This Week in Startups - TWiST News: The Startup Redefining Sonos, Influencers, Elections and Google's next AI push | E2034

Starting point is 00:00:00 The next one that was incredible in this buzz arrow was, hey, Alex Wilhelm's new book, cautious optimism, the book, has come out and you call all the bookstores and say, hey, do you have any copies of cautious optimism? I got to get this book. And they'd be like, let me check. Yeah, we have two copies. Hey, could you put one aside for me? I'm going to come pick it up tomorrow.

Starting point is 00:00:22 You never show up. But you do that a thousand times once a week, twice a week, to all the bookstores. What do the bookstores do? They order more copies. Now they got more copies. They're stuck with them. They put them in the mirror and they just created buzz that you go into the store and they were sending agents into stores saying, hey, do you have this book? Because my friend told me it's the greatest book ever to influence the people at bookstores who are the reps.

Starting point is 00:00:49 Wow. I mean, this is creative thinking, but it's sinister and it's against the FTC rules. And they came out with rules. You have to disclose. That's why you see paid partnership. a hashtag in a lot of paid partnerships. This Weekend Startups is brought to you by DigitalOcean. Whether you're just getting started with AI or seeing your project take off,

Starting point is 00:01:11 DigitalOcean is the cloud platform that lets you focus on what matters, building killer apps. Right now, get up to 100K in free credits at DO.com slash twist. Terms and conditions apply. Cloud devs. Building the best remote team is. tough, but you don't have to do it alone. Visit cloud devs.com slash twist for an unbeatable offer on hiring elite Latam talent today.

Starting point is 00:01:38 And Kelshi, the largest regulated predictions market now lets you trade on U.S. elections. Visit Kelshi.com slash twist to see live election odds, place a trade, and get $20 when you deposit $100. All right, everybody. Welcome back to this week in startups. I am your co-host, Jason Calacanis. I invest in startup companies here in the Silicon Valley, Austin, New York, basically English-speaking startups since I don't speak a foreign language with me.

Starting point is 00:02:10 Alex Wilhelm is my co-host here. Alex, how are you doing? Well, I'm good, although I think I need to find like the Duolingo FAQ got in wacky with it because you can solve that, Jason, and you can learn any language you want through the power of technology. No. if I had the time. It is one of my great regrets in life is that I didn't learn how to speak another language when I was younger. I wish some of that French had stuck. But, you know, it's, it is an interesting thing. People always want us to invest in their company in Japan or

Starting point is 00:02:41 India or France, Germany. And you really should go with a local investor who speaks the language, understands the culture, can use your product. And vice versa. If the person, doesn't speak English, can't use the product. It's really hard to give advice, right? And so we tell folks, like, we would, you know, this is a really interesting idea from China, from the Philippines, from wherever, Germany, a lot of times interesting stuff. And we say, just let us know when you launch the English language version or if you go after U.S. customers, if not, get those local investors to invest in the early stage.

Starting point is 00:03:17 I do appreciate that you're sharing Alpha with the European Metro Capital community, Jason. That's very generous of you. Before we get into the rundown, though, and we have an awesome guest today, just where everyone knows, I want to point out that this is earnings week. So when you see us later on, we're going to have data. Jason Tuesday, alphabet, Snapchat, Reddit, AMD. Wednesday is meta, Microsoft, DoorD, Robin Hood, Coinbase. And then Thursday, Uber. And Apple, Amazon, and Intel.

Starting point is 00:03:46 Yeah. So I am a shareholder in Alphabet, Meta, Microsoft, DoorDash, Robin Hood, Apple, Amazon, and Uber. So I have those eight of the group. So I have a vested interest in some of these earnings more. Are you feeling good? About earnings this cycle? How do you feel about like Q3 results?

Starting point is 00:04:06 You know, I have a thesis here. And I think it will be something we'll talk about on today's show. My thesis is, and this involves a lot of politics as well, we will continue the big spending train. I know there's Doge, Department of Government Efficiency, and I know Ewan said he was going to cut $2 trillion. That would be amazing if it happens. I don't believe it's going to happen under either administration

Starting point is 00:04:34 because I think what we're going to see is tax cuts from Trump and spending. And then from Kamala, if she wins, we're going to see slightly more spending and no tax cut. What that will net out to, I believe, is $10 trillion added to the debt. by the time this four years is over. If that does happen, where does that money go, Alex? Where do you think all that money winds up? Ultimately, winds up in the hands of investors, I think.

Starting point is 00:05:07 Equity holders, correct. Yeah. So if you were to spend money on government contracts, well, those government contracts might be with corporations. Those corporations service the government. They make the money. It goes into equities. okay if that money were to trickle down and go to new jobs for or raises for government employees,

Starting point is 00:05:27 where would that money wind up? Well, two thirds of our economy is consumer spending. So consumers, if they got a raise, what do they do with the money? They spend it on vacation. They spend it. Where does that wind up? It winds up in Uber, Uber, Carnival Cruises, you know, Virgin Airlines, whatever it is, Airbnb. Or maybe people treat themselves to a door dash more often, you know, once more a month, twice more a month. And so ultimately, equities is where I believe the money will wind up. It will not wind up in savings accounts. It'll wind up in 401 case, which is equities. It'll wind up in people buying equities. Or homes increasing in value. What does that mean? It means people who are the folks who own equities, which is the top 60% of the United States,

Starting point is 00:06:13 will benefit most in the bottom 40%. They live hand to mouth. They will not benefit. They're I believe corporations with static team sizes will have greater earnings, even if their top line contracts because of consumer spending, you know, let's say we go into a recession, which seems unlikely right now. Everybody seems to have baked in the soft landing has already occurred. But let's just say we go into a recession. If it contracts a little bit, then you just fire the bottom 10%, which, you know, CEOs don't have a problem with doing anymore.

Starting point is 00:06:49 boards don't seem to have a problem with it. So you just right size your spending. Earnings remain the same or go up. So that's my prediction. I agree with a lot of that. Yeah, well, Stonks go up was the old, the old riff. You know, it's interesting because that was partially kind of a cansyn argument, like government does this, economy does that.

Starting point is 00:07:11 But I wonder if there's not another side to that coin of what does it feel like as a nation to have net 10 trillion more in debt, because that's a non-insubstantial fraction of our total national debt. And I wonder how large that number can get. It's growing faster as a percentage than the global GDP is. And I wonder when those fall out of sync and then we end up in a worse situation. Do you think that happens in the next four years with 10 trillion? I think right now it's on people's mind, right?

Starting point is 00:07:46 And so the fact that there's even a dialogue of this doge and the government spending being out of control, the fact that that's even part of either platform's discussion means it's becoming acute. So we did not talk about it over the last 16 years because it was not acute. We had balanced the budget under Clinton and it was like a little bit of debt, you know, increasing didn't feel like our percentage was that off. Now the percentage is hitting an all-time high, and the interest payments are crazy. So I did a back of the envelope. I think the interest payments are now $3,500 per person per year. So you're a family of four. $14,000 is your share of the interest on our debt.

Starting point is 00:08:35 That if you were to put it in those terms, people would get very scared. And if it doubles over the next, you know, our debt doubles over the next three administration, Now you would be at your family afford $28,000 a year in interest. No principle. That will change a little bit, though, when rates come down because the interest payments skyrocketed with rising rates. So this is the chart I just pulled up. This is the federal surplus and deficit.

Starting point is 00:09:04 You can see that first real collapse was the 2008 era, I believe. And then we had a slight recovery through the later Obama years. And then the last couple administrations have been a little bit freer with the checkbook, you might say. If your startup is all in on AI, while DigitalOcean is the cloud platform that's going to help you scale faster without all the headaches, you're going to be able to do incredible things with their GPU powered virtual machines like training your models or building AI agents, or just crushing massive datasets all in seconds. And when you're done, you could spin them back down just as quickly.

Starting point is 00:09:39 With options from single GPU configurations to setups with eight GPUs, DigitalOcean lets you scale computing power based on your precise. needs without having to make upfront hardware investments. DigitalOcean is going to keep your infrastructure simple with an intuitive UI, robust product documentation and on-demand support. And with all that on-demand support, you're going to spend all your time building the next big thing, not managing infrastructure. DigitalOcean is your best choice because it's simple, scalable, and cost-effective. So here's your call to action right now. Approved listeners can get up to $100,000 in free credits on us. Visit DO.co slash twist to get started and view terms and conditions. That's

Starting point is 00:10:21 d.o.c.o slash twist. On the show today, we have a, we have Sonos's Matthias Sheik. We're going to talk a little bit about Google joining Anthropic and the AI PC race, the influencer election. And if we have time, some eye candy in the form of Chinese EVs that might literally take you to the moon. But let's start with Matias Sheik from Sonos, Jason. You really wanted to have Mattias on and I want you to tell people why. Well, I had just seen this great, Brad, welcome to the program, Matias. And you had built a speaker company and raised venture capital for it. I know the degree of difficulty in raising venture capital, better than almost any money, because I helped founders do it.

