Alphabet Breaks $100B Barrier, OpenAI's Rumored $1T IPO | Grant LaFontaine, Chris McGuire, Max Junestrand, Christina Cacioppo, Lin Qiao, Ilan Twig, Taranjeet Singh

29:47I'm thinking Bezos because they had a cash machine with Amazon. What did he do? He didn't return a dime of capital and had gross net profits of zero for a decade while he was building AWS. Yeah, but he built AWS. Yeah. And the last time that Mark invested this heavily in something, it did not deliver ROI. Sure. And so I think people are just a little bit scared. That's why the stock's more volatile. That's why it traded down 10%. This is fair. Google is up. Google's up 10%. Microsoft traded down briefly after earnings. But ultimately, I think if you're investing in this company, you have the sort of protection knowing that this is fundamentally just a fantastic business.

30:36and you can pause, you know, they can sort of pause this like crazy spending at any point. But again, they're doing a bond offering that got announced this morning for$25 billion. And that's after they did another deal with Blue Owl for like something like$30 billion a couple of weeks ago. And so Zuck is levering up. He's going all in. And investors need to be tolerant of that risk. And there's plenty of people that are going to say, well, it's fine. I'd rather own Alphabet. I'd rather own Microsoft, et cetera. Bucco Capital bloke is also, seems like he's on your side here. He says, excited to buy Meta in the 500s.

31:24The stock, of course, is trading in the 600s. So he's joking that it will fall down. $666 a share. Is that a good thing? Is that a good omen? Is that a good symbolism? It's a devilish omen. It's rough. And Buko Capital says, the only challenge will be having the courage to do it when it gets there. And he shares a pretty good visualization from Gemini of what's happening right now. Mark Zuckerberg is in some sort of convertible, driving off a cliff, wearing a VR headset with piles of cash. Wall Street in the back. Flying out the back of the car. I mean, to be honest, I think this makes him look really sick.

32:04It does. And this is exactly what I want out of a big tech CEO. I want risk. I want risk. And I don't know. I feel like you're making a ton of good points that I agree with this year. I'm not saying, to be clear, that I'm not riding with Zuck. I will always ride with Zuck. Exactly. I'm just saying I understand why people are like, you know what? But I can't really stomach a drawdown. Some people are going to just rotate into the other hyperscalers. I'm just trying. Tyler, what do you think? I mean, people were calling for Sundar to be basically fired for a long time. I know. Until basically this year.

32:41Yeah. So I feel like you're exaggerating how much Google has been this winner the whole time. No, no. No, no. I completely agree. This was my point about when they rebranded to Alphabet. like everyone was like you're not gonna cure cancer and create self-driving cars and also build agi and build a cloud platform people were still saying sundar like should not be for 10 years for 10 years yeah people were saying like it's not really working and now finally people are like okay it actually worked you have a diversified business congratulations okay okay breaking news yes this is gonna rock your world no way taylor sheridan and peter berg team on paramount and activision's call of duty movie let's go let's go hit the gong for that the creators that made yellowstone hell or high water and lone survivor and sicario i'm so excited for the call of duty movie you know they should also make a vanta movie automate compliance manage risk prove trust continuously vanta's trust management platform takes the manual work out of your security and compliance process and replaces it with continuous automation um that's incredible uh wait so what what we need to get someone on the show who understands media because i like this all seems like it's happening very very quickly we understand neo-factual media new media corporate media legacy new legacy media no i don't understand neotrad media the neo conglomerate is entirely new to me i don't understand what ellison's doing I don't understand what Taylor Sheridan is doing in reaction to it.

34:18It seems like this is a reaction. Basically, Taylor Sheridan left Ellison's new empire, left Paramount, and said, hey, I'm out and I'm going over to NBC Universal. I'm riding with the Comcasters. And now he's saying I'm making a Call of Duty movie. Is this all tied to him leaving Ellison? Or is it something else? I don't know. I'm sure this was in the works for a while. Potentially. Was it? I don't know. There have been like a flurry of Taylor Sheridan, like little leaks and news items because it seems like he's in a negotiation with David Ellison over, I mean, he was until he laughed, but he's still with Paramount at least for a couple more months, maybe a couple more years.

35:11He still has projects in flight. You can't just like, you know, completely dip on your production company. But we really should get someone on the show to figure that out. Let's go to more haters, more, you know, posters on the timeline from Jay Bultard. This AI stuff is so dumb. Zuck is repeating the 2022 sinking of a great money machine in meta with wasteful spend. The metaverse sucked, and AI is nearly as dumb with all this CapEx as investors tell him they're sick of it. Yeah, and part of the, I think, part of the blowback here is that the first AI product that Meta announced from their new team was MetaVibes.

36:01Incredible. and uh i know you're i know you've been putting in hours and hours a day on the meta vibes app but that didn't send a strong that does not inspire confidence yeah right because it was they if they had launched sora people would have been saying like wow yeah they're absolutely cooking and i think there's two audiences that the meta vibes app didn't hit with one was okay the extremely online technical teapot whatever you want to call them those folks are like well this looks like white labeled mid journey because it kind of is. And the app just doesn't feel like super polished or like crazy.

36:37Oh, wow. I've never seen these like UI paradigms. So there's that. And then there's the, the, the, you know, the, the teapot critique of like infinite jest and slop and all this stuff. So like they weren't hitting with like the insiders, but then simultaneously, it's not like you walk down the street and you were like, oh, wow. Like this is actually breaking through with just like everyday random people in the way that threads has. And so So I don't know, maybe in a year we'll be like, oh, wow, they actually figured it out. They funneled a ton of people to MetaVibes, and it's as successful as Threads.

37:08That would be impressive. It won't. It seems like a tall order. Yeah, it won't be. It won't be. But the thing about Sora that stood out, and even though I'm not bullish on Sora as a media consumption platform, it had this innovative, super viral feature that was fresh and new. And so I think if Meta had launched that, everyone would have been like, yeah, spend the money. Because at least you're on the frontier of the slop. Like you have the purest slop. Yeah. And Sora benefits Instagram weirdly because people are creating more content over there and it's driving engagement. And also just the Sora app.

37:47It did not inspire confidence. The idea of cameos and putting those things together. Yeah. And the way it was like all the little details. Yeah. It was like, okay, this is unique. and like there's clearly a ton. Like this doesn't feel - Did you see Cameo suing OpenAI? Like Cameo, the celebrity video. Ah, that makes sense, yeah. By a celebrity video platform suing OpenAI over the feature, which I don't know. Wait, the feature or just the name? The naming, the naming. The naming of the feature, which feels like a stretch to me. Yeah, I don't know. The other thing - The problem with brand names is if you create a name that is too descriptive of what you do, you can't trademark it.

38:27So if we were to call this like the news show, we couldn't have any intellectual property around that. Did you see this engineer said a couple days ago, just submitted my resignation from Meta. Some might say that one year at the company is not that long, but I survived like three rounds of layoffs. So I think it's fair play for what it's worth. My team is great and it was an overall positive experience, but there are other better places to build AGI, which hopefully won't be used for stepmom chatbots and infinite slot machines. Wow. Scathing. I wonder what the play is. Do you think that's like grow your account so you can be a founder, launch something?

39:08The whole, I don't know, the slot, it doesn't seem, I don't know, it's tricky. Yeah. So Zuck is certainly, I don't think, going to take his foot off the gas yet. No, no way. The vibes will have to get about 10 times worth. There's new Reality Labs hardware shipping every year. Like, there was massive pushback, and the stock, like, sold off by, like, 50%. It was, like, really, really rough for a while. Wasn't it – didn't it trade down to something like – Yeah, like$300 a share or something. No, I think it was – I mean, so – Yeah, look at the five-year chart. At the bottom of 2022, it was$90 a share.

