$NVDA's earnings, media economics in the AI era, Farcaster's mega crypto round, and more! | E1954

0:00All right, everybody. Welcome back to this week in startups. It's our news roundtable. We have two breaking news stories. Nvidia's earnings are out and open AI and news corp have done a big licensing deal a quarter of a billion dollar licensing deal with me to discuss it all Alex Wilhelm the new co-host of this week in startups welcome back to the program Alex hey good to be here I freaking love breaking news I love earnings I love media economics it's like this show is designed for me today I am the happiest kid in school. This is why I wanted you as co-host because I know you love this stuff and I love this stuff.

0:40Let's get to work. This Week in Startups is brought to you by Zendesk. The best customer experiences are built with Zendesk. Qualifying startups can join their startup program and get Zendesk products for free for six months. Visit Zendesk.com slash twist today to get started. Mercury. 90 % of startups fail. Just 10 out of every 100 make it. Mercury exists to close that gap, helping companies succeed with banking and credit cards engineered for the startup journey. Join over 100 ,000 companies banking with Mercury at mercury.com. and hidden layer generative ai is revolutionizing industries hidden layers ai detection and response solution secures your generative ai and llms from malicious attack helping you generate more by enabling seamless and secure generative ai visit hiddenlayer.com twist to learn more nvidia is the bell of the ball it's the most important company in our ecosystem and in on wall street as well so this is a rare thing that occurs where wall street is obsessed with a company and it's that same company that's driving our industry uh so let's talk about it how did they do overall what's the reaction stock up stock down did they meet expectations did they break expectations.

2:16Everybody wants to know what the hell's going on with NVIDIA's incredible run. So somehow they managed to beat expectations and gain more value. And just keep in mind, everybody, that NVIDIA was worth$2.33 trillion going into earnings today. So when I say that it was up 4.5 % after hours, we're talking about tens of billions of dollars of additional value out the company. Jason, the top level numbers are$26 billion in revenue, which was ahead of expectations of 24.7. And it's EPS of$6.12 per share, best expectations again. But the thing that I think really kind of solidifies what's going on here is a year ago, the company had $7.2 billion in revenue, which is nearly a little bit more than a quarter of what it pulled off in this quarter.

3:04So that's the scale of growth here and profitability. It's just, I've never seen a business story like this since i don't know like google rolled out ads for the first time like it's just it's crazy well yeah and just to pause here uh on on this important observation you have 7.19 billion a year ago now 26 billion so it's essentially quadrupled revenue but it's not quadrupling revenue on 10 million in revenue and going to 40 this is adding you know if we take the 7 billion out of the 26 billion they've added 19 billion in revenue yes year over year it is extraordinary we've never seen anything like that and this is all driven uh by one thing ai and large language models and this incredible competition between not only upstarts uh like claude but incumbents like microsoft tesla google you know people who've been around for 10 20 30 40 years this is what's driving it is major purchases of h100 essentially yes i mean so if you want to think about the biggest driver here it is data center revenue as jason mentioned and that was up 23 from the previous quarter and up 427 from a year ago which is it's just a piece of growth that a seed stage company would love to have like you would love to bring that to your board you just don't see at a company that's like the fourth most valuable in the world maybe third now um the thing though that i that i'm stuck on in all of this because i'm not really that shocked that nvidia had a good quarter we all kind of expected that but the fact that it was priced in so it beat expectations but only gained four percent you would think jason that with this much outperformance and overperformance that there'd be more stock market reaction but it seems that people just kind of expected this level of excellence so i'm curious if you're shocked by the lack of a stock market pop here kind of post our needs yeah i mean zooming out if you were to look at the five-year chart uh for this company and i'll pull it up here just uh so people can see it that really is the story if you look at the five-year chart this is up 2500 percent 250 times your money so one dollar would turn into 250 dollars right if you're doing percentages yeah and so it's up 913 dollars if you were buying this stock in just 2019 not that long ago you would have paid 36 for what's now a 949 stock by the way that's what seed investors are attempting to do on their outlier bets so there was this public market company sitting there you know at the time uh its market cap was a fraction of what it is now it's a 2.3 trillion market cap um it's nothing short of extraordinary um and it really is based upon uh the language models created specifically by open ai yes and once people got inspired by you know chat gpt 3.5 it was really what i think people started to 3 3.5 people started to realize hey wait a second it's spewing a little bit of gibberish here but when it gets it right um man it's so compelling right if you're like it feels like magic yes and you ask it hey make me a marketing plan tell me the history of china you know or tell me the history of china as compared to this and whatever all of a sudden you're like wait a second that's knowledge work and knowledge work is being done instantly by a computer and if you've been in the industry long enough you know enough to squint and what i mean by that is you give a little squint and you're like okay where's this going and we all watched dial up i was on a 300 baud modem then a 1200 2400 4800 9600 you know and then eventually you know isdn lines t1s and you're like yeah it just gets faster it only gets better it gets faster better cheaper and when you looked at that every single company said okay it's going to get faster better cheaper therefore what does that do to our core business And that scared Microsoft.

7:17And it scared Google, Apple. Everybody got scared and greedy. When you combine, Alex, fear and greed. Yes. These are the two most powerful, two of the most powerful human drivers and corporate drivers. When a CEO, when a board gets really scared, they take action. Yeah. And when they get greedy, they take action. These are major drivers in business, right? gordon gecko whatever greed is good yep you know and then fear you know only the paranoid survive there's a reason why those two phrases are so iconic right it was intel the head of intel said only the paranoid survive and that's the fear part and greed is good that's the the you know the basic tenet of capitalism this one hit both yeah if you're google we're going to lose the franchise or we're going to 10x the franchise yeah we're going to finally be able to give the answer Those two things combined equaled purchase order.

8:15Yes. And that's going to continue because if you look ahead, they're going to do$28 billion in revenue next quarter. So they're going to add another$2 billion in revenue. Now the growth rate is slowing sequential quarterly, but certainly they're still expecting a lot of growth. Here's the question though that I'm stuck on. I don't have the answer to it yet. All of this revenue at NVIDIA is just cost at your Amazons and your Microsofts and so forth because they're buying these chips on Maas. but they're also working to build their own chips. So is there any chance in the next 18, 24 months that homegrown AI hardware from the major tech platforms, these major cloud providers, can supplant some of their demand for NVIDIA chips?