Starting point is 00:11:01 And hardware is hard. And you were able to land the plane and be acquired by Sonos. And I'm also a fan of Sonos's products and audio products in general. So I thought you had a really great story that we'd start with. maybe you could tell us about your startup and how you found that there was an opportunity to improve speakers. I mean, this is a space where, let's face it, people were, I think, pretty sure that we'd have incremental gains in speaker, but you found a way to maybe have more than just incremental. So maybe you could explain to people what your product was as a startup and then we'll get to the acquisition by Sonans. Yeah, sounds good.

Starting point is 00:11:40 So thanks again for having you on the show. So let's start the beginning. My brother and myself, Timothy, and myself, Matthias, we have always been very passionate about audio. Even as kids of like four and five years old, we would like take secondhand speakers, take them apart, pick out the woofer, put it in a bigger box, and crank it all the way up. And as soon as the doors would start to rattle, we would be like happy with what we were doing. So it's a deep passion. It's not just something we started out of college or anything. So fast forwarding to 2013, we had a high-end speaker company together,

Starting point is 00:12:15 and although we had something which was quite unique from a design perspective, it wasn't that unique from a technology perspective. So we were looking at what the others were doing in the high-end speaker space. And funny enough, if you'd look at a woofer or a tweeter or a mid-range from a tech perspective, so looking at the magnet, the membrane, and all the components, they were really quite similar. And we also found that a speaker that you would buy, let's say, fish, years ago is quite similar to a speaker you'd buy today from that hardware component perspective. So you see some changes in materials, you see some changes in tolerances, but not something

Starting point is 00:12:48 that's really drastic or revolutionary in that sense. And I think it was our naivity, like being very naive in a way that we found something that was quite different. So we started the company Mite in 2016 after our high-and-speak company that we had in 2013. And our sole purpose was to find something new in the hardware space and to be specific in the transducer space or the speaker transducer space. Mattia, I want to pause you there because we pulled up your patents and you guys had a patent for distributed transducer suspension cones in October of 2020. Tell people out there who don't know what a membrane is, transducer, just give people

Starting point is 00:13:29 some background on the terms here. Okay. So I'll just bring this into the screen as well. This is a speaker or what we call it a transducer, and this is the one that we actually built before Sonos required the company. So as you can see, it's roughly three and a half inch in diameter, but it can get the same performance as a conventional 8-inch buffer, which is, of course, much larger, heavier, and everything.

Starting point is 00:13:55 So membrane, that's the surface, which moves up and down, back and forth. We have this rubber suspension that makes sure if it stays in place. And then the internal consists of magnets, which are in this white casing, and there's a coil up here. So as soon as you, like, there's electricity going in the coil, it moves up and down because there's a magnetic field around what the coils producing around the magnet. So those are the main components of a speaker. So you were asking about this distributed motor thing that we were doing. So conventionally, like, as you can see, all around, you can see these magnets and coils. Conventionally, you'd have one magnet in the center and one coil.

Starting point is 00:14:39 The funny thing about that is that if you put something behind the membrane, you are actually limiting how far this membrane can move backwards, because you put something behind it, a magnet and the coil. And in the end, what we're trying to achieve here is as much air displacement as possible. So moving that away from the rear of the membrane to the side, which we did, it's now in the corner, actually creates more space for this membrane to move up and down, ergo to displace air. Got it.

Starting point is 00:15:08 So that gives a fuller sound in a smaller space? Exactly. So some people think you need a large enclosure for large sound. That's not the case. You need large air displacement for large sound. And that's what we're doing with this very compact product. And then this manifested itself in a new Sonos product, which I haven't ordered yet, but I'm definitely going to order any moment here,

Starting point is 00:15:31 which is this new sound bar. Now, I've had the old Sonos sound bar. This is in an ad. Although I think they did advertise a couple years ago. I've just been a fan forever of Sonos. And this is the Arc Ultra sound bar and sub four. It's a thousand bucks for this Arc Ultra. And it's a thin sound bar that goes under your TV.

Starting point is 00:15:53 And then inside of it is your technology. What does this experience do for somebody versus, not having a soundbar and just having a regular TV. Yeah, that's like a world of difference. Yes. So, where to start? So, for example, I think one of the things that I really dislike with television, like the building speakers, is vocal clarity or dialogue clarity, which is a huge upgrade

Starting point is 00:16:17 with a soundbar in general. But then what Sonos Arc Ultra adds is spatial audio. So there's 9-14 coming out of a single object, which is really quite unique. And next to that, of course, we introduced our woofer technology. sound motion, which then adds like good good and very solid bass, which is also something that's generally missing

Starting point is 00:16:39 in the TV. But Jesus, I'm sorry, again, my ignorance here, but you said 9-14, and as a music fan, but apparently not a sufficient audio file, I have no idea what that means. Yeah, good point. So, 9 actually means all the speakers which are around you on air level.

Starting point is 00:16:55 They've got your front channels, your rear channels, your side channels, etc. surround channels. Then one, means the supple, and four means your height left, so stuff that happens above you. For example, if there's a helicopter flying on top of you, it can happen in front of you, left to you, right of you, behind you. And it's all coming from a single speaker, which is really pretty cool.

Starting point is 00:17:14 That's amazing. So you have the ability to point some of the speakers up in the air, some of them straight out, and then create this sort of surround. Sound bubble. Yes. Oh, a sound bubble, yeah. And I know people call it soundstage in the, when people listen to music, but now you can get sound today with like these $1,000 sound bars that used to cost, I don't know, $15,000, $25,000 to do a room 20 years ago, yeah?

Starting point is 00:17:42 Yep, it's like it's just the next generation of what you normally, I'm using conjunction with TV or just to listen to music. How did you get venture capitalists to back a speaker company 10 years ago? Yep, that's a good question. So, first of all, it was really hard to get money for hardware. because people always say like hardware is hard and that's true. So and then if you narrow it down, try to get hardware funding in the Netherlands. And it's a tiny country. We're great with software, we're great renovation, but hardware isn't that much.

Starting point is 00:18:14 And then to make even more narrow, we're talking about consumer tech hardware. So it's which is really hard because it's a race to the bottom. If you make something great, some Chinese factory is going to make, you know, the same factory that makes your product is going to to make three knockoffs on the weekends or when you end your contract with them with some local founder in Shen Chen, correct? This is what happened. Yeah, yeah. So that's one of the bigger worries that they have. And also, if you look at the multiples in consumer tech, in hardware, it's quite a bit less than in software. So you were able to get VCs to come on board because they were audio files already and you just found, let's say, product VC fit. We talk about

Starting point is 00:19:00 product market fit, but there's also product VC fit. In other words, there are VCs who have private jets who backed Boom, which we've had on the show Blake a couple of times. And I consider that the ultimate like product VC fit because they all want to, they think their time is so precious that getting somewhere two hours early and then sitting in bed watching Netflix as opposed to getting there two hours later than watching Netflix in the plane is a difference. Yeah, yeah.

Starting point is 00:19:27 I think the main fit actually found with our VC company was disruption in a way. So what we presented was something that disrupted or disrupts the audio space as we did something that's very different from what has happened

Starting point is 00:19:43 in the past 100 years. And that's what really spoke to them in a way. So they weren't really audiophiles. They did have backers of their fund which were audio files that we met and Shorotech and which they really loved. But in the end they really felt for the part that we can disrupt the industry.

Starting point is 00:19:59 Like big tech at the time that we were trying to raise money. Everybody was doing smart speakers from Amazon to Google to Apple. And all of them were trying to figure out, how can I get this small speaker into people's living rooms in order for them to order stuff through my speaker? But also it should sound great, so people really feel like they have a good speaker in their house. And that part really spoke also to this VC,

Starting point is 00:20:23 as it, of course, would mean large numbers. I got to say points to the VC because they backed a company that ended up doing you very well. But Jason, to Matias's earlier point, the multiples in question, Sonos currently trades for a price to sales ratio of one. Not easy, yes, it's hard. That's tough. But Matisse, you guys ended up selling for what was reported to be about $100 million. So I'm curious, take me from Mites founding to when Sonos starts knocking on your door saying, hey, you guys have built something cool and we want it. Yeah.

Starting point is 00:20:51 So 2016, we were just a company of two, my brother. and myself, starting to develop the tech and we really started like a skunk's workway with practically no money, actually with the money we had from our previous company. I think one year in, we tried to attract a little bit of capital, like really a little bit, like a couple of 10,000. And we did that from the university where I studied. We made a proof of concept, which they really liked, which was really impressive. So they invested a bit more money in that same year and with a bit more, I mean, like around

Starting point is 00:21:24 50K. Thereafter, we needed, like, we took really actually small steps now to think about it. Most of these stories are like, we raised 500K and then 10 million and then 26 million. But we really went from like 20 to 50 and then the next step 300, which was convertible debt, which never converted because we actually had the choice to convert yes or no. Yeah. Oh, nice. Yeah.