39:51That's what I'm saying. Rough. And it was up at 300. So it traded from almost 400 to 90, which is down 70%. Like that is an insane drawdown. And then it just built right back up and it was bigger than ever. And so I don't know. I feel like people might be filling out apology forms for him at some point, even though it looks rough right now. just something's going to wind up clicking and also the core business is just going to keep ripping so you just get so many more shots on goal it's just an endless an endless you know stream of shots on goal when you have a cash machine like the family of apps yeah also the stock's up 20 in the last six months even even you know factoring in this 10 drawdown so i don't know i think i think uh investing in ai is not the craziest thing you could do in 2025 as a hyperscaler CEO.

40:51Anyway, Polymarket has some OpenAI IPO market up already. So no IPO by December 31st, 2026 is already at 55%. We will certainly be tracking this. But if they do get out, the market is pricing right now between 1 and 1.25 trillion in market cap, but 6 % chance that it's up at$1.5 trillion, which would be crazy to go out at. Not technically the biggest IPO of all time. I believe that belongs to Saudi Aramco. So oil is still dominant relative to AI, but still would be one of the most insane tech IPOs in history, obviously. Quickly, before we move on, graphite.dev, code review for the age of AI. Graphite helps teams on GitHub ship higher quality software faster.

41:45Did you want to watch this video? Toby yesterday. Oh, yeah. Wait, Lootkey? Yeah, Toby, we covered his post where he said, having some AI follow you into your Zoom meetings or Google Meet for taking notes is the digital equivalent of showing up to a meeting with your fly down. I said, I never let them in. I just let them sit there in the waiting room. Oh, I forgot. Sorry. He responded and said, I'm actually very pro meeting recording and AI summarization, but I'm not okay with bots joining as fake humans to accomplish this. It's a meeting between you and me, not you and me and some startup's viral growth strategy.

42:19Granola is great. Gemini does this well in Google Meet. Hypernode is great and fully local, but use them with consent. My tweet is about how ridiculous and self-important it looks when you show up to a meeting with random bots as entourage. Yeah, I think... I do remember seeing there's a viral screenshot of somebody who showed up. Is this the screenshot where somebody brought like five bots? It's just hilarious. Entourage is actually kind of sick. Yeah. Wait, Tyler, this would be a good hack project. Like viral entourage on your calendar that when you join a meeting, it just joins with like 25 other people.

43:00And it just takes your name and just automatically does like your chief of staff, your head of security, your PR person, your secretary, your driver, your dietician. And it just joins every meeting with 20 different bots and just makes you look super important online. Maybe that'll happen. We should move on to this video of Jensen Wong, CEO of NVIDIA. Oh, Jensen. I was going to go for the guy from Snowflake. Oh, okay. Yeah. I mean, let's watch. The Jensen video is only seven seconds. So look at this. He's having beers with some absolute dogs. He's hanging out with the CEO of Samsung, the CEO of Hyundai, hanging out in South Korea, just having some beers, crushing some beers with the boys.

43:49This is how you do it. People will call it a top signal. I mean, this is going to go so hard in a top signal vibe reel that we will definitely create when the time is right. But I mean, there have been crazier top signals. And when NVIDIA was at like half the market cap it is at now, uh jensen was signing autographs people were like signing signing a woman's shirt i'd have done that too i don't know but it was a crazy moment and everyone was like this has got to be it so i think jensen is just consistently putting out crazy clips that read his top signals to confuse the market keep you guessing that's right that's right he's a doubt doubt me again He also says that Korea has the best fried chicken in the world, taking shots at Colonel Sanders.

44:39The rough out here for America. I don't know why Korea has the world's best fried chicken, but Korea has the world's best fried chicken. Let's go. The best fried chicken in America is Korean. Whoa. I have had Korean fried chicken. It is delicious. Something about the sauce is really, really good. Also, Jensen Wong in South Korea says, are there any NVIDIA investors here? The audience cheers and says, that's why Korea is rich, baby. Let's play this one. Play it loud. Did we get it? NVIDIA investors, that's why Korea is so rich. He's got bits. Do I have any messages? He's got bits. He's feeling good right now.

45:27I mean, on your$5 trillion day, I mean, let him have a little fun. What are you going to do? Have some fun. Let him have a little fun straight after GTC. It's so wild that they hit$5 trillion yesterday, particularly because they don't report earnings for another two weeks. So basically every other big tech company is on the October 28th, October 29th cycle. So as of tonight after Amazon and Apple report earnings, we should we have like the book closed mostly on all the mag 7 except for nvidia but nvidia is still in the news and so uh alphabet got crammed to the business section microsoft got crammed down to the business section meta made it to the front page of the journal because they sold off uh but nvidia got the full chart of their progress they got the the glazenator 3000 the a1 from the wall Street Journal because they just hit$5 trillion.

46:27And they have this interesting chart of how many days it took NVIDIA to create a trillion dollars in market cap. The first trillion took them 6 ,138 days. The second trillion, they did it in 180 days. The third trillion in 66 days. Then the fourth trillion was in 273 days. And then the fifth trillion in 78 days. So from 6 ,000 days, it took him 20 years to make that first trillion. They say that the first trillion is the hardest. They do say that. But then the next one comes in just 66 days or 78 days, which is remarkable. So he is on a tear. And congratulations to Jensen Wang and all the folks over at NVIDIA for powering this AI revolution and absolutely crushing.

47:14Let's pull up this video of the Snowflake CRO doing a little street interview. Excuse me, sir. How old were you when you became a millionaire. This guy. Yes, sir. This guy. I'm just forgetting your Instagram feed. How old were you when you became a millionaire? Oh, boy. Four years ago. Okay, pause. Yeah, wait. So who is this guy again? He's the CRO of Snowflake, which like Snowflake's been like$100 billion company for a while, right? Snowflake Market Cap. I thought it went out like years ago. I mean, maybe he wasn't liquid. It's$91 billion market cap company. And it's down too. Like it was grading higher.

47:58So this was the line that people were most confused about because I don't want to try to guess this guy's age too, you know, I won't pretend to know exactly, but I assume he's in his 40s, maybe his 50s. Yeah. He was at VMware for 10 years as the vice president of U.S. Enterprise Sales and then the senior vice president, general manager of Americas. Yeah. And he was also the president of Broadcom from November of 2023 till now. Yeah, it seems like an absolute dog. And became the CRO at Snowflake this year. But still, everyone, I think, is shocked of how he didn't possibly pull together a million dollars by somewhere.

48:5225 years in enterprise technology. That's what it says. 25 years. What is he spending his money on? He must be just obsessed with buying depreciating cars. That could be it. Maybe he gets a new Bentley Continental GT every year like clockwork. maybe he's the one that's buying all the lamborghini maybe he's buying multiple maybe he doesn't like to drive the same car twice yeah right maybe he gets a fresh car it must be it i don't like to drive cars with more than 20 miles on the odometer yeah so i'm just going to cycle it out i'll take the lot yeah yeah yeah every 20 miles they say once you drive it off the lot you lose half the value well it's a it's the price i'm willing to pay yeah because i got to keep my net worth under I got to keep it in six figs.

49:34To inspire. It's got to be six figs. It worked. It clearly inspired him to grind harder. He's a CRO at Snowflake. It's incredible. No, this must be some sort of he misspoke because that. Okay, let's play the rest of it. This does not make sense. What are you doing tech? CRO for Snowflake. Did you guys just go public? We went public five years ago. Congratulations. During COVID, couldn't ring the bell. So this is our moment. Do you recommend entrepreneurs take their company public? Because there's a risk there as well when you bring in all that capital, you bring in voting. What do you think about going public?

50:04Is it such a funny question? I have to snag him. I'm so sorry. I promise we got 17 million dollars. This is the best marketing he can do. I know. All right, just keep it rolling. Hey, we're keeping this shit rolling now. Sir, real quick. Could I finish real quick with you real quick for one minute? We'll do 30 seconds. 30 seconds is cool. All right. What do you think? He does not do 30 seconds. He definitely does another like five minutes based on the edit. Cash flow is everything until profitability comes. Then you get S &P 500 and you get a whole different kind of return. What does that mean too?