8:53And if so, what does that do to NVIDIA? I think that's probably a multi-trillion dollar question that no one knows the answer to. I don't think they're going to catch up that fast, but certainly they want to try. Yeah, so there's this concept, it's all priced in. It's all priced in. I think the growth is all priced in here. that's why a four percent pop not a 40 pop if you were looking for those pops that was you know you would have had to have your finger on the pulse of ai back in you know 2022 2023 and then realize the size of the purchase orders being put in and then predicting the fear slash greed dynamic that occurred going into 2023 all of these orders are in there is there are two possibilities of what could slow this down because i think your question is when does this juggernaut you know uh slow down there's two things one you're correct everybody realizes this and there are people making chips and alternatives right and the second is do we overbuild capacity okay so competition's coming but you're correct it's going to take a couple of years for competition to be able to put a dent into this.

10:06So let's put that at a 24 to 48 month reality that people start maybe replacing or considering other options compared to Nvidia. Then

10:20overcapacity. Overcapacity is the more likely short term scenario. I think what's going to happen is the models are getting so good so fast that the open source models, the cheaper models will be able to get certain tasks done on cheaper hardware and then if all this gets built out you need a use case so for self-driving yeah building simulations of self-driving or folding proteins it's never going to be enough capacity but to tag a blog post with what's this blog post about and put the tags on automatically in wordpress like you're probably going to be able to do that with like an open source product that can run on your laptop, or that's close to free, and doesn't require an H 100 or one H 100 could do every blog post in existence.

11:07So that's the more likely scenario is, are there jobs to be done that require this massive infrastructure? And is the software and the companies pursuing that software? Do they have those jobs queued up to leverage all this? And I think that's going to be the question, because you still have to build software to run these tasks what if you build up so much fiber that you don't have netflix or youtube to utilize it right and so there could be a trow or a plateau capacity is so high you built three nuclear power plants and nobody has plug-in teslas yet right like and nobody's running air conditioning global warming hasn't happened yet okay you're burning off all that extra capacity I think we might hit a dark fiber moment in AI.

11:55I don't think it's the majority chance, but there is a chance we could hit a dark fiber moment. Too much fiber, not enough bandwidth being used. Yeah. And that's why I think plateau is the right way to think about it, because we may end up with overcapacity. But if you think about how much bandwidth we use right now, I mean, I have essentially a gig of it up and down, and I blast through my total internet usage. It's insane. Essentially, my usage of compute power and internet have expanded to meet whatever I can purchase. And so I think long term, that's where we're going to end up with a lot of this AI stuff.

12:26But in the meantime, if you are building models for the edge that are smaller and so forth, then yeah, you can use different chips and so forth. But investors are really excited about NVIDIA still. I mean,$28 billion next quarter, tons of profitability. And also, Jason, improving gross margin. The business is not only getting bigger, it's getting better. And that, to me, is just crazy. I love watching history get written. And right now, NVIDIA is doing something that we haven't seen maybe ever. All right, startups. When you're a business, you know it's crucial to treat your customers right. Unreasonable hospitality should be your goal, right?

13:02Great book. And you know who's great at helping you get delivering unreasonable hospitality? Zendesk. We use it. We love it. The Zendesk suite arms your business with all the tools you need to deliver exceptional customer experiences. And it's made just for startups. And they care about startups. It's built so you can build strong relationships with your customers through amazing support experiences and without growing your headcount. So do what Shopify, Squarespace, Uber, Instacart, and JCal does. Rely on Zendesk. They're your partner. Software company Unity, they saved$1.3 million with Zendesk automations and self-service.

13:41and they saw an 83 % increase in their first response time. These companies love Zendesk because it's so easy to set up. It's made to scale as you grow, and they give you all the metrics to make reporting easy, keeping you and your business agile and investor ready. And because you are a startup, they created the Zendesk for Startups program. When you join the Zendesk for Startups program, you're going to get unlimited access to all Zendesk products, expert insights, all the best practices you need, and entry into their amazing community of founders all at no cost with the first six months free.

14:14So there's no risk to you and there's all upside. You're going to learn so much from them because Zendesk does this better than anyone. So apply to the Zendesk for startups program and use Zendesk for free for six months and get ready to scale with the best customer support. Go to zendesk.com slash twist at zendesk.com slash twist. Going back to your chart, the 2019 to today's growth. Yeah. I think it's really important to keep in mind that if you missed Bitcoin and a dollar, you also missed nvidia at 35 a share so congratulations everybody we all missed it twice and this is why i'm placing uh placing bets on promising companies it's pretty good pursuit i have to say because you never know exactly what's going to happen we didn't know that there would be breakthroughs in ai here at this velocity right um but yet here we are and people are ordering these there's a there's a very interesting nuance here too and i want to hear about the profitability and the earnings and how margins have changed if you have that information i do but the thing that i think is also interesting is um the game on the field has made it so large companies with huge amounts of cash cannot buy other companies right we put a regulatory environment that says these companies are too big let's not let them buy you know whole foods let's not let them buy irobot or like a robotic vacuum company right amazon got blocked so you have all of this m a getting blocked so then you have google apple microsoft amazon sitting there with tens of billions in some cases over a hundred billion dollars in cash yes and they're still some of these companies are giving great dividends they're doing stock buybacks like apple and they yet they still have tens of billions of dollars laying right so for them to not be able to deploy it hey you know what let's build some server farms let's spend a billion dollars let's spend 10 billion dollars second mark's like yeah i'm going to spend 20 30 40 billion building out this infrastructure who cares uh elon's like i'm going to build out you know tens of billions of infrastructure so that circumstance led to i think boards and visionary folks saying gotta deploy this capital somewhere they're not going to let us buy it's not like they're going to let google buy airbnb or apple by uber or amazon by uber which would be an incredible combination if amazon prime and uber and airbnb were one company can you imagine uh like how crazy that would be all your commerce all your deliveries everything is just merged together and and because you can't do that maybe reasonably so to get the playing field fair this is just another way to spend 10 billion dollars and deploy it yep for you know that money sitting there in your pocket i remember at one point i was talking to larry and i think they had hit 10 larry page they hit 10 billion dollars in revenue and he said what do you think we should do with all the money and i was like you should invest in startups and you know what they did google ventures yeah and then they were doing google fiber sergey was literally running google fiber in the backyard of google and he was testing ways to rip up the earth and lay fiber faster so he had like teams of people you can look it up he's like pictures of it where they had drills you know those big circular drills that you know rip up the earth real fast he was having competitions amongst engineers to see who could rip the earth up and lay the car fiber faster to try to just understand from first principles how to do fiber and they were literally going to fiber to everybody's home because the the um infrastructure players the cable modem people they weren't moving fast enough and sergey had told me that was their way of telling it was a shot across the bow of the cable companies hey if you're not going to do this maybe we should do it yeah and you know that's how desperate they were to try to figure out how to spend the money so well back then at that time interest rates were very very low if you go back to the last 10 12 years now cash does have some value which probably changes the calculus a little bit but i will say we need more google fibers like I feel like that level of like, hey, we've got a lot of money.