Starting point is 00:21:47 Oh, so that's, this is a very unique thing. If it doesn't convert, you pay it back and you don't have to give up the equity. depending on the deal, that can work out really well for the founders or not. And for the investors, sometimes they don't mind doing that. I wouldn't do it. Are you looking for an unbeatable deal on senior talent? Well, if so, you need to look outside of your borders. According to G2, Cloud Deves is the best place to hire top Latin-Am talent.

Starting point is 00:22:16 That's Latin America, a market with millions of developers. Cloud Deves is the largest pool of Latin American talent with over 10. thousand senior engineers who are rigorously vetted for communication and problem-solving skills. And now, Latire by cloud devs can even fill non-technical roles. Very interesting, from accounting to graphic design, with elite Latam talent. Giants like Coinbase are choosing cloud devs because its talent is matched with your time zone. And it's up to 80% cheaper than a local hire with the same Silicon Valley standards. So here's your call to action. Because they're such big fans of the show, CloudDev's is offering

Starting point is 00:22:54 Twist listeners an exclusive pay what's fair deal. Instead of their usual fee, only pay what you feel the talent is worth for a limited time only and exclusively for our listeners. Just go to clouddevs.com slash twist to pay what's fair for your next five-star hire.

Starting point is 00:23:11 So tell me about how the audio industry reacted to what you were doing. I know it's a very precious insular industry. When they saw what you were doing, did they think this is BS and try to deride you? Did they copy you? Did they celebrate you? Because it is a bit disruptive and it's lower price. And man, I use some focal headphones and I got these key speakers. These

Starting point is 00:23:37 people are really, really like serious audio files and very serious about product. I'm curious how they looked at your entrance. Yeah. So first, of course, there's disbelief. Especially like we're in in the Netherlands. So I can really demo anything through Zoom or whatever video conference. So first of course, we tell a story and we tell them we can do something 10 times better than you, which is, of course, a pretty tough story to sell. The marketing people like it, the tech people a little bit less. We spoke to a lot of companies where you'd find that they would feel like we didn't invent it so we don't really like what you're doing, that kind of feeling that we had, they had. So over Zoom was pretty hard or any video company was pretty hard.

Starting point is 00:24:20 But as soon as we get the chance to demo what we had developed, it was a way different story. So we'd get in the room, they would get all their smart people in the room, all the guys that had infinite patterns or made big inventions, and which are patented. And we'd demo our speaker. Funny enough, there's one here in the corner. I think it's already visible. Quite small. There, next to the Air 300.

Starting point is 00:24:44 And again, disbelief, like, is this really coming from that tiny speaker? which was actually in the beginning was really, really funny to see that, to see the reaction where even companies bring in their own speaker because they thought maybe it's the acoustics of the room. So they wanted to compare it and make sure what they were assuming as real. So let's go, you know, as we get to the third act here,

Starting point is 00:25:06 getting acquired. How does that happen for a startup? I've always been of the belief that great companies get bought. They're not sold. Did you go out and try to sell the company or did people start knocking? Yeah, it's the second. So we, of course, we kind of set it up in a way. So we always stayed in the background until CES of that year, 2022, where we really made a lot of noise.

Starting point is 00:25:29 So we won best of CES at that time. We started talking to reporters. And of course, we reached out to all these companies that we already were in discussions with for a couple of years. So even from 2019 onwards, we've been in discussions with like almost all of the audio and tech industry and just like keeping them in the loop of what we were doing. So this is a key, key important part here, Alex. What great founders do is they keep the competitive set and the partnership set aware of what they're doing. Now, this seems like maybe you wouldn't. You'd want to be stealthy.

Starting point is 00:26:03 Well, once the products out there, they're looking anyway. And, you know, most of the acquisitions you'll see are partnerships to acquisition. In other words, somebody says, oh, you know, we want to do some sort of partnership with you. And then they offer to acquire you. So that's what happened to me with Weblogs, Inc. AOL said, hey, we want to invest. Then they said, we want to buy half the company. Then they said, we want to represent your ads.

Starting point is 00:26:27 And then at some point, Ted Leonez has just said, you know, well, we'll buy the whole thing. This is getting too complicated. Why don't we just buy the whole company? And I think that's probably what happened with you, too. Yeah, Matthias? Yeah. So, our strategy was to sell licenses to all of the audio and tech world. So imagine getting a couple of bucks for every speaker sold with hard motion or sound motion

Starting point is 00:26:47 technology. That's pretty cool. But in the end, it's also a huge competitive advantage to have a technology that all the arts don't have. And that's the part that Sonos really fell for. Next to, of course, a great cultural fit and in general, great product technology fit. So that's what I want to double click on, because people talk about selling their company, but we often don't talk about day two, like once you're there. So you guys sell to Sonos. What was it like going from being, you know, in charge of your own company to being part of a much larger organization and how hard was it to get your tech into their roadmap? Yep.

Starting point is 00:27:20 No, actually was really quite funny. Like from day one, so the company got acquired and there wasn't a break in between it being might and it becoming a part of Sonos. Like on Friday, we brought the news and on Monday we started. Oh, wow. It was pretty cool. But it actually went quite natural. Of course, there are quite large differences.

Starting point is 00:27:41 So the weight of changing a roadmap with such a large company is way different than with the startup, we could just say from the one day to another, let's stop doing this and do something else, which of course, sounds couldn't. But as they were already working on our Ultra when we joined, they were actually quite flexible and changed to design quite drastically once we joined. Are you happy you sold the company? Like, looking back, I mean, I know this is a public forum and you're talking about your employer, but like, you know, wasn't the right choice, do you think looking backwards to sell for that price at that time? Yeah. Yeah, I think like, I'm really happy with our choice, like, especially from a timing perspective, it really was the right time.

Starting point is 00:28:18 The market was doing great. And also, there was a great need for this technology. At the same time, also, our company, we are really great at disruption. And like the first part of having an idea and bringing that to life. And then Sonos is really great at building reliable, really great sound experiences. So marrying those two together to make beautiful products, it turns out. to be a winning combination. Yeah.

Starting point is 00:28:46 All right. Well, I'm excited by it. Well, and we're looking forward to buying the Sonos arc. Everybody can go to Sonos and just pre-order it. I don't know if it's for sale yet. Is it for sale yet the arc? Yeah, if you're in tomorrow, like tomorrow onwards, you can really like buy it. Oh, okay, good timing to have you on.

Starting point is 00:29:02 Continued success. And then, hey, the final question for you, these high-res flak files I've been getting into and playing them. Do you have the ability to play those yet on Sonos? I know Sonos has been working on their hi-fi strategy, and there were one or two speakers, but hi-fi doesn't work over Bluetooth. You've got to have,

Starting point is 00:29:21 like, Ethernet or something plugged in. So tell me about the hi-fi strategy here, and with your speakers, does it make a difference if you're playing like an MP3 off of Spotify or a higher-res MP3 off-title or co-buzz and using a proper flack like I do? Yep. So, of course, it matters.

Starting point is 00:29:40 Source or content quality quality matters. What I personally do if I want to listen to a flag file or even like a master rendering, I use title. Yeah. Probably also familiar with that. Yeah. Yeah. Play high-res audio.

Starting point is 00:29:55 And so does the speaker support that high-res and the input of it now? Title is supported on Sonos? Yeah. Title supporter. Absolutely. Yeah. Great. What about COBUS?

Starting point is 00:30:05 Do you guys do COBUS yet? Oof. I mean, I wish I could answer that. I feel like I should know, but I don't. You know about COBuzz, right? Yeah, I do. I do. Like, I used cobes before title, but then I found that they have less content available, so I moved to title. I may have to move to title then.

Starting point is 00:30:22 Yeah, I am really getting into it. I can't wait to get this. I have an LG TV, and they've gotten better with dialogue. So I was using it last night. I was watching Inside Out 2 at my daughters. And the dialogue was much better than previous TVs, but I know it can go up a level. Why is the mix on dialogue so? problematic today.

Starting point is 00:30:46 Don't understand why everybody is complaining now that Game of Thrones, Breaking Bad, whatever it is, they seem to make the sound mix not, um, they seem to make it much different. Maybe it's because they're making it for like movie theaters and these and they're,

Starting point is 00:31:04 they're listening to it in a way that 90% of the audience isn't listening to it. Yeah, I can imagine like, what I would recommend is like, for example, with Arc Ultra, we can render spatial audio, which means you have a dedicated center channel and we have dialogue enhancement. So all the details that we're put in the mix, you can really get that like I'm a dedicated speaker, which I believe is quite nice. So this is a known industry issue, the dialogue issue and everybody's kind of attacking it at once. Yeah. Yeah. I think this,

Starting point is 00:31:37 do you have this issue too, Alex, where you can't hear the dialogue, but you hear explosions everywhere? So I was just actually chatting to the YouTube crew. I wonder if this issue is why people in Gen Z and the Millennials love to have subtitles on. Because one thing that people say is, oh my God, these kids always have subtitles on. Maybe our media is just poorly mixed and no one can actually understand what people are saying, especially if there's any background noise. The only other hypothesis I have, Jason, to your point about Game of Thrones and Breaking Bad, shows that use a lot of ambiance and maybe sound engineering to create a lot of background noise to set the stage,

Starting point is 00:32:09 might have an issue of that clashing with dialogue. And I think people expect more of a sound script than we did back in the, you know, pure dialogue film era. So I wonder if that's an element of it too. But my TV is terrible. So I don't know if my TV is trash or if I'm commenting about an actual issue here. Well, we're going to get you sound bar. And that should solve it.