50:30Do I not understand that? What did you guys do in revenue this last year? So we're going to exit this year probably just over about four and a half billion. Let's go! Disclosure! That's the disclosure that he's not supposed to make. Keep track of us. It's amazing. Do you have some good equity in the company? We're doing okay. We're doing okay. As somebody who's an executive at a billion dollar company. You're awesome, by the way. I watch you all the time. Congratulations on all your success. This is amazing. The best financial advice you can give to the younger generation. Best money advice you can give to the younger generation.

50:55This guy shouldn't be giving. He's like, step one. You can save your way to being financially secure in the future. You've got to build up wealth. Don't worry about W2s. W2 is not necessary. W2 maybe will have to pay for that. It took me 30 years to pay for your college tuition. Next five years, what do you see in the future? Million-dollar savings. This is crazy. Five years is tough to predict, my friend. In this world, this is actually hilarious. It's up to 18 months. That's all you got. Yeah, okay. Give me the next 12 months. What's the number one thing people need to be adopting in their business right now?

51:22They're going to be using AI. He's still going. 30 seconds. 30 seconds. Accelerate revenue. It's all about moving fast. Do you think if business owners aren't leveraging AI, they're going to be left behind? Absolutely. Why? It's like the internet. If you're not on the internet, wave, you're dead. It's going to replace Uber drivers. Potentially good self-driving. It's going to change the world. It's so funny. Oh. Wild, wild interview. Yeah, so they had to file an emergency 8K disavowing unauthorized statements. The projection exceeded the official guidance by$100 million. so uh they they said that they were projecting 4.4 billion he rounded up to 4.5 and uh that's a no-no so he got himself in trouble stay safe out there i just don't understand how he has on his on his linkedin he has president at broadcom full-time since november of 2023 he's not listed on the Broadcom website.

52:17I don't know. But it is one of the funniest. Maybe he's kind of like so Humphareek style moonlighting over there. Yeah. Who knows? We have the founder of founder CEO of Whatnot coming in the studio today. Is that correct? In about 20 minutes. About 20 minutes. We have Grant. We're very excited. People have been debating Whatnot because it sort of came out of nowhere as this Dacacorn e-commerce company. Chris Pike, former guest on the show, is taking shots on the timeline, putting the timeline in turmoil, calling it unregulated gambling. Well, we're going to get to the bottom of it. Is it gambling?

52:56Is it not? He says he feels like someone needs to write a deep dive on this, but we will dig in. I'm not exactly sure, but we'll try and understand the business, how they make money, how it all works. Is buying groceries gambling? Sometimes you go buy a watermelon you're like i don't know that's it looks pretty good you can't crack it but if i once i cut it open once i get home and cut it open it might not be the best watermelon but it look good you're trying to knock on it you know you're trying to feel out kind of the vibes of the watermelon with pumpkins right now around that's right that's right halloween's coming out people are gambling on pumpkins it's gonna be rotten inside you gotta play the game yeah play the game on the field um we uh droove over on x uh released a video i guess he says amazon will sue me for leaking this.

53:42He says, worth it. This is so funny. Putting the tiny Amazon package in the missile launcher that just flies in and I love it. It just demolishes the door. It just completely destroys the door. The door just gets broken into a million pieces. And this is innovative because everyone's experiences, right? Even though Amazon gets it right to your front door, it'd still be kind of mildly annoying to have to go out and carry boxes in for a swirl problem. But this solves that. Yes, you need to replace the door once a day. Some people are taking shots at DoorDash with a new DoorDash robot. Will it get into your door?

54:25Well, here's your answer. Maybe it needs a missile launcher on it. I also love that this video, it looks like it's in part a real video. Those first frames with the human hand, that looks like they actually built the missile and then they switch over to just a 3D render. And so there's like maybe no AI here, but still some CGI. It's very, very hilarious. So funny. Should we move into some of the more Neo reactions? Of course, yeah. But first, let me tell you about Julius. What analysis do you want to run? Chat with your data, get expert level insights in seconds. The Julius.ai. Yes, me after loading the dishwasher instead of paying a robot$20 ,000 to do it for me, and it's Trump holding a ton of cash.

55:12Very, very funny. Also, people are planning to drip them out with gold chains or silver chains, and I think a fade on here too. This is very funny. I mean, underrated the idea that if you get the 1X robot, But I wonder how much you'll actually be able to dress it. Because at a certain point, you can't put a crazy body kit on a Tesla because you'll just block the cameras and eventually you won't have self-driving. And I wonder... I'm putting Chrome hearts all over my knee. Yeah? You're going to? I think it's going to trip and it's going to shatter more. No, no, no. One t-shirt? One t-shirt? I don't know.

55:52You get a baggy t-shirt in there. If it can't handle one baggy tee. Yeah. If it can't handle... It's just you're going to have to pay for two Chrome Hearts T-shirts because you have to send one to the teleoperator because the teleoperator doesn't feel the exact weight of the Chrome Hearts T-shirt. They need to immediately make an account dripped out Neos, and it's just a daily feed. The vibes around this company are just great. It's such a moonshot, and it's so capital intensive. It's so difficult. But I think everyone's having fun with it. MKBHG actually dropped a whole video. Rob, we've got to pull up this video.

56:25Rob Frude. He says 20K to simulate the experience of a roommate with a ketamine problem. That's ridiculous. No, this is what we were saying on the show yesterday. It's like, hey, Neo, clean up all those wine glasses. And it's just like. Yeah, just smashing everything. I mean, it's going to be crazy. Neo, get ready to be good at cleaning up glass. For sure. That's really the final challenge for humanoids, cleaning up glass. One of the greatest challenges for humans. Yeah. Cleaning up glass. Oh, yeah. It's a mess. It's an adventure. It's a mess. It's a high-stakes game. I mean, some of the anecdotes from this robot are going to be hilarious.

57:08There's going to be wild stuff. Well, if you're looking for a robot to generate some media for you, head over to Fall, the generative media platform for developers. The world's best generative image, video, and audio models all in one place. Develop and fine-tune models with serverless GPUs and on-demand clusters. another post here with neo with got the glock out now tell me where the seed phrase is gotta make sure that the just another reason why they put the experts locally you definitely have to arm them this is important yeah should humanoids be able to open carry there's an old corridor crew video about this that went super viral boss town dynamics they made a whole series out of it, all CGI, but they took renders of real Boston Dynamics robots and put them in like shooting ranges and this whole environment.

57:59They tell this whole story and it just goes viral constantly and people think it's real, even though it's CGI. What were you saying about MKBHD? Oh, MKBHD put out a video about Neo 1X and he's beating the drum of like, this is teleoperation, it doesn't count. And he's like, he's going pretty hard on them. He gives us some caveats and he says, you know, hey, I wish, I hope this does work out. But some other people were kind of levying the critique on him that he's not in the arena, not trying to build things. But I mean, at the same time, we talked about this yesterday, I think MQBHD really has seen a number of these consumer.

58:37He's been burned. Yeah. And he's been burned on a couple things. You know, multi-hundred dollar devices that should be a lot easier to deliver a great customer experience and so i think it's fair for him to be concerned yeah about a twenty thousand dollar product that would probably like i i i'm every day i'm more bullish on one x yeah i think i think uh they're they have a great understanding of the internet yeah they're folk they've been focused on like a specific use case the home for a while they're focused on teleoperation which is the right thing to be focused on. Because even if it only is good at teleoperation forever, that still could be valuable.