18:20Let's make the world better and business better. At the same time, kind of like, dare I call it a moonshot? I feel like we've lost some of that mojo in tech as the consulting people have taken over. But you wanted some NVIDIA numbers. Let's talk about it. So the thing that I am most stuck on is that the company's gross margins improved from 76 % to 78.4 % quarter to quarter. and that's up just 2.4%, but 13.8 % year over year. The company's revenue up 262 % year over year, while operating expenses, Jason, were only up 39%. And that yielded just some crazy stuff. Operating income up nearly 700 % year over year.

19:06Again, just, yes, yes. Basically what they're doing is they're taking money from Amazon, Microsoft, and Google shareholders, and they're putting it into their own pockets. and laughing all the way to the bank. It's the best business I've ever seen. Wealth transfer. This is it. We've redistributed wealth from Apple, Google, and Microsoft to NVIDIA. The crazy thing is investors are also rewarding the major platform companies for investing in AI. So essentially, they're spending money and not losing value in the process while NVIDIA takes that money and grows value in the process. So if you own stocks, which we all do, i mean it's not a bad day to be around i'm just again curious about when this amazing growth will slow down and then we'll see nvidia be uh stuck a little bit looking for the next growth because there won't be a similar boom in gaming hardware jason i don't think so what's the next act but they are doing a 10 for one stock split and boosting the dividend at the same time so lots of goodies in there uh for shareholders and fans like 90 stock when when are they doing that 10 i don't have the exact date but yeah they're planning it yeah what matters is dow jones industrial average is i believe price weighted not market cap weighted and so i'm curious if this is a play to get on to some bigger indices and so forth as a company that's speculative i mean it's uh one of the great stories of our time and one of the great things i think about this uh whole ai revolution and what's occurred is it's got everybody you know i think the industry was in a little bit of a lull like what next and crypto was kind of a head fake all these promises it was going to be a paradigm shift we can get into that we got a billion dollar uh valuation coming up uh with farcaster uh in a moment on the docket today we'll talk about that but crypto didn't work right If you look at how inspired people are by using ChatGPT and how much value and how many use cases there are, and then you look at the promise and the deliverable of crypto, lots of promises.

21:16Hey, it's going to be better. We're going to have a better Airbnb because everything's on the blockchain. The reviews, the things, there'll be less fees. You'll pay gas fees or whatever. Transactions will be free. But still, Airbnb is Airbnb. Didn't get disintermediated by crypto. but you know ai god everybody has all these ideas for use cases like maybe we could make better images or videos or audio or we could talk to this thing we can make my writing better or you know maybe make me a better developer and they all happened the time between the promise and the delivery and ai is like it's days weeks months like somebody comes up with an idea we have a company howie that like uses ai to book um it helps people book their uh calendars so you and i email we cc howie and it's like uh yeah we should get together for lunch and it's like here are some proposed dates and and then all of a sudden it's booked it's calendared and nobody had to be involved like pretty magical kind of like a faster version of calendly um and the time between when the entrepreneur pitched that and then people are losing their minds over it is a couple of months.

22:23So here we are. While at the same time, the underlying tech itself is also improving. So the use cases are coming quickly, the new ideas are coming up fast, and also the underlying models are improving. And that's kind of, you know, it trickles down to Nvidia's shareholders at some point in time, but it is really fun to see. But on the AI front, Jason, the other breaking news story that we have today is that News Corp and OpenAI have come to terms. And right before we hit record today, you and I were just like gobsmacked by the dollar amount here 250 million dollars over five years not all cash some other things are mixed in there but certainly this is a bigger deal than i expected to see from uh news corp and open ai okay so i have questions obviously news corp rupert murdoch's company they got in a big big fight with google rupert murdoch had those famous quotes so you know google's the enemy kind of thing then he got into it with steve jobs over the ipad and who would own the subscriber and so i would say rupert murdoch is the most old school protecting content of all content creators and all publishers um second only to say the music industry which protects their rights in a major way and scarlett johansson who also protects her rights in a major way apparently so for them to be able to iron out a deal with newsquart who really understands the value of ip um you got to give credit to both parties here to come up with a framework 50 million is like a crazy number a year that's a lot of journalists like the average journalist at wall street journal or whatever you know i think what else do they own at dow jones so they own barons too i mean they own a lot of you know newsquart owns a lot the average journalist maybe at the wall street journals some of the highest paid people 150k 200k sure um the high you know and the average across their local papers probably brings it down to 100k or something like that journals are not like super well paid let's be honest no um they're not a journalist at vox is what 45 but it was 45 now it's 55k i think yeah because they have a union so you see these numbers so lest you think journalists are making a killing the starting salary at vox which is a major publication shout out to jim bankoff who bought weblogs inc for me at aol like 55k yeah like if you're a plumber you make more than a journalist just believe it at that like a lot more electricians make a lot more um you know you probably make similar to a barista or an uber driver being a journalist at most places yeah at most places 50 million that's a lot of journalists yeah a lot of journalists.