Starting point is 00:32:29 All right, Mattias. Thank you so much for coming on the program. Congratulations on your startup. Hardware is hard. You made it happen. And now you get to, you know, have this great. partnership with Sonos and the team over there, which is awesome. My friend Julius Jankowski is on the board and, uh, you know, I'm a big fan of their products. I use them

Starting point is 00:32:48 in every, uh, one of my homes, which I guess sounds obnoxious. Um, in both of my homes. I have sold the other two, but I love Sonos because it just works and it's so elegant and beautiful and it's awesome that it's supporting high res now. Uh, all right. Thanks so much, Mateh. Thanks, Jason. Thanks, thanks Alex. See ya. What is a prediction market? Let's talk about that to start. It is a way for you, just individuals, to buy or sell a contract on a future event. Okay, what's the future event that you could buy or sell a contract on? Who's going to win the presidential election in 2024? Hal She, the world's largest prediction market, just made it legal to trade on the upcoming U.S. elections. That's right,

Starting point is 00:33:30 for the first time in 100 years, U.S. citizens can now legally place trades on election outcomes. election is less than, I don't know, two weeks away. Wow, so close. Now, political polls, we know can be highly inaccurate. But the markets, that's where the sharps are putting their money. And so the big question is, who's going to get it right? And that's where the arbitrage comes in. And that's where you have to make a decision. And election outcomes matter. They impact your taxes, the economy. And of course, it's going to impact your investments. So you can use this as a hedge if you believe that one candidate is going to be great for your stock. And you might bet one way. If you believe that they're going to be terrible for the stocks and the economy,

Starting point is 00:34:12 you could bet another way and that could hedge your exposure in another category, like say, equities. CalShe allows anyone to protect themselves if their preferred candidate loses. So go check out the real-time election odds. It's really fascinating just to see it go up and down based on the daily news that's coming out or somebody had a rally, somebody had a great interview, somebody had a bad interview, a great debate, a bad debate. I think it's very exciting and this is the future because, hey, people have skin in the game. I want you to check out Kalshi. That's Kalshi.com slash twist, K-A-L-S-H-I, K-A-L-S-H-I, Palsh-C-com slash twist, and get in on the action. You will get an additional $20 with your first deposit of $100. That's calshy.com slash twist. What's next on the docket? I love having these

Starting point is 00:34:59 live guests and it adds a level of like raw honesty. to it. I love the fact that you're asking hard questions, Alex, well done. But let's keep going. Polite edge. I try to be like, hey, I know your employer is going to hate this, but, and that way it gets them all relaxed. Okay, real talk to us. Somebody was saying last week when we did our great, we had heritage on to talk about voting irregularities. And we had Hans on and people were like one or two comments in YouTube, oh, you have TDS or, you know, like he's never going to come on the program again, blah, blah, blah. And I'm like, at the Hans and Heritage want to take hard questions. There are,

Starting point is 00:35:33 partisan organizations in the world. This is a nonpartisan show run by two human beings who also vote in the election. But I thought I asked him, I don't think any of the questions were particularly hard. I don't think he would say that he got hard questions. I think the issue is fan culture and people getting very tribal in what they expect questions to be. And I think in the era of people going direct, any question to someone of note that is, let's just call it sharp elbowed is considered to be hating

Starting point is 00:36:07 or being biased but I think the hard questions are what made that entire section good because you and I could have gone over the data ourselves and shown the statistics but it was pressing Hans on a couple of things that I thought was in fact useful I just want to say you end up

Starting point is 00:36:23 with it by the way where did you because we're sitting here a week later six days later where have you wound up in terms of your assessment of voting irregularities and people cheating on voting. Obviously, there's some amount of fraud in every system. So where do you net out specifically with the presidential election that is occurring in 10 days? Yeah, I think we're doing pretty good.

Starting point is 00:36:48 And I think we're doing better than we deserve to, given how our system is set up and given how spread out handling elections is around the nation. I would expect there to be more fraught, given that there does seem to be ways to impact the system. But the penalties, as we discussed with Hans, do seem to keep people away from voting. Going through the heritage database, prepping for that segment,

Starting point is 00:37:16 because I went through a lot of the entries. It was a lot of people that were like, I didn't know that in my state, a felony precluded me from voting. Right. Now, some large percentage of it are individuals who registered to vote who weren't citizens or had some other issue or were Fallons or something like that.

Starting point is 00:37:38 The conclusion I've come to is because we have a union of 50 states, which I think is this brilliant feature of the United States, it's like having 50 different operating systems or forks of an operating system. And so what that means is, if you have one, if you just have Windows, there's one attack vector affects every Windows machine.

Starting point is 00:38:04 But if some people are on Mac, some people are on Chrome OS, some people are on Windows, and that's just three operating systems, you can only have one third of users go down, right? Or one of the three. But we have 50. So to hack a national election would require getting into 50 different states,

Starting point is 00:38:25 or let's say six or seven swing states, and having to do different things in each state because they run them differently. And because everybody, and the comments are always, well, why don't we have like one thing, right?

Starting point is 00:38:38 We'll have one system of record. It's like, okay, one, then you have one system you can hack. So in some ways, the fact that there's 50 different systems keeps it from,

Starting point is 00:38:50 at least on the federal level, when we have a national election, impossible. And if you extrapolate, late, and this is where I think maybe people thought I was a little bit sharp elbow or whatever, he wouldn't answer how many were federal.

Starting point is 00:39:06 You know, he wouldn't answer, like, does he think Trump won or not? Yeah. Do Trump lose? So I understand partisans don't want to answer that because they lose favor with Trump and, you know,

Starting point is 00:39:17 like saying your team lost is kind of tough for people. Like I would, when the Knicks lose, I'm kind of like, we got robbed. me and the U.S. women's national team every Olympics or World Cup. Yeah, sure. I'm with you.

Starting point is 00:39:31 You're like, I got robbed. The refs got me. But then also, like, he wouldn't answer the question of like, how many votes do you think are actually stolen? If you have 1,640 a year, would we times it by 10 or 100? And that's where I took from that baking in his partisanship, which is fine. It's a think tank. That's a partisan think tank. Yeah.

Starting point is 00:39:57 Even if it's 10 or 100, if it's 100 times, if there were 100, instead of 1600, if they found 160,000, over 40 years when we have 100 to 200 votes a year on average, I think we're in great shape. Now, every vote should count, but if I told you there were a billion credit card transactions a month and we had a thousand cases of fraud, you'd be like, awesome, that's what we're having in voting. that's my impression of it and it's I find it very encouraging because there are people out there who want to influence impact and change our elections I mean the government has spoken about Russia, China and Iran in particular

Starting point is 00:40:36 about where they're looking to change things or at least influence things and I think we're doing okay I just think as well that it's such a it's such a bummer to me how how partisan voting rights are today and frankly

Starting point is 00:40:53 the Heritage Institute our foundation has been muddy in the waters a little bit. They did this clip that I found when I was prepping for that segment. They were going around asking people in Spanish, you know, are you a citizen? Are you going to vote? And some people said that they weren't a citizen and they were going to vote.

Starting point is 00:41:07 And then the post went back in fact, checked these people and no, they're not going to vote. They're not registered. They just didn't understand the question that was being put to them. And so that is a way to engender fear of essentially people who are around. Your uncertainty and doubt is the strategy, I think,

Starting point is 00:41:24 you have on one side, FUD, and then on one side you have, hey, you know, you're making it impossible for this, you know, a group of individuals, pick race, gender, socioeconomic status. You're making it impossible for them to vote. Folks, to get on an airplane, you need an ID. It is not too much to ask for ID voting. Let's call it what it is. Give you a national ID plan that will make sure that everyone has one, and I will sign that day one. It's such a simple thing. So I just find both sides are so insincere about solving this issue. Maybe one side, 60, 40 or 70, 30, we could debate that. But let's keep moving through the docket here, you know? Let's get moving through the docket.

Starting point is 00:42:10 And let's actually start with the state's point, because one thing I found while I was learning today about how influencers in our elections are being paid under the table, Jason, is that this is something that's not been decided on the federal level. and in fact, states are stepping up. So in this post that was written by Katzikreski, an excellent journalist over the post and former tech cruncher. This is Washington Post. Oh, yes, Washington Post. Washington Post did a story about influencers being paid by PACs or super PACs, correct? Yeah.