59:23I'm bullish as a futurist. Not as a consumer advocate. And I think that everybody should be going into this with the understanding that you are going to be participating in trying to pull the future forward. I agree. Yeah. It seems like, I don't know if they're calling it a beta but it feels like you should go into it uh assuming it's a beta what are you laughing at didn't you have a joke did you ever post your joke at least tell your joke on the on the stream if you didn't post i'll tell my joke tell your joke a woman shoots her 1x neo then holds him underwater for five minutes next she hangs him right after they enjoy a lovely dinner how can this be the what was uh what was john posting about that before the woman john john This was inspired by John Palmer's post last week of the AI agent being like, this is not my son.

1:00:13Oh, yeah, yeah, yeah. A father and son create an AI agent company that gets to$20 million. They go to San Francisco. Yeah, it was a father and son build an AI agent. The father tweets that they have$10 million ARR, and the son has rushed to San Francisco to pitch the seed round. At the pitch, the VC looks at the boy and says, I can't fund this boy. He is my son.

1:00:37And so the answer, of course, to my riddle, a woman shoots her 1X Neo, then holds him underwater for five minutes. Next, she hangs him right after they enjoy a lovely dinner. The answer to the riddle is, of course... Turbo Puffer. Search every bite. No. Serverless vector in full text. Search to build from first principles on object storage. Fast 10X cheaper and extremely scalable. Major cliffhanger. No, she killed the robot, but is hanging out with the operator. Oh, okay, okay. At first, my read on it... I mean, my read on it was just like, it's ridiculous and it's a play on that whole... riddle that confuses LLMs generally.

1:01:10I mean, there's some very funny examples of those. It's the strawberry thing all over again. It's just very easy to confuse these things. P.E. Cooper over on X went incredibly viral. 25 ,000 likes, on half a million views, saying, my Neo robot, after I learned the third world or piloting them is mad chill. It's not going to be a third world or remember. it's gonna be an american tele-operated yeah robot what is this will manitis post he posted this neo robot in this like uh in this image and i understand it's photoshopped in there but i don't understand why it went so viral people really like it what is this reference to it's some i mean if you scroll down brandon jacobi is you think it's this you think it's They don't know.

1:02:28but it's such a real image. porches, I suppose, but to each their own. Dev had an absolute banger. Wait till you find that MF 20K clanker in your kitchen acting too friendly. And I guess you followed him after this post and he says, damn, after so long. Oh, that's funny. I got caught following him, but I did like this post and I was like, I gotta follow this. There's a dev. I don't know. Wait, after so long. You're not that big of an account. I feel like I'm not that behind. I don't know. I'm glad I found your account, though, Dev. You're very entertaining. You guys found each other. It's very sweet.

1:03:08Oh, Jordy already following. Okay. It's me, you, David Holtz for the followers I know. And Jaya Seed. Very funny. And I like that Dev replied to the post, banger. Just giving yourself a pat on the back is fantastic. Okay. Eric Katana has a less not-so PC post in here. Is it gay to have sex with a female robot teleoperated by a man? Elon Musk comes in with a crying emoji. I don't know. We'll leave that to the philosophers, I suppose. We'll need a whole new branch of philosophy. Eric Atana replied and said, Elon, what did you mean by this? Which is, Optimus' new hand upgrade includes 22 degrees of freedom.

1:03:51What does it mean? Anyways. Did this Tane post get deleted? I don't know where this one went. I don't know. Let's move over to some Bub Talk. We have a post here from USV with a clip from the legend Fred Wilson. While we pull that up, let me tell you a profound get your brand mentioned in ChatGDT. Reach millions of consumers who are using AI to discover new products and brands. Let's play the clip from Fred Wilson. The thing about Bubbles is that there's just two things I think that people need to remember about bubbles. First of all, bubbles provide the financing for new revolutions. Like that book you described.

1:04:37I love revolutions. And they provide it at inflated valuations. So people can, I mean, look at what OpenAI and Anthropic and others are doing. They're raising money at astronomical valuations and then throwing all of that money building. We got to get Fred doing it. This reminds us of the telecom overbuild of the late 90s, whatever. But the net result of it will be there will be advances that would not have been possible without that investment mania. So that's one thing to remember about bubbles. The second thing to remember about bubbles is after the crash, there's a renewal. And so if you can just ride out the crash and then stick with it, things will be even better afterwards.

1:05:17So the thing about bubbles is that there's just two things i think basically just survive survive getting the winners riding the winners invest in the companies that are durable and generation multi-generational yeah skip the others skill issue skill at all just just don't make mistakes don't make mistakes uh use google ai studio the fastest way from prompt to production with gemini create an ai powered app faster than ever that's the second time I did that I read but sometimes we skip around Dean Ball says my sense is that selling Blackwell chips to China would quite possibly be the most unpopular tech policy move of the Trump admin especially on Capitol Hill it's plausible that the long term really even near term result will be much more compute regulation I like this guy Dean Ball Tyler do you know Ball?

1:06:16I think you've already made this joke but have you Have you gone down the ball rabbit hole? Have you read any of his work? I've read some of his longer posts. He's coming on the show tomorrow. Pretty spooky day to come on the show. Yeah, it is. That'll be an interesting interview. Yeah, I feel like the... Didn't Trump went to China just yesterday, right? Or where did he actually meet Xi Jinping? I don't remember. Trump's not going to China. No, no, no. But he met with Xi Jinping and immediately let up... South Korea. Oh, South Korea. Okay. Oh, Jensen's there too. That's fun. I have no idea anymore.

1:06:56I feel like if you were anti-selling GPUs to China on the basis of AI doom, you have to update towards it's actually kind of chill to sell them to China, right? Like debate this with me. If you're a doomer. One of the reasons people said, let's not sell GPUs to China, was if they get GPUs, they will get super intelligence. They will steamroll us. They will be all powerful. And super intelligence will be communist. It will be state capitalist. It will be not loving free speech. It will be all the things that are un-American and we don't like. Then the last couple of years have happened and played out.

1:07:42and it feels like the fast takeoff scenario feels less likely. And so that should update you to be a little bit less worried about it. No? Less worried about selling to China because it's not nuclear weapons. It's not the Manhattan Project. It's nuclear slop. It's slop. It's erotica. But it's also great knowledge retrieval, really cool image generation, really cool audio generation. It is definitely a great productivity tool, but it's not going to be, okay, they have it, they got the GPUs, they got a GPT-5 level model, and now America just loses. And China is just dominant, right? Yeah, I mean, I think, you know, the Leopold take is that it's still like, I don't think in his scenario, like, the labs haven't been nationalized yet.

1:08:44And he's like, I don't think the capabilities right now merit the nationalization. But I think a lot of people still think that, like, it's like, you know, two more years, then we'll be at a point where, like, it would make sense for there to be, like, major geopolitical conflict just over the AI question. So I don't think it's like the people who were super like bullish on that idea, I think are still, I don't think they've been fully like disproven yet. Yeah. I don't know. It seems like we'll have to talk to more people about like how that's shaping up because there's also the weird dynamic of like, even if we wanted to sell Blackwell to China, would they even buy it?

1:09:22Because they seem very interested in their own supply chain. I think they can do both. you know they can buy blackwells today and they can still i don't know i mean the more you buy the less incentive you have to build locally like that's just true like we buy solar panels here there are no good solar panel companies i don't know there aren't they're probably they're like america has been like yeah we're buying them so cheaply like how do you get a company off the ground in a market that's just getting destroyed we ran this example with with gopro like america had a company there and got steamrolled.

1:09:58And so it's like, you need to be over investing like crazy. And America just doesn't have the appetite for that. Yeah. And I'm just saying China has more of an appetite to have a, you know, entire industry, you know, backstopped and, you know, money losing for a sustained period of time. Yeah. Um, couple more posts before we bring in grant zero hash. We talked about this briefly yesterday. MasterCard is set to acquire crypto startup zero hash for nearly$2 billion. Imran Khan over at Alliance says zero hash powers most crypto to fiat on off ramps holding money transmitter licenses and MSB licenses across the US.