25:04If the money does go to the journalism bucket and does not just go to the share buyback bucket. But I do think what it shows here is that if you are going to build large language models, ingest lots of data, and you want to ensure that you have up-to-date information that is trustworthy so you can build search engines, et cetera, or answer the history of China to your earlier point up to today, you need to pay for it. And I'm hoping that this implies that the New york times open ai current legal spat will resolve in a similar monetary fashion and that also that there's a way for smaller pubs to also do well because it's going to be a little bit crappy if rupert gets 50 million a year and everyone's smaller than rupert gets nothing that would be i think an unfortunate situation i don't know how to solve that jason i do okay we have a start we have a startup we're going to announce an investment in shortly um that's going to be a clearing house for groups of publications so on a very granular basis and i won't steal the thunder but we talked about this on uh uh all in i think a couple of weeks ago and so sax and i are are involved in this company and basically you know imagine you know i don't know you own 20 blogs you could put them together and say these are the rights i'm looking for and then people could then you know programmatically included in their long tail of content right without having to get on the phone and do a deal kind of like ascap does for the music industry or like if you have a bar they come to you and they're like oh you're playing music in your bar do you have a license and you're like what's a license and they're like oh let me explain to you you're committing a crime yes by playing music in here um if you just you know you're you have this many covers in your place if you give us 800 a year of 100 bucks a month you can play music in your place and people are like okay and they just pay it because they don't want to get sued yeah the music industry knows how to you know sort of scale this um there's a billion dollars in subscription revenue at the new york times now right they got 10 million subscribers something in that range so they think they're doing over a billion dollars and it's 20 bucks a month so i don't know what their average revenue per subscriber is but i think it's probably 100 or 200 bucks to subscribe to the new york times a year maybe 200 so i just pulled up their numbers q1 2024 total subscribers is 10.6 million 9.9 of those are digital only right um they had about 557 million q1 revenue so a little more than uh call it 2.2 billion a year yeah so they're over 2 billion a year so you know these they would probably have a bigger deal i think maybe than newscorp i don't know but this could wind up being low single digits of their revenue mix but it's edited additive you don't have to do anything to service this it's free it's all profit yes so the way to look at this is let's say there's 50 million in revenue a year maybe it costs them a million dollars to service this so it's 49 million to the bottom line shouldn't cost more than two or three percent to just give access to this stuff right it's not an expensive proposition um that's going to be super accretive if they got a 20 margin if they're making 400 million a year or something and then you add 50 million and it could be like 10 boost to their profitability oh that's why i think that's why i think we could see a virtuous model here which is okay these language models are paying us we've got multiple ones paying us it's like this whole new licensing business for them and if it if it holds up which is a big f but if it does hold up and there's a clean model here then there's no lawsuits amazing um and then those companies can be like you know what we're not as dependent on ad revenue let's we have a five-year deal here okay next deal let's do a 10-year deal and all of a sudden it's like stability in journalism and you and i in journal have been journalists for or you know acted done i've done acts of journalism for a long time you're forgiven don't worry random acts of journalism you've been an actual journalist.

29:06Like we've watched this. What a show it's been. And we've watched all of our friends lose jobs, you know, consolidation. I mean, it's a disaster. And so this could be a path out, I think, along with the Apple news path out, which we should talk about as well. Most startups have tight runways. We know that. And you got to keep your finances in order. And it's complicated, isn't it? You got all these tools, bill pay, invoicing, reimbursements. You got so many financial workflows you got to stay on top of and they're all connected to one thing, your bank account. But the tools aren't integrated. Luckily, Mercury solves this by powering all the workflows from your bank account.

29:45Having these workflows powered by banking is going to let you get more accurate visibility into all the money moving around your account. For example, rather than monitoring all these different tools, you can just check in on your bank account and see a real-time view of every outgoing payment and every incoming invoice. And you really want to control how and when the money moves in and out of your business. Why would you send money to another platform to pay your bills? All that third-party processing, you can pay your bills the moment you need to inside of Mercury so you can maximize your cash flow.

30:16And you're going to close your books faster. You can avoid all the manual data entry and you can minimize all the errors by just categorizing a bill the moment you pay it. All these details will sync with your accounting software, easy peasy lemon squeezy. In the end, Mercury will simplify your workflows and make everything tight. It's going to give you more control. It's going to help you work faster than ever before. So apply in minutes mercury.com to join 200 ,000 ambitious startups that trust Mercury with their business. It's elegantly simple yet complex. And it's going to simplify all your finances so that you can get back to running your company.

30:51Yeah. So on that front, the latest news is that Apple News Plus, which came out and that I think most of us forgot about, is actually driving some serious numbers. Now, the scale, Jason, is a lot smaller. The news here from Semaphore is that the Daily Beast, which I'm sure we've all read at least once or twice, is doing single digit millions a year now in revenue from Apple News Plus, which is Apple's new subscription service. I think they said they're doing about three to four million this year and what really hit me in the semaphore article we have up is that that's more than it was doing on its own subscription business and that's the key thing here this is a really big lever for them yeah see this is a big deal again it's passive revenue they don't have to worry about finding the users etc and uh what a great licensing deal apple uh is making a ton of money from services i pay for some apple one you know i don't know if you pay for that but i have like a family apple one thing that might be like i may pay 500 a year or something it's a big number it might be 50 60 bucks a month for me to have apple tv apple news apple arcade two terabytes shared across five accounts whatever i mean i got a nice apple music and i'm like this is easy for me to spend it's easiest subscription i have and i like apple news i like it because i use apple stocks and apple stocks is powered by apple news so i don't really and then i'm finding i don't know if you've had this experience in the group chats or with your friends people are sending the apple news links i'm starting to see that the daily beast link so i'm getting like hey check this out you know and it directs me and opens up apple news so when i start seeing shares from apple news you know it's getting some kind of critical mess oh absolutely and i the question that i have though is the daily beast doing well i'm glad they have any revenue stream huzzah three four million dollars a year back to your point about what journalists cost and what you know journalism margins are is huge how many pubs will this work for because again if it's only top five or ten we're going to end up pretty top heavy in the world of media and i think media is often best when it's weirdest and therefore most niche so i'm hoping that this kind of works out for everyone else the thing is the ad market is not as bad as people think it is in general but for media that is journalistic it's garbage still and so i'm hoping that these things do add up um did you see the vox news vox turn 10 vox is going towards subscriptions i'm curious if you were surprised by that yes yeah you know what vox was formed out of engadget at aol jim bank off left he i don't want to say stole but he stole the it kind of he he evolved my playbook the weblogs in playbook but stole a lot of the gadget and joystick team and created the verge and i guess the verge was like kind of combining gadget and joystick into one thing like all tech kind of geeked them in one uh and did a great job and uh obviously still does still does um and so when i saw they're going to subscription i did do a tweet about that um it really just has to do with how efficient advertising is in three places the first is obvious google meta tiktok twitter snapchat you know there's like five or six at scale ad networks sure then you have a new emergence of what i'll call checkout shopping cart basket uh e-commerce related ad networks amazon is up like 40 billion i think uber instacart has a billion each um also um doordash crushing it on advertising yeah so these marketplaces are starting to generate my understanding is like five to 15 of their revenue they expect to come from advertising already and i think that number could go up and then finally there's podcasting um which i know i'm like kind of tooting our own horn here because we're on a podcast but podcasting is like really uh started now it's smaller numbers in aggregate though advertisers really love it so you put those three things together and i'll put youtube and podcast together shows sure host red ads whatever those are going to win most advertising and they all have like great tracking and they're very unique in the world uh-huh for different reasons the the scale of the ad networks from google and meta being close to checkout and then host reading ads right and like just really passionate target audiences that's hard to beat that's hard for engadget tech crunch you know the verge it's really hard for them to say like buy our ads and they're like if you're samsung i think i want the amazon ad i want my tv and my receiver and sonos yeah maybe sonos will put a little bit on the verge because it makes them feel good but they're probably better off being at the checkout on amazon or during search on google not on the verge the verge is going to have the least track ability engage is going to have the least track ability which wasn't always the case right those were like the premier places to spend your money so i think that's why subscriptions and these licensing deals they could wind up winning the day i agree with all of that the counter argument is somehow to this day television advertising remains simply enormous and people spend ungodly amounts of money to spray very poorly targeted stuff all over television.