Starting point is 00:42:40 Yeah. So the folks on your TikTok or on your YouTube or on your Instagram that are talking about a particular issue might be getting a check, Jason, on the side. and at the federal level, we as a nation decided we're not going to regulate this mostly. And so people don't have to disclose that they're being paid by a PAC or a super PAC to post about a particular issue to their audience.

Starting point is 00:43:02 And to me, that just seems like an odd exception to make for normal paid disclosure rules. But I want to get your take on that first. Well, if you're doing an advertisement, if I were to say right now, hey, go to Athenawow.com, get a virtual assistant. It's the greatest thing ever.

Starting point is 00:43:20 I would always say, before you go with Theenowow.com, understand their advertisers on this pod. And I was the first investor in this company and their last company. So you would have that context. I am biased in the fact that if they do well, I do well. I have an investment in the company. So that's the kind of disclosure you should make. And I try to do that. Anytime I talk about one of our startups, I will say in just natural language,

Starting point is 00:43:48 I'm so excited to be investor in density. Here's their new product, this new waffle when we had them on. So transparency equals good and integrity. So what's the issue here? The issue is there was a company called pay-per-post or whatever back in the day. Yes. And they were covertly getting influencers, you know, in the YouTube era before TikTok and Instagram blew up.

Starting point is 00:44:20 And they would do really nefarious stuff. And they called these Buzz agents. There was another company called Buzz agents 15 years ago. This is how sinister it was. I mean, sinister in a clever way. Okay. So I give props to the marketer. They would have the Sony camera.

Starting point is 00:44:35 They would have a beautiful couple, like unrealistically gorgeous couple in Manhattan on their honeymoon. The couple would say, hey, would you mind taking our picture? you would take a picture. They would say, yeah, no, no, I just my, I got this wedding gift. It's the Sony VX 1000.

Starting point is 00:44:56 Can you hit this button so I can get a Y shot in 4K? And they would take your picture and they would just do this all day long. Oh, wow. Is this? And they wouldn't disclose it. Then they had like Jimmy Dean sausage or somebody. I'm making it up. Yeah.

Starting point is 00:45:12 And then they would send you sausage. They'd say, have a party, take pictures of the party, share it on your social. of you and your friends eating Jimmy Dean sausage and now you're being used as a prop. Yes. Your friend who's serving you sausage is in on it. The next one that was incredible in this like buzz arrow was

Starting point is 00:45:31 Hey, Alex Wilhelm's new book, Cautious Optimism, the book, has come out and you call all the bookstores and say, Hey, do you have any copies of cautious optimism? I got to get this book. And they'd be like, let me check. Yeah, we have two copies. Hey, could you put one aside for me?

Starting point is 00:45:49 I'm going to come pick it up tomorrow. You never show up. But you do that a thousand times once a week, twice a week to all the bookstores. What do the bookstores do? They order more copies. Now they got more copies. They're stuck with them. They put them in the mirror and they just created buzz that you go into the store and they

Starting point is 00:46:07 were sending agents into stores saying, hey, do you have this book? Because my friend told me it's the greatest book ever to influence the people at bookstores who are the rep. Wow. I mean, this is creative thinking, but it's sinister and it's against the FTC rules and they came out with rules. You have to disclose. That's why you see paid partnership hashtag in a lot of paid partnerships. But in Kat's story, discussing this, she points out that people want to make sure that there is plenty of room for political speech and that we don't have two onerous of rules for political speech versus commercial speech. And I can see that being an argument you might really buy into. But when I watch an ad that's paid for by one of the major two campaigns, at the end it says, you know, I'm Kamala Harris and I approve this necessary whatever, which I think is good. I think it should have a name attached to it.

Starting point is 00:46:59 Super PACs and PACs definitely muddy that. Super PACs impacts paying influencers, muddies that even further. So to me, Jason, given the examples that you've put out there, and given that marketers will do anything to hit their message out, I think more disclosures here make sense, which is why back to the state's point, Texas is kind of an outline example here,

Starting point is 00:47:20 because Texas decided that if you're going to do this, you have to disclose that you're being paid to make a political post. So shout out to the Texas ethics. And the detail is, how do you have to disclose? You have to say it up front at the end. Do you have to say it in audio or do you have to bury it in a hashtag? And at these, the devil's in the details, because people could give a whole thing and then just put it in the hashtags and nobody would see it. Yeah, you should always have disclosures.

Starting point is 00:47:49 It's so simple. Even like the New York Times when they bought wire cutter, they put a little sentence, if you buy stuff on this page, we get an affiliate link. does that make anybody think twice about, you know, the advice in wirecutter? For me, no. Because I trust it because I know the writers there or whatever. But, you know, on a schlocky website, they don't do it. And you think, hey, the schlocky website has their 10 best, whatever, you know, webcams. And they don't disclose it.

Starting point is 00:48:20 And you know they're getting affiliates. So they don't care what advice they give you. They just churned it out using chat GPT and make a thousand. pages all over the web with top 10 lists, the top 10 list is meaningless. And so you really have to, it's buyer beware in all this cases. And I do think super disclosure,

Starting point is 00:48:39 like really upfront disclosure is important in all these platforms. And I do applaud Texas, if in fact they're stopping this shenanigans. I don't like shenanigans. No, shenanigans are not good. But I do think the fact that we're talking about this does underscore how things have changed in the, not just get out the vote, Jason, but get out the message side of campaigning.

Starting point is 00:49:02 I mean, think about 2008 was, you know, pretty email focused. And this is also mentioned in the Washington Post story we're describing. But over time, things have changed. And I think today, social media broadly is where a lot of people who are potential of young or even kind of middle-aged voters are getting their news. So having your reach there in a more authentic way, aka from influencers versus branded accounts, makes a lot of sense.

Starting point is 00:49:27 But if we're going to do this as a nation for the future, disclosure rules seem to be something that we should be able to pull off. And yet, not to disparage my favorite form of government, democracy, but we can't seem to get this one over the line.

Starting point is 00:49:41 And that is to me a disappointment. But both political parties do this and both at the federal level don't want more rules around it, which I think is a real le-time. And there are more pernicious ways to do this, I can tell you. Really, you know, the,

Starting point is 00:49:54 what happened with RT backing that third party podcasting company and then giving those four podcasters a hundred thousand an episode and then they look away and don't ask questions of why they're getting a hundred thousand an episode but they previously made five thousand an episode so it doesn't add up but they just take it anyway these are the kind of things that i think are super important to consider there are other ways to do this as well you could pay for somebody's trip to the dnc the You could pay for their hotel room. You could give them a stipend.

Starting point is 00:50:29 Now, they're not paid to say anything, but you got them there. You paid for their camera equipment. You paid for their trip, first-class travel, whatever it is. And that's where junkets come in. And in journalism, we had a junket culture, and it was pernicious in the automotive space. Very. To the point at which automotive journalists would get a car for a year or two. And they would have two or three cars in their driveway because somebody might say, yeah, you know, you can just hold on to the car.

Starting point is 00:51:02 It's insured. Uh-huh. All you got to just put gas in it. Uh, and I remember when we did auto blog, somebody was like, hey, yeah, we'll give you this card to review. I was like, all right, great. So then we did it. And I was like, oh, wait a second. I just got $2,000 of value taking this thing to Tahoe.

Starting point is 00:51:19 Uh, and so I looked at what the automotive review, uh, policy was at. at the New York Times, and it was, we return it with gas full, if you give it to us gas, and we can take it for whatever number of days is necessary to fit the story format, which in their estimation was one week. Yeah. So we'll take the car for a week, we'll return it with the gas you gave us, that's it. And other people, you know, would take them forever. And then they would fly them to the Autobahn or to Kauai to do like a racetrack in Hawaii

Starting point is 00:51:55 or Germany or France and you're like, okay, so how's the review going to turn out from this mega corporation that just spent $20,000 sending you first class somewhere, putting you in a great hotel for a week to drive their cars? If you write a bad review, maybe you don't get invited to the next one. And that's the thing. If you dangle all these perks, you might say yourself, look, I am an independent person. I'm not going to go to the auto bime to drive the new Mercedes SL $40 billion. and be biased, but then you're sitting there your computer, you're like, that was lovely. And I'm a freelancer, and I make $4.

Starting point is 00:52:30 And they've only first class to Bund or something, you know? Yeah. And that's, yeah. Well, that's why Junkas are generally every publication and everyone that you can name their band, because they don't want that kind of bias. And all this is to say that if you want to have trust with people,

Starting point is 00:52:47 and I think influencers really trade in trust with their audience, you should disclose. So I think it's even better for the influencers themselves, Jason, to be upfront, to be up clear. Like when wirecutter tells you, we might get into commission, you know they thought that they should tell you that.

Starting point is 00:53:01 It says a lot about their ethics, how they approach business. All right. Next story. Let's keep moving here. All right. So you want to talk about this. Let's talk about trust.