1:10:40Most on ramps you use today likely run through zero hash. This is a pretty big deal. So MasterCard getting into the game. In other news, Kimball Musk says, please vote your Tesla stock. There is no one remotely close to my brother and no CEO on earth that would accept this package. It is a massive win for you as a shareholder every way you look at it. Elon says, thanks, bro. These are the original technology brothers. Oh, true. What does he mean by no CEO on earth that would accept this package? I can give some more context. So Gavin Baker says, my firm, Atreides Management, will be voting in favor of Elon's performance-based compensation package.

1:11:21I would prefer that every company I invest in have a comparable plan. Sure. I believe shareholders should generally support thoughtfully structured performance-based CEO compensation packages because they incentivize CEOs to create transformational growth and value. The basic logic of these plans is the CEO gets 10 % of the incremental value creation if they 10X the stock with a 2X hurdle. So shareholders get a 900 % stock increase, and the CEO gets an incremental 10%. The first Tesla plan incented Elon to 10x the value of Tesla. He achieved all the milestones and may never be paid for this. The Axon plan incented Rick Smith to greater than 5x the value of Axon, and he did.

1:11:57And when milestones aren't met, such as Farfetch, then the CEOs don't receive the payouts. The new Tesla plan is arguably better than the first, as it also includes financial and technological milestones in addition to market cap-based incentives. And again, so Atreides is riding with the brothers. Yes, but what do you think Kimball means by saying no CEO on earth would accept this package? I understand that I completely agree with the first part. There is no one remotely close to Elon for Tesla. Like I'm totally in on that. I'm in on the idea of incentive packages. I like that as a structure.

1:12:37Targets are so aggressive that most CEOs don't think they would achieve them. Yeah. They'd be like, no, pay me. How about I do it? Yeah. How about I get this compound at 10 % a year? Yeah. How about I do this or that? Like if I go to the CEO of Target or something and say like, okay, you won't make a dime unless you 10X the stock, they're going to be like, what are you talking about? No way. Yeah. That makes sense. Thank you. Well, let me tell you about Linear. Linear is a purpose-built tool for planning and building products. Meet the System for Modern Software Development, Streamlight Issues, Projects, and Product Roadmaps.

1:13:12Let me start building. Let's go to Cheng Liu. Have you heard this Scott Forstall stans? There's a lot of Scott Forstall stans out there. He was the Senior Vice President, iPhone Software. 13 years ago, he was apparently fired from Apple after the disastrous launch of Apple Maps. What mapping app do you use on your phone? I am an Apple Maps guy. Me too. I use Apple Maps as well. I pretty much always have been because I used to be. I switched to Apple Maps, but only like a year ago. Yeah. Then I was on Google Maps. I was on Waze and Waze was incredible because you'd find all these like secret behind the scenes Waze.

1:13:57And then I think a lot of local neighborhoods basically said like, hey, we actually live on a nice, quiet street. We don't want Waze funneling 1 ,000 cars through our neighborhood during rush hour. So can you route around us? And it kind of put everyone back on the normal routes. I just always like that you can see the maps on your home screen without unlocking your phone. And that was enough for me. I agree. I always stuck with it, even though sometimes it would take you on a little bit of a goose chase. Yeah. But I also was never a fan of the goose chase. I'm like, I don't want to. You're anti-goose chase?

1:14:34Like in the Waze context. Oh, I love a Waze goose chase. I don't want to go on a street race through a bunch of 25 mile an hour. No, the best is when you're on Waze. No, I agree with you. But the best is when you're using Waze and it's like, hey, you're on the freeway. But why don't you get off and get back on the freeway seven times over the next six miles so you can save 45 seconds? It's like Waze. I don't want to do that. At one point, Waze. They really should have a slight, like an aggression slider. They basically do. So if you go into the Waze settings, there's one where it's like, okay, avoid toll roads.

1:15:09It's like, I understand that. It's like, I'm driving across the town and I don't want to pay five bucks. Like, I want to save money. Yeah, avoid toll roads. Then one was like, avoid freeways. And I don't know, I guess it's like, maybe if you get motion sick or something, you don't want to be on freeways. I met someone recently that doesn't drive. They didn't grow up in LA. They moved here and they just avoid all freeways. So they only take street law. I mean, the self-driving cars often avoid freeways because higher speed accidents are more risky. But so for a while, I don't know if this was actually in ways, but there was a moment where you could go in and check.

1:15:43I want to avoid toll roads. Two was I want to avoid freeways. And three was I want to avoid dirt roads, which was hilarious because I can just imagine driving through Los Angeles. And all of a sudden it's like, yeah, if you cut across this person's front lawn, you can save 30 seconds. And so I don't know. Hot take is that Scott Forstall, like Apple Maps was this kind of disastrous launch where it really didn't work as anywhere near Google Maps. But eventually they turned it around and Apple Maps became a really great product for the iPhone. And Cheng Liu says, as someone with a bit of insider info, sorry for baiting, I can tell you that the current trajectory for Apple would likely have been better off if he ended up CEO, look up his background and interests, I'll forever be a forestall simp.

1:16:34The guy can code and coded the right things. And so it's very, very interesting. I wonder what he thinks about iOS 17. Yeah. I mean, you can run this counterfactual and say, okay, if you put the software guy in charge, maybe the software is better. But what does that mean for the hardware, the supply chain, the trade war like would scott forestall have been able to get through the crazy trade war i mean tim cook could have been a goaded coo right hand to scott that's true that's true but maybe you wouldn't have maybe ios 17 wouldn't be such a dumpster fire i don't think it's that bad actually i i it i i get multiple times a day a reminder of how how how bad i think i think it will i think you'll get used to it over the next like six months i've gotten used to it but i still can appreciate how bad it is.

1:17:28But we have Mark Gurman's coming on the show next week. Yes. And I'm excited to talk with him about it. Well, we have our first in-person guest of the day. While he hops on the show, let me tell you about numeralhq.com. Sales tax on autopilot. Spend less than five minutes per month on sales tax compliance. Thank you so much for hopping on the show. Great to meet you. Likewise. Good to meet you. Thank you. uh can you fellow la technology brother yeah not many of us yeah many uh here i'm gonna have you uh why don't you introduce yourself uh and and kind of describe the shape of the business today and then i'd love to go back and kind of hear the whole the whole entrepreneurial journey sure uh so i'm grant lafontaine uh co-founder ceo of a platform called whatnot whatnot's a live stream shopping platform online marketplace the way to best way to think about it is we sort of enable anyone to create like a digital brick and mortar store.

1:18:26So as opposed to, you know, any seller can come on, they can then sell through live video, which means you sort of like this, you can stand up, you can be in your store, you can walk around your store, sell goods, sell anything. And so we have people on the platform selling comics to paintings, to plants, to women's fashion, everything in between. And it just ends up being an incredibly good way to build a business very similar to like a brick and mortar retail shop. It's so interesting that that digital commerce has just and doesn't look like, yes, there's items on the storefront that you can buy.

1:19:03But imagine if stores had no employees in them, and you couldn't like, in real time quickly ask, yeah, ask a human like, hey, like, you know, what size do you recommend here? Or do you have this? Do you have this in another, you know, all these different things that ultimately drive like higher purchase volume in a retail setting that we just generally don't have online, but do exist on whatnot. Yeah. What was the, uh, uh, take me back to the start of the company. What were the, what was the competitive landscape like at the time? What was the, uh, what was this something that was like working internationally and you were like, Oh, of course, of course that was the whole, that was the whole thing.

1:19:40It was like, everyone was like looking at China and you look at how some of China's retail giants got started. It was like selling fruit and vegetables. Yeah. Like in live, and it's something that Americans just don't understand at all. You know, buying like bananas through like basically a farmer who's just like, yeah, check out these bananas. I actually imagine you buying bananas through. I would, I would do that. In Malibu. But anyways, but yeah, walk us through kind of like the early days of discovering kind of the opportunity. Yes. So one of the funny things with us is, so the live shopping market in Asia is like monstrous.

