often dashed. It's okay, it's it's. It's okay, I'll, I'll be fine. Thank you for being here. We'll see you next time on intelligent machines.