Starting point is 00:53:10 Trump Rogan YouTube search. Jason, what is the... Oh, okay. So there's a kerfuffle today. David Hanmeyer Hansen from 37 Signals did this tweet. He searched for the Joe Rogan interview of Trump,

Starting point is 00:53:23 which I guess everybody wanted to watch. watch. I got to watch the first half hour of it. I'll watch the other two and a half hours of the greatest hits at some point, I'm sure, when I go for a hike or something. But if you search for Joe Rogan or Rogan Trump interview or Rogan Trump interview full, it does not come up in the search results. So I did the search for it and it just doesn't come up. Now, there are multiple issues that play here. But end of the day, Google's the greatest search company in the world.

Starting point is 00:53:58 YouTube search is normally pretty good for me, but I know it's not perfect, and I do know there's a bunch of clipping accounts that clip a lot of stuff, but it should know A, this is the hottest video on YouTube, 30 million plus views in whatever, three days, and it should

Starting point is 00:54:16 know that this is the right video. So this will make conspiracy theorists. the technology industry is trying to suppress the video. Now, the video has got 33 million views, so it can't be suppressed. It's being viewed, but it is being suppressed in search results. Is it being suppressed or is it being,

Starting point is 00:54:38 is it not coming up in the top views because short clips are preferred and there's so many short clips of it that the zone has been flooded by short clips. What's your take? Yeah, so YouTube search is bad. And while I really agree five years ago that Google was the best search company in the world, I don't think that's the case anymore. In fact, Google search is pretty poor. And often you have ads throughout the whole top of the fold and they often want to push their current thing. So on YouTube, to your point, they're very much into shorts.

Starting point is 00:55:11 And they have changed search in such a way that when you search for something, it gives you three or four or five entries. And then it says people also liked and then just shows you whatever it wants to. YouTube search is not search per se so much as it is a conduit for them to give you stuff they think will keep you engaged and that is very different than find me this specific video and I've run into this problem for years

Starting point is 00:55:35 on YouTube search, it's terrible and just for example, for this show today, I was trying to find the 32nd Sonos clip showing off the arc soundbar. I did variations of the search before we even had this conversation and it took me five or six tries, I think, to find that actual video

Starting point is 00:55:52 because it was always reviews, reactions, you know, AI voiceover, stolen video clips, and just bilge. All right, so here's a Google search. What is the search term here?

Starting point is 00:56:04 What do we search for on Google search, not YouTube search? This is Trump Joe Rogan. Okay, Trump Joe Rogan, great. If we look, you get news because it was a big news story.

Starting point is 00:56:15 There's a lot of news coverage on these keywords. And then if you scroll down, I'm sure the first link is going to be the, YouTube video or not. Let's see. Nope. Yes, the first actual web search does seem to be... Oh, so the first organic web search is correct,

Starting point is 00:56:30 but the videos, it starts with MSNBC, and then some other account... Flash post and Robert F. Kennedy Jr. So this is the issue. As you can see here, Google search, organic search starts after the video, is perfect. So this would lead to the theory that YouTube search is broken because the YouTube results here for videos

Starting point is 00:56:54 if you go to videos in the tab on Google and you do a search there it's YouTube videos plus everything else and so if we go to videos actually it came up number one so there you go but the little snippet was incorrect so something is going on I think around the other theory I have is

Starting point is 00:57:13 they talked about election interference during the show and they talked about COVID I know from this week in startups videos, all in videos, if you talk about those two subjects, they can label you. So let's click on the video here and see if it got a label. If this has a label on it, if we scroll down, does it have the warning about COVID or elections? Nope.

Starting point is 00:57:36 So that's interesting in and of itself. This should have been labeled about elections in COVID because I know they talked about elections extensively. So I wonder why that didn't come up. and I'm pretty sure they talked about COVID. So that should have come up. I think I know why that's the case is because they don't want people saying, well,

Starting point is 00:57:55 they're censoring this. Oh my gosh. Bias and so forth. I bet they're giving really, really free range to anything Joe Rogan wants to do. I was just trying to figure out where it ranked on trending because often when you hold a YouTube clip that's new and hot, it'll say,

Starting point is 00:58:09 you know, number 34 on trending. I don't see this one on the current trending list. I don't know if that means anything. But I don't think they're going to be trying to, point out that Trump might have told a couple of whoppers in there because I don't think they want Trump to try to ban YouTube when he becomes president if he wins. Well, let's do this.

Starting point is 00:58:26 Let's go to Bing. Oh, that's a good idea. And type in Joe, the same exact thing, Rogan Trump interview. And let's see what Bing thinks, because this would be a way, and then we could do duck, do. So if we do those two,

Starting point is 00:58:40 we would get like to some ground reality here of if this is a YouTube issue, a Google issue, and bias. So on Bing. So here's, here's Bing. This is what comes up.

Starting point is 00:58:53 I haven't used Bing in a minute, uh, later. Um, so news first. News first. Yeah, explore more news.

Starting point is 00:59:01 Top 10 moments from watch mojo. Okay. Um, and then it goes to the organic results after news. The first one is New York Times. Scroll up. Okay. So you have news.

Starting point is 00:59:12 Then they do a video. The first video, they have watch mojo. So they got it wrong. then for organic results are starting here. First one, BBC's seven takeaways. Second, we at one is did YouTube censored show Rogan's Donald Trump interview from Newsweek? The Hill five takeaways, then variety.

Starting point is 00:59:34 So here, if we're scrolling down here, where is the actual interview? So let's click this button up here where it says see more underneath the first video, just to see what we get. Yeah, see if there's more video here. The answer is nothing. When you push that button and Bing, nothing happens. Okay, so there you know. So let's scroll up and just click the video tab. And let's just see if it actually, I mean, this could just be the fact that this thing got so much coverage for so many different outlets that it kind of broke the algorithms on the entire web.

Starting point is 01:00:06 So if you were to click the video tab, what happens? This is it right here. Oh, I'm sorry. Yeah, you have search co-pilot video. So. This is the, this is the, this is. the Bing video search and very similar to YouTube it's clips, excerpts, and

Starting point is 01:00:21 smaller bits. Interesting. And let's go to duck, duck, go. This would be very interesting and then we could also do Brave's search engine, the Brave browser, but if we search just up there for Trump,

Starting point is 01:00:38 Rogan interview, boom. Let's see. Okay. The news is in, they put news first and then the BBC Is that and taking it brushing it on SEO apparently yeah apparently BBC is very trusted

Starting point is 01:00:52 I think as a news source so they get it wrong too and just for fun videos really quick oh yeah same issue all clips not the original so what we've learned

Starting point is 01:01:04 three of those four get it wrong and Google video search gets it wrong so out of five searches four out of five get it wrong the only one that gets it number one in organic results is Google search which I mocked at the very start of this segment and I feel slightly silly.

Starting point is 01:01:17 So shout out to Google search for having the first organic result be correct. But you know, these search products are not designed to help you find something. They're designed to keep you watching. And I think that's... That's an interesting insight too,

Starting point is 01:01:33 is if you are keeping... If they are designed to make you hunt and pack through all these clips, maybe the algorithms because they're designed to if you wanted to extend the session time, showing people clips and making them fight to find the original, the algorithm could be doing that

Starting point is 01:01:54 to extend people's time on site. That's a really interesting, scary thought. The algorithms are not your friend, Jason. The algorithms want to waste your time and get you to click more, because if you click more, you would click on more ads. I think that might be actually what's occurring here is that they make more money. If the algorithm was told make money and increase session length, those are your top two orders,

Starting point is 01:02:24 and third get to people to what they want. Like I said, distant third, eventually get them to they want. The algorithm would say make people click and do four or five searches to increase their time on site. So it purposely frustrate you. Wouldn't shock me if that's the case. Although I will say, though, what's interesting about this entire situation is we're not taking into account subscriptions. And people, you know, we love to say on YouTube, like and subscribe. But Joe Rogan's clip with Trump, which is several hours long, got, as we said, 34 million views in a couple days.

Starting point is 01:02:55 People found it. The thing is, they just didn't find it through the search box mostly. So I guess that just goes to show the power of people actually subscribing to a channel, because you don't get 34 million views that fast by accident. the link was definitely shared everywhere on socials and people eventually found it. So if you're a subscriber, it probably came up on your for you page,

Starting point is 01:03:17 essentially your logged in version of YouTube. So fascinating. I'm so excited that the election is almost over, Jason, because I feel like we have to go around and like everything is so amped up. Everything is, oh,

Starting point is 01:03:30 they're censoring this. Oh, they're stressing this. Oh, oh, I just, well, I think we've done our service

Starting point is 01:03:36 as Captain Obvious and I'll be Captain Obvious, Captain Nuisance, and you could be like your own... Second Lieutenant Obvious or something, yeah. No, nuance. It would be like a superpower, or we'll be of this like super team, because we looked at election interference with Heritage, the people who have the most bias of anybody, you know, and are totally... like literally writing Project 2025 as the suggested agenda for Trump.