1:20:12You know, like nearly a trillion dollars. And it's been big for a long time. So I think it took off in like a little after 2015 or so. And it was just like ramped to a crazy extent. And when did you get started? What year? So we started Whatnot in 2020. Okay. Okay. So there was like some, there were examples out there. Yeah. Like this can work. Although we had no idea. Yeah. We literally had no idea. So when we started Whatnot, it was me and my co-founder, Logan. Yeah. And we were working at some bigger tech companies. We'd both started businesses before. We're like, we like to build. Yeah. Like that's sort of the thing we enjoy to do.

1:20:42Yeah. And we're like, we don't like what we're doing now. Let's go build something new. Yeah. And we were both really into buying and reselling things online, collectibles. So at one point we were selling sneakers and that whole world. And my first business was when I was seven and I started selling Pokemon cards online. There we go. There we go. I love it. I was selling DJ equipment. He was selling skateboards. At the last YC batch, we were asking every founder that we interviewed how they started their first, what their first business was. And a lot was somewhat collectibles. It was like running Fortnite servers, collectibles.

1:21:15Sneakers, yeah. I don't know. I have kids. I don't know if you have kids. I want to get my kids to do business online ASAP. It seems so valuable. People that are like, I want to start a business, but I don't have an idea. Business is just making something for one price or buying something for one price and selling it for a higher price. That is just business. They're just sneakers or collectibles, Pokemon cards, whatever. It's just very simple. So I feel like a lot of the media platforms, they'll start with just like the idea of like, I want to share photos online, Instagram, or I want to have a place to host video online, YouTube.

1:21:51And then over time, the commerce aspect comes in and becomes more feature rich, whether it's tagged products with affiliate codes or sponsored content or an ad network. Like how mature was your thinking when you went to write like the first pitch deck or the first, did you go through YC? We went through YC. Yeah, so the YC application. Did you apply with this idea or did you guys like take a little bit of a. So I'd say we had like a quasi pivot. Sure. Which was, so early on the thesis was there's something around collectibles and social commerce. Because a lot of why you collect is not necessarily for the object, but to sort of hang out and talk to people about your interests.

1:22:33And so our initial pitch was basically a social commerce-like marketplace for Funko Pops, which are like little vinyl dolls. And that was it. That's what's in the YC application. And we didn't have any idea we were going to do video. It sort of looked like a standard marketplace with a follow button and a cool handful of features. So was it creator-led in the sense that I, as the collector, come and share pictures and inventory and details about particular products that I have in stock? Yeah, it would basically be that. So you'd have a profile. You could follow it. You could post your stuff. There's DMs.

1:23:10Got it. You could chat with people. Yeah. And how many investors, obviously, you managed to pull around together, but how many investors passed because they were like, sorry, I'm not backing your niche toy marketplace? marketplace. Yeah. I mean, it's, it's funny now because the amount of money we've raised is like ungodly large, um, you know, but not even, not even, not even relative to the valuation. I mean, you've raised like just under a billion dollars. You're sitting at 11 billion. It seems like also the stakes right now for, for raising an ungodly amount of money are like, uh, if you, if, if your business doesn't work, does everyone in the whole world lose their job?

1:23:48And as long as you're not there, like have fun. Or people are like, I have two GitHub stars on my projects. I'll take 2 billion. But yeah, so, so, so when did this, when did the video aspect come in? What were the, what were the thinking? And like, how did you, how did you think about the first design and then start iterating? Yeah. So I'd say like I'm a purist when it comes to building product, which is like, you, you know, stay close to your customers, you talk to your customers and you figure out the problems they have. And so as we were building our Funko Pop marketplace, we just knew everyone on the platform.

1:24:21Yeah, yeah. Like we knew all the sellers. We were friends with them. Sure. And we could see them hacking sort of social media video in interesting ways, in particular, trying to run like auctions in live video. Oh. So maybe going live on Twitch or other platforms and then selling on whatnot. Yeah. So what we'd say, well, the actual real initial example is we were following, I was following a seller on Instagram and I tapped into their Instagram Live and they were trying to hold auctions in the comments. And the latency on Instagram Live video, at least at that time, was like 30 plus seconds. So they basically have like put up the item, run a clock.

1:24:57And then once the clock stopped, they'd start taking the bids in the comments. And I invited my co-founder in and we got into a bidding war and it was just like really fun. And we're like, huh, why has no one built? It's sort of broken. Yeah, like why has no one built mobile video auctions just sort of felt obvious yeah and then people don't realize unless you've bid on an online auction you have no idea how thrilling it is so i've i've i've i've uh bought cars on on uh bring a trailer and once you start bidding it is just there's nothing more entertaining on the entire planet the more i get to know you the more you're like actually gamification really works on me.

1:25:38It does work on me. No, it's incredibly fun. And that's even in the context of Bring a Trailer, which anybody that's used Bring a Trailer, maybe it goes on the site all the time, but they're like, this is so underdeveloped in comparison to every other application that I use. And so even just the dynamic of bidding against someone else and looking at their bidding history and saying, okay, this guy just bids a lot. He doesn't actually buy cars okay this guy's bought like six cars in the last six months like i've got a real real competition here so um but i still yeah it's still the percentage of people that have actually like done like bought things online via auction i think is relatively it must be relatively low still yeah and so so yeah we basically how quickly did you ship the the your version of the live auction product after that four weeks okay yeah correct they were good apis around that time because I remember, uh, it wasn't Clubhouse doing live audio and they were partnered with some new company that was able to kind of do it like turnkey.

1:26:44Yeah. It was, uh, this company Agora, I think that's it. We actually didn't, again, we made a bunch of mistakes. We didn't know anything about China. We didn't know anything about the low latency live video APIs. So we actually stood up our own like low latency live video surfing infrastructure. If we hadn't, we probably would have built it in a week. Yeah. And it was not very good, but like good enough to get good enough. And the product was so clearly had product market fit. Yeah. I'm not entertaining as you guys can tell. And I did the first live stream because we, you know, we got to like make sure it works.

1:27:16And I sold$5 ,000 worth of Funko pops in two hours. And so you're like unentertaining human being selling, sells that much. We basically dropped everything the next day. And we're like, we don't know what this is, but this is really fun. This is it. Drop everything we're all in. Yeah. Okay. And$5 ,000 doesn't sound like a huge amount of money. But when you look at that in the context of an individual seller that probably has decent margins on that$5 ,000, and then they start realizing, I can do this every single day. And this quickly just becomes like a real business out of like, could be out of nothing.

1:27:49So from there, help me understand the growth mechanism. Is it people screenshotting and sharing to other platforms? Are you running ads? Is there a referral program? Is there SEO? Like what's the flywheel? Because it's - Well, and part of it, because I imagine like the people that are obsessively buying like Funko Pops are like pretty, like they know where they're being sold, right? Like these are people, like it's not like - Well, it's the same thing with GT3 RSs. Like people know, I'm going to go find that and bring a trailer. Yeah. Or their handful of other platforms. Yeah, it's like people that are obsessive about a category are going to know all the surface area where selling is happening and just go to where the inventory is.

1:28:30Yeah, is that roughly? Yeah, that's roughly correct. For almost the first three or four years of the business, we basically just focused on collectibles and enthusiast categories, which we sort of broaden out. Maybe like if you're into cars and things like that. And it was fundamentally because these categories, people like to talk about what they're buying, where they're buying it. And so we lean for the first four years, we basically just grew through referrals and word of mouth. And what was it like with the investors once the product was clearly working? Because I remember there were so many VC like think pieces on live shopping and how it had broken out in China.