36:41And so to me, I'm just holding everything you said in my hand and going, yes, you want to be close to intent. You want trackability. You want the ability to really understand the ROI here. And then I turn on my television. And if I see one more state format, I'm going to rip my TV off the wall and throw it through the window because I can't stand it. Here's the thing, Jason. I'm not in the market for insurance right now. And yet I have seen in the last week, probably 200 paid insurance ads on television because I watch sports. And so I just I struggled to see like, why can't some of that dumb money land in the hands of your in gadgets and tech crunches?

37:18We could talk a while about why that's the case. um but they do have big scale they do have big scale uh on tv you know and there's like an old school you know group of people who believe in that scale right and it's just it's hard for people to make small bets it's the same thing with venture capital you know it's this week in startups we're talking about it you see some billion dollar fund and they're like well why can't they give like a million dollars to a thousand people and it's like it's a lot to manage for the management over there they would rather do 25 million 40 25 million dollar bets than$1 ,000,$1 ,000 bets.

37:53Now, in a logical basis, there could be an argument that you would hit more outliers, but there's also the human constraint of what a human can imagine and manage. Oh, yeah. I mean, the dynamic of larger funds wanting to put checks into earlier companies at higher dollar amounts is partially why the seed in the Series A market is so wonky right now. And if you're going to do the scale bet, you're going to have to try to beat YC because they are probably the most active, super early investor out there. And think about how much work goes into each of those cohorts, the brand, the promotion. I think they have 150 people making 500 investments.

38:31So on average, they're doing three, four investments for each. And that's just the, I think that's the total staff is 150. Wow. So, you know, if there's 50 partner people, you know, it's like they're each doing 10 investments a year, one a month. That's still a lot of work, right? That's a lot of paperwork. Just a lot of forms. That's why the forms are standardized and they don't negotiate them, by the way. Yeah. And we do that here. I forgot to tell everybody you're here at launch doing the show with me. So you're going to get some inside information, as it were, on startups and some education.

39:00So I'm really excited for that for you. Here to learn. AI has changed the startup game forever. And companies that are slow to leverage this technology are going to risk falling behind their competitors. If you're not deploying AI inside your startup and your competitors are, I think we know what's going to happen. They're going to get more done. They're going to move faster. And generative AI is so powerful, but we all know it comes with security risks. There are tons of threats that can cripple your AI investment. So how do you accelerate AI adoption while also managing these risks? Well, Hidden Layers' AI Detection and Response Platform is your secret weapon.

39:34It's a security solution designed specifically for AI and large language models. It gives you and your team the tools to see and stop confidential information leaks and to prevent malicious code injection and safeguard your models from theft. It's trusted by leaders in finance, technology, healthcare, and even the U.S. Department of Defense. Hidden Layer helps you generate more by enabling seamless and secure generative AI. So here's the easy call to action. With Hidden Layer, go from pause to possibilities and step into the future of secure AI innovation. Learn how to protect your generative AI today by visiting hiddenlayer.com slash twist.

40:11That's hiddenlayer.com slash twist. where are we going to go next? We got NVIDIA. We got the News Corp. There's a lot of other things on the docket. What do you want to go next, Alex? I think we got to talk about Humane AI and the company potentially being up for sale because there's a lot baked into the story. It's a media story. It's a review. It's a hardware story. It's AI. And I got to say, I'm not that shocked the company is looking for an exit path. After its pen came out, everyone's seen those little pens that Humane shipped. They were expensive. They had a subscription. They were ambitious.