Starting point is 01:04:11 We had them here. We came to our own conclusion with their data with them. It was a wonderful discourse. And here we are looking at search. It turns out four to five search engines didn't get this right. So then you have to ask yourself, is it Microsoft, duck, duck, go, Google's team and YouTube's team, which are different teams, all working in concert against Trump. And I think the answer is, I don't think so. I think this is an algorithm problem.

Starting point is 01:04:40 Those teams aren't even working in concert with themselves half the time, and let alone across different organizations. There's no, there's no cabal, guys. It's just a bunch of nerds that are trying to make one KPI slightly better. And what that means is you don't recommend three-hour video clips top in search results, as we found out. No. And listen, I'm doing Brave Search, which is my favorite search of all,

Starting point is 01:05:01 and I love the Brave browser. and if you go to search.brave.com shout out to my friends at Brave, who have been a sponsor of the show, full disclosure at some point. Although I don't think they are now. Literally, they start with news in its independent, San Antonio current,

Starting point is 01:05:19 Newsweek, independent, real clear politics, discussions from Reddit. And then on the videos, MSNBC, a short from YouTube and CNN. And then back to organic results, they do not have the actual video

Starting point is 01:05:41 in the top 10, just like everybody else. And even, you know, I tried doing like full interview and you can see when you have the suggested drop down that everybody's putting in full. If you put in full, it does take you to episode 2219 on Spotify as the number one result. So, you know, maybe it's just search is doing what search is supposed to do, which is to get you the trending things, the clips,

Starting point is 01:06:09 and the recaps are more appealing to people than the original. Maybe that's what the algorithms think, because now we have six different search engines, and one and six got it right, five out of six didn't get it right. So you have to believe Brave, duck, dot, go, independent search engines. Then you have to put Bing, YouTube in that group as well that they got it wrong.

Starting point is 01:06:33 So yeah, I'll just call, I want to close with this. YouTube, your search is miserable and I hate it. And it makes me so sad that every time I do a search, you think you know better than me after giving me four wrong links, that then you're going to change the subject in the back half a search. Stop it. Stop it.

Starting point is 01:06:48 Just do what I say. Thank you. It's so obvious to me now that we did this exercise and talked it through. If Captain Nuisance and, if Captain Nuisance and cautious optimism both work through this, and it seems like the algorithms are designed to frustrate users to increase session length and to make the most money by getting you to click on things that you didn't intend to click on in your frustrated search, if you stop searching and you found, they don't make as much money.

Starting point is 01:07:24 I think that's what's happening here. The conspiracy theory that I will allow is this. If you make the little clips more popular than the full thing, people will watch three of them instead of one. And there's probably a higher ad load for watching three, 10 minute summaries and there is watching one three hour clip. Done. Done.

Starting point is 01:07:41 Precisely. In sentence. All right. Well, I do think Joe Rogan turns off YouTube monetization. So YouTube makes no money on Joe Rogan's clips, Joe Rogan's full episode, and they make a ton of money off the clips. That's what it's doing.

Starting point is 01:07:54 We should have started there. We should have gone, all right, forget everything else, follow the money. we've figured it out the clips are monetized and Joe Rogan's interview is not this happens to this podcast and all in as well

Starting point is 01:08:07 if we don't have monetization on I bet you we don't get as much and we should run a test here on this week in startups and turn on monetization so I'm going to ask my team to just turn on monetization for this week's shows

Starting point is 01:08:18 on YouTube and we'll run we'll just test it because we'll still have our ads in but we'll turn it on for whatever this week's episodes are turn on monetization and let's see if they get better search results. People are going to point out that we're running ads twice,

Starting point is 01:08:31 but it's okay. We are going to pause everyone right now in time a quick second for Jason to talk about the launch Cloud Kitchens Incubator. And when he's done with that, we are going to go back to AI on your PC because Google has big news to take us out. Okay, so my friend Travis, who created Uber has a company called Cloud Kitchens. We're back at it collaborating.

Starting point is 01:08:50 And what we've done is we have created Lick, the Launch Cloud Kitchens Incubating. And what it is is the 14 weeks. program so that food entrepreneurs can get a little bit of investment from us. I think we give them 50K at the start and 50K at graduation to help them build and scale their vision for a cloud kitchen. What's a cloud kitchen? What's your favorite food, Alex? Tell me your favorite food at the moment. You got two favorite dishes? Three favorite dishes? Rattlema. Pork geosa, I would say really, really good chorizo tacos. And I'll round it off with a

Starting point is 01:09:27 a healthy peanut butter protein smoothie. Perfect. Let's go with dumplings. Gioza is those dumplings, right? Now, what a great concept if somebody had, let's say, 12 different homemade dumplings, pork, mushroom, maybe duck, maybe some blend, maybe something fun like pushromi in it. And so you get this incredible menu. You have this great idea.

Starting point is 01:09:55 You're going to do. get Gioza. Am I pronouncing it correct? You're pronouncing the way I pronounce it, though I'm now terrified that we're going to get some some mail. Oh, okay, fine. Sue us if we get it wrong.

Starting point is 01:10:06 But get Gioza is your idea. And it's 12 different ones, and you order it like chicken wings where you can get a 12 pack and you can pick in groups of four, right? This is a great idea. You're not going to open a storefront. If you opened a restaurant to do Get Gioza

Starting point is 01:10:22 or Gioza group, the Gioza Group restaurant, man, that's going to cost you a million dollars minimum to test that in a city. But what if you could test it in a $5,000 a month, all in, maybe 10 with all the food, in a cloud kitchen, and then get feedback on your recipes. You could spend $10,000 a month, 5 to $10,000 a month, maybe hit break even immediately or even profitability, or maybe you lose a little bit while you're testing the concept for three months. That's what a cloud kitchen allows you to do.

Starting point is 01:10:54 Now, once you figure out your format, then you could say, you know what? I want to have this. I got it in Manhattan. Let's have it in Brooklyn, Queens, Staten Island, and the Bronx, and New Jersey. And now you have your Gioza made in one of those kitchens, and then somebody ships them to the other five kitchens. And now you've got six locations for $5,000 a month each, all in. You're spending $30,000 a month, and you're making some of these cloud kitchens, Alex, are making one to $5.5. million dollars per kitchen. There are brands like, there's a Goop Kitchen brand now by Gwyneth Paltrow.

Starting point is 01:11:31 It doesn't exist as a storefront. It exists only in virtual kitchens. And if it works, you can sweep the nation with these once you perfect it. Okay. So I, you know, people often joke that like technology has invented something that we already had before. Like, we're going to put self-driving cars together on the road. You invented a train. Okay. Hear me out, though. Is cloud kitchens kind of like a non-mobile food truck? Exactly. Okay. Now, here's the thing.

Starting point is 01:11:55 When you do a food truck, that costs $200,000. I've looked into it. Okay. Just to get one of those set up. Now, I'm sure you could buy a used one for $25,000 somewhere. But I think on average, you're talking about $25,000. Now you've got to park it somewhere. You've got to put it somewhere.

Starting point is 01:12:09 You can be in a cloud kitchen. You could tour one today and start tomorrow. And you don't have to buy any equipment. You're renting it. You're renting the equipment. You don't have to sign a lease. You can come and go as you please. Plus, all of the software is designed to take orders through the Cloud Kitchens platform,

Starting point is 01:12:29 which lets you do direct, let you do DoorDash, lets you do Get Eats, whatever, Postmates, Uber Eats, whatever. And so they manage everything for you, and then they also have salt. So if you didn't buy salt, you can buy wholesale salt from Cloud Kitchens in the pantry section. So all of this equals, oh, and they place them all over the world at the best locations for delivery. So they've already optimized, this is at the 10 and the 405 freeway exchange. This is at the 290 and the 1. This is at the 280 and the 390 or whatever in the Bay Area, right? They put it exactly where the density is, so the drivers know where to go.

Starting point is 01:13:12 And the drivers sometimes will pick up from two cloud kitchens or three cloud kitchens. at the same time, so things get routed easier. When you have 20 Cloud Kitchens in one location, if you ordered the Gioza Group and I ordered like the Peking Duck, they might both be virtual restaurants there and they both go to the same block and it makes everything faster, better, cheaper. Anyway. Okay. So Cloud Kitchens are essentially the infrastructure side of AWS and the distribution thereof,

Starting point is 01:13:39 combined with the ease of distribution that Disturikid brings to you, Gist or Kit takes your music to Spotify, Apple, title, et cetera. CloudKist and take you to Uber, it's both ofments. Okay, I get it. All right. Yeah, that's it. And so we're looking to invest in startups. If you want to apply, CK, as in cloudcchitchens, dot launch.co.

Starting point is 01:13:58 CK. dot launch. com. And you can get to it at launch. com just in the menu bar as well. All right. So if you want to come, start a brand, apply, and we'll invest, and we'll change the world together, yada, yada.