1:29:10And I imagine when they met you, they were like, lisan al-gaii like i finally i finally but there was also there was also like uh in the early days like i remember there i won't uh there was like there was like pop shop was also very hot i think they raised around from like benchmark around the same time as you guys so it seemed like very competitive at the time also amazon owns twitch they have selling infrastructure like there were lots of reasons why someone else might do this the classic like what if google builds it what if twitch does it what if someone else does it yeah i'd say probably at least for the even the first couple of years after we did live we weren't like particularly hot yeah um a lot of people like looked at the side that we started in which was you know all collectibles and then sort of compared it to asia and they're like is that really the same so like sure you know a lot of it was people holding you know doing auctions and they're not showing their face on camera they just show like the Funko Pop.

1:30:06And a lot of the other companies... And that's like they'll put the product out and wait till it sells and then put a new one up there? Yeah, they'll like talk to you and you'll just like look at the product and you can... And meanwhile in China, if people haven't seen these videos, it'll be like, they'll show like, they'll just take a product, hold it up, move on, take a product, hold up, move on, right? Yeah, we have those people now. It's crazy that that works at all. Like that feels like such a worse experience than just an e-commerce app with like a bunch of dropdowns and filters. I don't know.

1:30:36Something, something about seeing a product in real life, like makes it, I think it inspires like some element of like trust even than just seeing like a digital, you know, photo or render. Yeah. I think what we see is that, um, people start to build relationships and trust the seller. So as opposed to like in the previous e-commerce world is where is it coming from? Who you don't know people come to whatnot. Cause they build a relationship with a seller. It's sort of like, I don't know, a local shop where you go and meet someone, talk to them. And so that just builds a lot more trust in the system.

1:31:06So in that sort of like, it seems like there were a few years where like you weren't as hot as like maybe you should have been. Was that driven by just not fully monetizing as well as you wanted or like choppy growth? Were you cyclical at all? Was it like the interest rate, like crisis and SUV collapsing that like put a damper on the hype or was it all pretty smooth yeah because i because because my my memory of of the whatnot hype cycle was like okay wow this company has like a massive valuation but i've never used it and i think that's what a lot of people like recognize like people are like something can't possibly be real unless i'm using it every day but you guys had this insane valuation and then every once in a while you're not you don't go on a lot of podcasts at least i I haven't heard you on one before, but you just pop up with a massive upround and then just pop up with another massive upround.

1:32:05And so I guess what we're trying to get at, I think, is has the growth been relatively consistent and smooth or has it been unlocking new categories, unlocking new use cases that have just really continued to drive the business? In general, it's been pretty smooth. We had one slow patch probably coming out of COVID where growth during COVID, if you were like a consumer entertainment company. No one had anything to do. And so I think we were sort of lazy in our growth motion because it just, basically we didn't have to worry about growth. It just sort of like, boom. It was happening to you. And then we had to get a little bit more deliberate, but I don't think we've ever grown under like, you know, top line sales.

1:32:44I mean, revenue has probably never grown under 90 % a year. Yeah. In the entire history of the business. That's amazing. And so it's been pretty good. And there's, when people, like there's moments and times where people get excited. So for a long time, we're like, oh, this is niche. It's collectibles. And then we got fashion going and working. And I'd be like, oh, that's exciting. That's a big category. And is fashion driven mostly by reselling existing inventory? Or are you partnered with specific brands or fast fashion companies? Like what's the shape of the market? Yeah, I mean, it's pretty diverse.

1:33:16So there's some brand selling now. There's brick and mortar jewelry shops selling. I would say the dominant sort of theme in fashion is probably like old season inventory or overstock. So people, the reason it's really great is it's great for whoever owns that inventory because they can collect higher margin by selling it on whatnot. And if you're a buyer. And that's the individual brand. A lot of times it's just businesses. Yeah, yeah. Think like TJ Maxx style stuff. Yeah, yeah, yeah. So people getting access to TJ Maxx style inventory and it's multi-brand. It'll be like Lululemon, but from last year.

1:33:55And so you're going to buy it. So it's like 60 % less or something. Yeah, 60 % less. And the deals are incredibly good. Yeah. Yeah. So people are kind of building their own business on whatnot from scratch, but then scaling to a point where they kind of have an operational relationship with different brands and they have a whole kind of strategy and probably pretty stable earnings as a creator. What's the competitive landscape been like over the last couple years? because I know it's been widely reported that TikTok's been willing to just burn billions of dollars trying to get shopping working on their platform.

1:34:29You have, there was another one, I can't even remember the name that popped up with crazy incentives that was like, I don't think it was live, but what's the one that? Flip. Flip. So Flip popped up and they were like, free money as a service. We're just going to get, and that was like, I remember people were like, people would send me a text And it's like, I just sent you$200 on Flip. And so I'm like, this can't, that's not going to last that long. And so you've had a number of people kind of like, you know, trying to at least like enter this like overall kind of like entertainment meets commerce market.

1:35:04But how have you got, it sounds like it's just been continuing to execute against your roadmap. Yeah, we were pretty, like, as you said, we, I sort of don't go on much podcasts and, you know, maybe we'll pop up every once in a while with a fundraise. We just try and stay really heads down and just deliver. Our thesis has always been it's going to be a competitive market. And so how you're going to distinguish yourself is consistent delivery and then, you know, continuing to add value to customers every day. And, you know, our belief has always been there's room for a purpose built platform here.

1:35:33Like, say, TikTok, amazing product, but people are going there to like watch videos, not necessarily shop. And I think that pattern is really hard. So we've always believed there's opportunity for a player who's purpose-built to be world-class at shopping and sort of, you know, shopping meets entertainment. And I think over the past six years, we've sort of largely proven that out. Is there a Mr. Beast of whatnot? um a true power law creator that's just a whole order of magnitude above everyone else in terms of seriousness and operational excellence size of team like that type of thing there there are a few people who are like you i think whatnot's very similar to other entertainment platforms so i spent like six years at youtube yep you know it is power law oriented And so the handful of people who are in the tippity-top tier, you're looking at operations with 100-plus employees at this point.

1:36:28So pretty big. That's pretty remarkable. Why do people say Whatnot is gambling? It feels like collecting things that just like, I mean, obviously there's a little bit of risk. You buy GTTRS and you find out that it is a broken axle. Yeah, and I think to be more specific, people are saying certain sellers on Whatnot are focused on. Mystery boxes? Hard mystery boxes, right? Like the power packs from GameStop. Yeah. I've seen a lot of this stuff. So you would have to also say that, yeah, power packs. Do you have policies around it? Do you have a personal stance? How does that look? Yeah, because gamification, too, is like that is something that sellers want.

1:37:06Totally. They're like, make it easier for me to provide an entertaining shopping experience. Yeah, yeah, yeah. So first off, there's no gambling allowed or whatnot, right? Like gambling is illegal. Flat out, like it's illegal and whatnot. Okay. You know, our trust and safety and customer experience team is probably our biggest team at the whole company. So we do take this stuff, like, very seriously. Let's hear it for no gambling. We love that. Let's hear it for trust and safety. Let's hear it for the trust and safety team. They work really hard. We know it's a serious expense. Very good. But, yeah, so none of that's loud, I think.

1:37:35You know, people open packs of cards. It's fun. I opened packs of cards when I was a kid. I sold Pokemon cards when I was a kid. And so we try and give sellers the flexibility to run their business in the right way. while making sure it's safe for customers and nothing illegal is allowed. Yeah. Stop. Yeah. And so there is an element of randomness if you're selling just, like I used to buy the booster packs of Magic the Gathering cards as a kid. And like sometimes you get something that's a foil or something. But that's very different than putting money into the whatnot platform, participating in some sort of ecosystem and then pulling out dollars directly.

1:38:09Like that's what gambling would be, right? Yeah. Yeah, that makes a lot of sense. Thank you for clearing that. what categories are you guys working towards that you're kind of most excited about? You said you unlock fashion. So the one, the future category I'm most excited about, it's still early innings is food and drink. Okay. That feels like really hard to sell online. Well, so it would be in a static world, but in a video based world where you can like, you know, so like the, we have a, there's a farm outside of San Diego who sells organic citrus. The bananas. Back to the bananas. Back to bananas.