40:43perhaps to a fault and the reviews are pretty negative and then what feels like 20 minutes later jason the company is trying to sell itself for 750 to a billion i don't think it's going to get that but first reaction hit me all right so um like you said there's a couple of vectors here hardware is hard consumer hardware is hard why is it so hard well you have somebody like apple who's just so good at it that they're going to own the category you got michael dell and Dell computers you know I love them you know I always buy Dell and they're just great and like when I buy a Dell or my MacBook Air I have like a like you I have my gaming PC over here I got my Mac here when I'm doing creative stuff like the show I'm on Mac when I'm working I like to use my Windows machine sometimes my giant 50 inch Dell monitor and I like to play a little Age of Empires 4 once in a while when the kids go to bed and I need to blow off a little bit of steam it's hard to compete with people at that scale right and then there's a long tail of whatever interesting thing you make like drop cam or nest cams you can go buy four you know cheap hardware versions of that for like 199 or you can buy the 199 one from google yes it's a race to the bottom so it's risk the bottom you know it's it's hard hardware is hard is what we say in the business vcs don't like to invest in hardware and all you have to do when you want to judge a category is look at exits and if i were to ask you what are the three biggest venture exits for a hardware startup in you know from 2008 till now after the great recession take a minute to think about hardware startups that you saw and when you were you know editors at crunch base and tech crunch go ahead uh nest nest which was drop cam 500 million to google right uh danger to microsoft back when they were trying to get their mobile life back together i forget the year on that that was that's earlier yeah that's earlier okay okay um uh it's not not a venture exit but i mean in microsoft and nokia back in the windows fun days okay yeah giants i'm really digging here uh i mean maybe framework the laptop company today is doing well hasn't exited but is doing well framework we should talk i had them on the pod i really love what they're doing conceptually um the ones you miss were two um and the reason it's so hard is because you kind of forget about them gopro and fitbit fitbit bought by google right they had a really hard time and then you had of course uh gopro which both of those i believe fitbit went public before being bought and gopro then you could add a couple more hardware ones that are still in play peloton went public uh raspberry pi raspberry pi is open source i think i'm not sure if there's a startup behind that they're going public on the london stock exchange they're raising 40 million yeah i stand corrected okay i just saw the other day well that's actually an interesting one for us to double click on because it's like a platform i wonder if they make it from the hardware who knows yeah um so you start to put all this together you can understand why most exits that have occurred are in the low single digits the biggest ones are in the low single digits gopro peloton fitbit drop cam which became nest those were all 500 million to 4 billion ish there was a little weird spike with peloton of course during covid where it went to 30 billion which made no sense so that's why people don't like hardware that's why this is such a tough space these people have a pedigree so then the question becomes does anybody want this product does anybody want it what what as a consumer do you want it no and the reason why is not actually the hardware because i think the actual uh thing they designed is small enough and sleek enough that i can see myself wearing it i mean i mean i'm not wearing an apple watch every day which is enormously dweeby yeah you are too yeah uh i think the problem here is that hardware is hard because you have to nail two things at once software and hardware yes and in this case i don't think the ai software that underpins what humane has built is sufficiently good enough to warrant the high cost for the device and the subscription and so look i did not buy the vision pro i'm going to buy the next generation of it because i have more money than sense and i like gadgets i don't have a lot sense just to be clear um but in this case i wasn't even like i wasn't even that excited about it and i think that's that's the lethal thing here and that brings to me the question of who wants to buy it the company what is the value that they have put together what's the ip i i don't know i don't want to be i don't want to be rude but i well okay so that is a good segue about not being rude there is uh i guess a debate or a micro kerfluffle i'll give it a micro kerfluffle it's like a little donny brook it's not a major one um that marquez uh from youtube and maybe the verge kind of trashed it as a 1.0 um and i think they put like in the headline the worst product i've ever reviewed i just want to let people know how how headlines work it's just a public service announcement announcement if the headline is the best ice cream i've ever had the most disgusting ice cream i ever have people will click on that more than i tasted ice cream yesterday so the most disgusting ice cream the absolute best ice cream i've ever had and so people can be a little hyperbolic in the headlines especially on something like youtube and so marquez did the worst car he ever drove which was i think the fisker garbage and i think they're out of business now and then this device and then people countered that with oh but they're being mean to innovators and startups and founders okay this is the big leagues folks if you don't if you make a movie you don't get to make a movie for the public and then not take criticism of it quite the opposite you need to embrace and love the reviewers for actually giving you attention now if you want to say this is a developer preview disclaimer disclaimer disclaimer this is for developers it's not intended the price point we're going for is 299 not 999 but that's not what humane did humane came out with it like this is a finished product and also the rabbit we had the founder of rabbit on the pod they came out with it saying like you're going to be able to do you know order your instacard and order your uber people got it They weren't able to order their Uber.

47:49You got to be careful in how you frame these things. If you want it to be a developer preview, say this is a developer preview. Put it on the box. Put a disclaimer. Say you're only selling a certain number of them. And then when you do the demo, say, this is things, these are the things we hope developers make for Rabbit, for Humane. And this is what it might look like in a couple of years. You basically play it down, play it down, play it down. They didn't do that in both of these cases. No, and they're going after things we already know how to use. If you think about the pitch behind a lot of this AI hardware, it's keep your phone in your pocket, which honestly, great.

48:28I would love that. But then you have to really own up to that and earn it because people do already take their phone out for a reason. Now, the Rabbit R1 was 200 bucks, no subscription. That gives them a lot more space to ship something that's not fully baked. seven to eight hundred dollars for the humane pen plus twenty four dollars a month in subscriptions so it's it's netflix every month for this thing that's a lot harder and i i do think that there is something to be said for people not understanding the role of media in the world of technology products because people have become so tribal that they think that any criticism from anyone in the world of media in general is evidence of bias or haters or whatever but marques is is well loved and trusted because when he talks to you in his videos he's telling you what he really thinks and the headline was spicy absolutely but i don't think his review was actually that far off i watched the review it was not like the review was disastrous it was just like uh he actually says hardware is incredibly slick there's a lot in here to like that's literally how he leads and then he's just like but it gets a little bit hot the battery you have to change constantly it hangs everything it felt like a iphone 1.0 review right if you go look at the iphone review it's like this thing's pretty stupid there's no keyboard you got a hunted pack the screen's tiny it lags it crashes like it wasn't as magical as it was it wasn't like you know even with the reality distortion field i think people were a little critical of it too they were i mean it didn't have 3g that was a big point and don't forget when the iphone came out it didn't have an app store the whole idea of the original original original iphone was yeah the mobile web is going to be amazing and everyone's going to build great websites and then we ended up with the app world but i mean really it was the iphone 3g that i that i think really took that from this is a niche thing to everyone's going to have one of these until they die and it turned out the app store and apps by developers using the platform they discovered like the really cool things to do with it like angry birds or a flashlight you know like there were like silly things but people don't remember that the flashlight was an app you bought it wasn't built into it right right and then eventually we saw a major platform company slowly eat bits of the market um which happens every time out there but i mean certainly apple is not trying to replace the entire app store um but we actually we have a bit of a startup story here on that note because we were talking about um the recent microsoft event bill and they're talking about ai pcs they're going to put neural processing units in these they're called copilot plus pcs uh i think joanna from the journal calls them ai pures we'll see which phrase takes off uh but jason they have a new thing called recall coming out that's going to let you go back through your computer and see what you have done in the past and as it turns out on this very show a couple of times we've had a company on that was called i think rewind ai and you've sat down and talked to them about this and so i'm curious do you think this is like microsoft coming late to the show and they're going to like take the market or is this one of those moments when startups are going to be able to fend off the big tech incumbent number one it's a pretty terrifying um concept your entire screen is recorded and you know like uh as i talked to uh the founder about this i had profound concerns i'm like well what if you have signal on your screen it's like well that's your choice it's like recording your screen i'm like yeah but the expectation is signal is disappearing messages or i don't know pick telegram whatever you use people are not expecting to be recorded so you're kind of covertly recording people and he has a different view of privacy um like many tech people do which is we don't care or you know uh like you know uh you gain so much that privacy doesn't matter as much because you just make that trade-off i guess that would be the fair way to say it's not that they don't care but they think the trade-off's worth it i disagree profoundly you know with these ai pendants that people want to wear because one of the things um and what was the name it was dan was his name dan stroker yes dan and he also was going after the pendant market there were two different companies we had on yeah the idea behind the pendant was you wear the pendant it records every conversation and i'm like that thing doesn't have a record light on it does it like nope it's like okay so you're covertly recording me like michael cohen like recording trump like nobody wants this by the way nobody wants 24-hour surveillance people who wore those goddamn google glasses the first time got punched in the mission for a reason yeah like nobody wants you walking up with a vhs camcorder and being like what's up you know you will get punched and nobody wants that covertly so i i think it's a little crazy these concepts okay i think consumers don't want them so i could be wrong we're going to find out because a couple things have happened since we had dan on the show first of all they have rebranded they're now called limitless ai and what they've done actually i i watched um dan this founder discuss microsoft's uh recall feature because people of course were like oh my gosh you know what do you think about this and we have a we have a clip we're gonna pull up in a second what i think's really cool is they have moved away from what microsoft is now doing microsoft is saying it's your device everything will be local and what you do on your computer is going to stay there it's for you and and okay okay i'm cool with that i mean that's how apples work right like your stuff is stored local encrypted even apple can't get to it as we saw with i think the san bernardito shooting yeah they wouldn't open it for the government and they had to use an israeli company i believe to unlock that iphone because apple's like tim cook's like yeah no we don't open phones sorry yep but now limitless ai the startup that's rebranded has built what they call a private personal cloud.