Starting point is 01:14:13 So one geosa at a time, everybody. That's it. That's what we're going to do it. All right, Justin, just take us out really quickly. I want to make sure I got to this because you and I were looking before the show at the recent news that Google is going to follow Anthropic and release, according to reporting from the information in Bloomberg. Later this year, a AI model that can actually run your browser. It seems that Anthropic was first out of the gate, got a lot of the early hype and press here, but Google is coming right behind them. So I got to ask, are you excited by you?

Starting point is 01:14:43 this or does this feel like Sundar is doing some copycat work? No, I mean, everybody knew that this is where we were headed, which is there's a lot of tasks that occur in a browser window, and if you put a language model and you let it control the browser window, you unlock everything. So what could be unlocked here? Shopping. You say you want to order coffee beans and, you know, show me the top three brands, you know, in this price range, and I need to get them today,

Starting point is 01:15:15 and it goes to Amazon, Walmart, and Target, opens three tabs and then let you pick, and you say, buy that one. Or you say, email everybody in my favorites list that I'm having a party, and it goes and does that. Now, the downside to this is, if a hacker were to get control of Gemini, they could launch an attack or this clawed one, and say, anybody who has this installed, please, you know, do this following attack.

Starting point is 01:15:44 Go to their Chase account, open up Bank of America, transfer funds to this. Open up Coinbase and transfer funds here. So you will see in the next six months, within six months of all of these things being released, you're going to see a major hack occur where people's private emails, bank accounts, crypto gets stolen.

Starting point is 01:16:07 It will happen. And so we have to, side as a society if we want to give AI the browser. Because it will get hacked and it will cause significant damage. Okay, but take the first thing you said, which was this is where things are going, with the second thing you said, which is, oh, cybersecurity. And to me, it sounds like if we're going to go this direction anyways, we should be thinking more about how to stay safe versus if we're going to need to have

Starting point is 01:16:33 protection. So I guess, you know, I hope Anthropic and the Gemini group and everyone else who's working on this have that in mind. but you did take some of the shine off of it for me with that last riff. Well, there is like what Claude AI can do and cannot do with their computer interface. So there is a list and think they do not allow you to send email. And so their computer interface has a list of things it can do

Starting point is 01:17:05 and what Sonic can't do when it controls your computer. and so you can look that up. It's a simple list that they shared on social and other places. So I don't think they let you use your password manager, email, or purchase tickets, but maybe it lets you search for tickets, you know, that kind of thing. Yes. I think this could have already been released, and there were startups that did release some of this already, so it's going to be super granular, but here we go.

Starting point is 01:17:32 Yeah. Here is the list of things. So to minimize risks, consider taking precautions, such as, use a dedicated virtual machine while playing with Anthropics computer use beta. Maybe don't give it internet access. Ask for a human to confirm decisions. We're going to need to have some best practices here. I'm not entirely sure if we know what they all are.

Starting point is 01:17:51 But at least Anthropic is trying. I hope that the Gemini team is as thoughtful. They will be because there's a lot at risk. I think you're going to have to click over like 10 disclaimers. And just like Apple has that simple thing in settings that we do. discovered to let people learn. There is a little toggle bar that says, like, let Apple intelligence learn from this app. It's in the settings of each individual app. You can open up your phone, go to your apps, in settings. When you're in settings, there's a button that says, let, I think

Starting point is 01:18:25 the language is let Apple intelligence learn from this app. So, you know, what would make sense is to have each website URL have to be whitelisted and have to be approved. So, you know, this is one of the great things about the App Store. And the value of paying 30% for apps from Apple is somebody's looked at it and said, can this thing take over your computer, download your videos and photos, use your location deferiously? And there's that interface layer, right? So that's why you have to give permission to each app

Starting point is 01:19:04 To know your location To send you notifications to use your webcam And to use your microphone It's a system level decision by Apple And you have to turn it on for each app Yeah, that's what we're going to have to do with browsers Because Chrome is owned by Google Google should allow Chrome or should have Chrome say

Starting point is 01:19:24 Do you want to allow Gemini to read and send email on your behalf. And it should say, yes, no, or with confirmations. Boom.

Starting point is 01:19:36 And if it just said with confirmations, you would come up with an API that on send or on compose, it would say, click here to approve. Yeah.

Starting point is 01:19:47 But with X and LinkedIn, we have seen kind of people being opted into some things at times. And so I wonder if we have the right norms in place kind of at the trust level, frankly,

Starting point is 01:19:58 across major social platforms around the web to do that. And I wonder if Google's going to have the the hutspot and the spine to say we would love to track everything, but we're not going to until you give us permission to, even though we're an advertising base company. Because they are risk averse and they,

Starting point is 01:20:16 when you file a lawsuit or when you do a hacking attack, you go after the deep pocketed folks. There's no reason to sue a small player that has a million dollars in revenue. You're going to go after the player that has $100 billion in revenue because they'll pay you a settlement, right? So they have the most at stake. And this is exactly what the arc of technology is.

Starting point is 01:20:38 Something disruptive like this comes out and you're going to have to be thoughtful every step of the way, just like self-driving. You've got to be thoughtful every step of the way. And I don't know if you saw the Tesla earnings last week. They've increased their margin. The energy business is doing great. Cars are flat. But the most important thing was Elon said,

Starting point is 01:20:59 listen, if unsupervised FSD doesn't work on hardware three, no big deal, we'll just upgrade you to hardware four for free. That was like a very big under the radar statement. What he's basically saying is, hey, you know, there's, I don't know, maybe 500,000 people who paid for self-driving. We'll give you a $2,000 new computer and cameras if we need to. Because four is working, but three maybe isn't working to the level that four is, and he just said, hey, no problem, that would be a $2 billion recall on a million cars,

Starting point is 01:21:36 which is what they made in profits in a quarter. And that's what I was about to get to, which is this is the statement that a company that is not cash constrained can make. Go back in time when Tesla was scraping nickels together from the couch, they could not have afforded to do this. No, but here we go. Your free cash flow does unlock a lot of good things that build customer goodwill because people were worried that they might get left behind, but if you say, we'll fix it, there

Starting point is 01:21:58 go. Well, you know, I use, I think this is, I was talking to a friend of mine and we're having very different experiences with FSD. I was saying, I'm getting three interventions an hour. And I would say one out of six is one that I would say would cause severe damage and or could be potentially life-threatening. And five out of six would have been fender benders or like somebody would give me the finger or something because the FSD did something weird, like didn't know which lane to go in or stutter stepped into a traffic circle. That's typically where my FSD breaks down is like making a left turn, it stutters a little bit or going into traffic circles or unpaved roads,

Starting point is 01:22:38 you know, unmarked roads. Obviously it has to make it has to make its best guess. Putting all that aside, people with the 4.0 hardware from the last year were reporting different things than people with three. And I think now we know why the computer is more powerful. and it's able to process more in the moment, and maybe the cameras are better.

Starting point is 01:23:00 I don't know, or maybe the number of cameras is better. So I thought that was like the most important thing I saw out of that Tesla thing. And then also they're making a fortune off of the margin on these battery packs. Yes. They discussed that very seriously. If you look at the Tesla earnings,

Starting point is 01:23:17 once you scroll past the car bits, you get into the energy storage, energy collection storage section, whatever it is. That's one of the most interesting parts of Tesla's Q3 earnings. If you haven't read it, go just Google Tesla IR and just go read through the report yourself. It's worth the time. For Friday show, I think we should do the crazy cars from China on Friday show.

Starting point is 01:23:38 So let's tease it here. I just want to show one of them. I found a bunch of different accounts. The cars in China that are being made today are bonkers. I'm not saying they're safe. I'm not saying you're going to be driving them here. But there are some really crazy visions. and they've stolen a lot of innovation from American companies,

Starting point is 01:23:58 but they're also innovating in their own right. Here's a sample of two or three things that will show you in detail on Friday. The minivan, I think, is the one I thought was really crazy with the doors that opened like a subway doors. Okay, so this is a car and a veto in one. We'll talk about that. The car separates from the veto. Let's show the next one.

Starting point is 01:24:20 The Ziger Mix is going to cost about 40K, Jason. and it's got really wild stuff. We will show this. And as another point, I was doing work on the land and air vehicle. X-Peng, G-A-C, S-A-I-C, and G-Lie-Ley-Technology groups, Aerofugia, are all working on wild EVTALs that merge with cars. So come to the show on Friday. We'll blow your mind.

Starting point is 01:24:41 Friday, we're going to do crazy car concepts in reality. In China, if you have any of them, I'm Jason on Twitter. He's Alex on the Twitter slash X. You can add mention us or DM us, what you want to have on the. show next live show i think is going to be friday um and uh but maybe who knows maybe we'll sneak one in on wednesday we'll see you all next time bye bye

This Week in Startups - TWiST News: The Startup Redefining Sonos, Influencers, Elections and Google's next AI push | E2034

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.