54:30So essentially, it now works across your devices, and gives you your own repository, which is great until it gets hacked, in my view. But let's go ahead and take a quick look at this clip, everybody, and we'll see you on the other side. It's easy to imitate the outside of a product, imitate what looks like the way the product should work. But we can't imitate as the learnings, what took to actually build the product, the innovation, the core problems you're trying to solve. You know, we've evolved a lot in the last year and a half from the learnings we've had from our users. In fact, just recently, about a month and a half ago, we completely launched a new product called Limitless that took everything we learned from Rewind, invested it in an entirely new architecture that we think is fundamentally better.

55:12In many ways, we actually felt Rewind, the approach we took, that we took, that Recall is taking now, we found it really limiting. And we knew that users really cared about being able to access their data from any device. they cared about being able to use the best ai models and they really cared about unlimited storage and at the same time if you could deliver on those guarantees and give them the same guarantees they had of privacy that they had with a local app like rewind but in the cloud then that would be a breakthrough oh that's great but back to your earlier point often is cloud is that an oxymoron uh kind of i think so just given my view of cyber security which is that everything's going to get breached eventually, forever and forever.

55:51The thing that's not fixed by having a private personal cloud is if you're recording with your pendant, whatever else is saying, you're taking their words and putting them into your cloud, you're taking them. And so I'm very curious about that. But the consumers will decide. And that's why I think this this recall feature is interesting for Microsoft, because they must have known that people were going to go as you did when you first heard about rewind. Oh, my gosh, this is invasive. And they went ahead anyway. So presumably they've done the research and the thinking. So we'll have to see if folks actually use it.

56:22I probably won't. I think, you know, every time somebody proposes the Truman show, somebody tells them, please go watch the Truman show. And we live in public. It's a nightmare. Go watch the Black Mirror episode. It's it seems like the Truman show is a positive thing. Then you realize it's a nightmare for Truman. yes it's great for everybody else gawking it's tragic and anxiety producing to put yourself under that much scrutiny that's why people right now love the brave browser i i made the brave browser my default on everything really i don't want to be tracked i i have it you know i have my shields up i have i pay for their vpn and i pay for nord vpn i have double vpns i pay for two vpns uh and nord vpn has a double vpn feature i'm just paranoid i'm going to get hacked i have like confidential documents you know from startups and stuff like that you just i'm vigilant about it so i think there's like a trend to take your cyber security seriously um and these kind of things there is there is a benefit to it we have an investment in a company called uh recall ai paradoxically and um it's a browser that plug-in that allows you to um make a summary of the page you're on and then put it into your knowledge graph so you have to actively do it and i ask them actually to um when i have a browser session put it on record everything in my browser session so that every page but what i would do in that situation is i would have a browser that's my knowledge one that i wouldn't do my personal banking in yes i wouldn't do personal communications and so i would take like a browser instance now i'm a vigilant user and i love that idea of my entire of this like powerful browser history where like oh you know i had read a story about that etc so i do like that you know delicious bookmarking on steroids with ai i do think there's something there for that but um i just don't like this idea of recording your whole desktop it's going to get hacked just like everything else uh there's no such thing as like a cloud that's perfectly you know private i think that will get hacked too eventually um and so i i do think that there's a big danger here and then you might also be educating microsoft everybody says the language model is not going to be informed by this but then i saw the slack kerfuffle i don't know if you saw that i did where slack was saying they were training their data on yours and you had to opt out of it so i think the industry needs to remain vigilant on all sides here like and be very thoughtful about this um because there is going to be there are going to be great gains and there's going to be great losses the fact that everybody knows where you are at all times because of gps gps is so powerful i'm not ready to give it up but i do have concerns about being tracked right did you see that article recently about how there's a new norm in like dating relationships to have like find my iphone turned on so you can kind of see where your partners are or partner is and you of course people have been caught cheating and and so forth but it just it just blew my mind because i don't know when i was a child we like left the house and no one had phones and we would just go and and that was you know you expected to not be not be a dumbass and get hurt and now like we're tracking adults like i just it feels too much today i am claiming you for the gen x uh in the gen x draft i'm claiming you alex i think we're claiming you and we're willing to give up like i'm not sure who in gen x we're willing to give up but i'm sure there's somebody annoying in the crowd we will give to the millennials in order to get you into the gen x bucket we i was explaining this to my daughters because i told them get out of the house on saturday you can come back in two hours and they're like what are we gonna do these are eight-year-olds and i said go play in the backyard there's a tennis court there's a tree house there's this go play you can come back they're like they came back every 15 minutes to ask me what time it was and i was like guys you're not getting your ipads you're not watching harry potter for the 17th time outside and you know what then all of a sudden i come out the magic not half an hour after they're painting they found the paint sets they got these are eight-year-olds the twins they found the canvases they're painting they come back they got paint on their face pants they got all they get into all kinds of good trouble and i said to them imagine when i was a child we were kicked out of the house at 8 a.m we were given a sandwich or 10 bucks and we were put in the streets of brooklyn when they told us to come back the rule in brooklyn was you can come back when the street lights come on yeah the street lights were time to come on at dusk so streetlights come on you come home it was like the bat signal for us and i do agree like we're the truman show will make you crazy i had a friend josh harris who did a company called pseudo back in the web 1.0 days and he did a experiment called we live in public um or quiet and there's a andy timoner documentary uh called we live in public everybody should watch it because he went mad because he put in his loft in new york 50 cameras and he put himself under surveillance and then people were watching him all day all night in his shower everywhere in the bathroom it just it was crazy and he literally went mad from it just like in the truman show yeah no spoilers but yeah i don't think humans do well with constant surveillance like you know just just to record this this morning i had to think about like do i shave today what shirt am i wearing like which glasses am i like just the preparation of making sure there's nothing untoward in my background and that's to record something in one place to one camera in a controlled setting now imagine your house everyone's weird right yeah of course yeah i don't want people want yeah i mean this should be obvious um be careful what you wish for and you always have to ask is the juice worth the squeeze and what's the worst possible scenario and you know the worst process is really bad so i i do like the idea though of you know some amount of watch me work and i'll get better at my work and so shout out to david sacks and his partner evan for glue.ai their new slack killer it's watching what you do at work so you can instead of people saying here's what i got done today you're you're now in our corporate slack i just ask people like please do your start of week and your end of week what you plan to get done this week what you got done start a day end a day it's just a ritual for a remote company five minutes in the morning five minutes at the end of the day 15 minutes at the beginning of the week 15 minutes at the end of the week you add all that together it's like an hour of time one hour of time just to have like a gsd get stuff done methodology and accountability to your team members and now you can just ask slack hey what is alex working on this week and be like oh alex was working on the new ticker alex was working on producing this episode alex doing this it will know that automatically in in glue.ai the slack killer that sax made it knows that you can ask questions to your corporate chat and be like who was the guest we had on or did we ever think about having this guest on it's like oh yeah you know we invited humane back then here's a link to that thread where we talked about inviting humane because we've been trying to get them on the program for two years um the end yeah they should come on they should i mean we i think they're gonna have to do a post-mortem at this point oh man not to be dark not to be dark so i actually um helped put together the notes for uh the glue show you're you're discussing oh and uh thank you for that oh yeah no worries but the thing that hit me while i was looking at that is this is like this is not something you could do with Slack.

1:04:25Because Slack right now is doing, what if we take an existing product and we try to thread AI through? Glue is, what if we just erase the whole freaking whiteboard and start again and build with AI from the foundation up? And that is, I think, a very compelling way to rethink enterprise software that might be more useful than what we're seeing now, which is a lot of people trying to put a fresh ai code of paint on an old car you know and i wonder how much of the b2b software stack that we all live in use day to day needs to be torn down to the studs or further and rebuilt and i'm i mean we'll see how glue ai does super humans new ai feature is also pretty dope but like we'll summarize your inbox kind of cool you can ask of questions that's just coming out notion has a built-in ai that's pretty dope um and so i do think it's going to be table stakes soon and it's just going to like things that would have taken you 15 minutes or five hours with your inbox or your knowledge graph or in your you know corporate messaging that you would just be like i don't have time to do that are just going to be automatically done so again the great efficiency of corporate america and our personal lives is going to hit an all-time high because of a ai it's going to be bonkers how efficient we all are uh just even researching stuff you're going to be able to like i don't know if you've noticed this but as a an editor and as a writer even on a great publication like tech crunch where you're hiring people based on their aptitude in those and their achievement in those areas everybody seems to be getting really good at writing like everybody's a good writer all of a sudden and i'm like you know what my whole life the majority of people were bad writers and what made me special was i was a great writer and the difference between me and other writers you know an average writer was just tremendous and people like hey can you read this for me and write it for me because you're a great writer and now i'm like everybody's pretty good everyone's pretty good the thing though the the the reason why i'm not concerned about my ability to send my child you're a great writer well one thank you but i mean the thing that i think matters is ai kind of all sounds the same it has this like it all tastes like chicken to me and if you look at like grammarly or even just like how notion wants you to write or docs, they all have a very, very straight line approach to writing.

1:07:05I like to mess around and have fun and abuse grammar a little bit and throw some words that they're going to say are too hard in there. And so I think that there's still going to be plenty of space for humans to like, to put some, some riz on it as the kids say. But another thing that I have learned is that people hate writing. And so they want to use AI to do more of it faster and probably better. so i agree that it's raising the floor i just think the ceiling is still pretty hot i think the ceiling's high you're right because i am a big fan of the m dash it's like my big comedic device i'm like blah blah blah blah blah blah m dash yep drop the bomb and everybody's like it's not a jacal essay without an m dash or two and uh you know like they're just like make it a new sentence i'm like no nope no nope it's not a new sentence and they're like you can't end on an ellipses i'm like oh yes i can oh and then a preposition and an ellipsy i'll throw brackets in there a long long time ago in a galaxy far far away ellipsis like we were raised on this like back to gen x like we can go with an ellipsis it's not going to kill anybody that you know no you gotta have fun man that's that's one thing that i'm really this is not on topic per se but like life should be fun we're here once we should we should enjoy it and i want to compound sentence make it make the sentence too long yeah compound it three times i'll make it make it that you have to read it twice sure why not give me the run-on one sentence paragraph that says exactly what you want and tells me how to say it as i read it don't just give me the bullet points yes yes um okay we're a little low on time but let's do so quickly a thing that you I are both now members of, which is Farcaster.