1:38:45And so, like, I'll order from them. And, like, the next day at my doorstep will be, like, organic tangerines and things like that. And so I'm, like, I'm really into – I'm a foodie. And I think everyone increasingly is more conscious around what they eat, where they get their stuff from. And so I think being able to eventually, you know, get fresh meat from a farmer, get fresh fish from a lobster man from Maine. There's a bunch of stuff we're going to have to figure out logistics-wise and all sorts. but that's like food is crazy but but wouldn't you rather like see it get caught and then at your doorstep the next day yeah with starlink with starlink the fisherman's gonna be out there and be like all right i'm just pulling up a crate of crab who wants it who wants it that's wild and then he pulls it up and and there you go uh yeah that that's uh i can see that i can also there's so many other categories of like collecting i don't know if these are popular yet but eventually I could imagine like a wine wine collector is going on there and wine is something that you want if you're buying it like unless you're like a just an insane expert you probably want somebody like part of going to a wine shop is being able to talk to the you know associate there and they can talk through like the origins and the winery and what made what made you know a certain um case of wine like special things like that so wine makes sense I think other categories of collectible alcohol, like whiskeys and things like that.

1:40:11What about cars? Cars? Dreaming of the day to do cars. I've wanted to do cars for a long time. We'll eventually get to it. The truth is we're growing so fast, and so we're having to make sure the core user experience is really good before we can get outside of our cities. How rigid are you? Because Twitch just has just chatting or like IRL for like it's a catch-all If you're not playing Fortnite or Minecraft, they have a catch-all, and then people kind of go and innovate in that category. Is it just like I will get deplatformed if I try and sell my car and whatnot? Or is it just like you're not ready to say it's actually a good experience, but it is possible to do?

1:40:52Cars are tough, too, because you can buy a new car. It could be a lemon. You can buy a used car. and you know even after you know on bring a trailer there's like you buy it and then you can get ppi you win the auction then you're getting the the the ppi done and then at that point you can back out you don't get your like i guess like fee back but at least there's like some protections there so when you up the when you up the price like you know a thousand x it just gets so theoretically today if you adhere to our platform policies you'd be able to do it you So theoretically, it would be almost impossible because you'd have to be able to pay for it with your credit card.

1:41:34You'd have to be able to ship it in two days. It's like, I will be driving this across the country. Someone's like, challenge accepted. Cannonball across the country. We have sold some cars on whatnot. There's been more for collectible cars for specific events. So like the creator of Spawn sold like the Spawn mobile. And so we just did some custom stuff there. but in practice if we saw people starting to sell cars they'd likely butt up against some of the just general policies we have today and so we'd have to go and make the platform suitable for it. How's it like is the majority of the team in LA?

1:42:13So we're all over the place now so the true COVID company? Yeah we're true even before like we started a handful of months before COVID and our initial team there was our designer was in Italy two engineers in Brazil and me and Logan in LA. And so we've always sort of been distributed on the introvert. Did you go to YC up in the bay? We did go to YC up in the bay. And then we couldn't raise money. And so we moved to Phoenix. No way. Yeah. Because we had like no money and we just ran it, ran the company from our garage in Phoenix. And then I would, I would, uh, off air, I would love to see who passed.

1:42:48I mean, even to, to a lot of people's credit. I mean, I saw Sheil, uh, from BTV posted yesterday, Like if I got pitched whatnot and they were selling like dolls, I would have like almost certainly passed. Yeah, look, you would have the only way you would have ever invested in whatnot in the early days was if you knew me or Logan deeply and had worked with us. Because that's the only thing that would have given you confidence. We always had the confidence because we, you know, we thought we were capable of doing a lot per period of time. But we're not. I didn't know that many people. I hate networking and you're coming here with like a Funko pop thing.

1:43:27Like, so I totally, do I harbor like a little bit of like, you know, you're getting the last laugh, but, but, you know, like, I still think it's early, so I don't even worry about it. I just sort of like blinders on, we got to deliver. But yeah, unless you knew us, you would have sort of been a fool to invest. How did the Logan Paul investment come around then? I think we got introduced to him through Andreessen. So Andreessen did our series a and then helped, you know, get us connected. And he was really into Pokemon cards at that point in time. And so it was sort of like a natural fit. He's, you know, been on the platform a handful of times and, and, and so it was a good add on investor.

1:44:03Yeah. I was reviewing his post, uh, uh, talking about the investment. And I was like, that makes so much sense. Cause he's been so immersed with even early stage, like vine was so primordial in terms of like what TikTok became. He's been so early to all these platforms. It makes sense that he could see the vision better than a lot of other people. How international is the business today? I think we're in nine countries now. So I think we got into international maybe about two and a half years ago. It's a really big focus for us. The international business is doing very well. And I think you'll see us continue to expand.

1:44:36We're going on the offensive. Are we going to China? We're taking the fight for them. We're going to TikTok's front door. Bank dance, you're on notice. We'll go anywhere where we think we can add customer value. Yeah, that makes sense. uh how uh how crazy has the labubu uh side of the business been uh we we covered it what is so so we i think we literally by the time we did a deep dive on it i think we like may have like top ticked that that trend yeah i don't know maybe they're still growing as fast as they were but uh i mean it's pretty much came out of nowhere and then was doing hundreds of millions of dollars in sales on whatnot.

1:45:15It still is, I think. Wow. Annually? Yeah, annual. Like if you were to annualize. You're like now daily. No, it's pretty big. You know, I think it's just like fun. You know, you sort of get a box and you see which one you get and it's kind of quirky and weird. So it like lends itself to making content. And so it sort of takes off. What's the relationship like with the company behind Labubu in that case? Because I can imagine that they want to capture some of the value created by the gamification and they have the ability to sell directly and they might even have the ability to auction things off.

1:45:53They want efficient pricing. And so how does the relationship work between you and the originator of the collectible? I think it just depends. There are some collectibles companies we work with directly. I think we're working with Marvel. We've done a bunch of stuff in the toys and comic space. We haven't specifically worked with Lububu, honestly. I think Lububu's stuff took off to such a large degree. Even their, like, again, I don't know. They're still probably catching up. But I imagine, like, holy, you know. So no matter what, they were going to sell a gajillion Lububus. And so I think they were just trying to make sure manufacturing hits and the product is still doing so.

1:46:32Totally, totally. That makes sense. uh what uh before you jump off what what is the next like couple years look like i think it's a lot of the same stuff one just continue to make the buyer experience better the seller experience better you're going to see us get into more categories more countries and then make sure we build the products for more and more types of sellers so more investments in brands is definitely in the cards very specifically yeah where does ai play a role do you need uh like a traditional machine learning system for recommendation algorithms or is there a role for generative AI in this business or is it more just like you're using, uh, you know, tab completion models to write, speed up the developer productivity of the company?

1:47:15I think everyone's like, there's, there's the obvious spots where like, yes, everyone uses cursor and you know, whether you're an engineer or not, you can prototype and build stuff faster. Sure. Um, I think we're seeing obvious applications and CX and trust and safety really good at like detecting context in core ranking systems. I haven't seen anything yet that's like much better than the state of the art machine learning models. So I'd say we're at least in core ranking systems, probably not a lot of LLMs. We're starting to use it more for like content understanding. It's very good. You can take a live stream and like tell me everything about this live stream to then feed into the ML models.

1:47:56So I think for us, we're just trying to be world class at using LLMs to solve our customer problems. So we're making deep investments, but it's, you know, wherever we can solve a problem. Where does the community stand on AI generally? Are they like anti-AI because it needs to be like an old-fashioned collectible? Or do they see it as an opportunity? There will be new IP that comes from some kid prompting a model and then creates a collectible phenomenon? Yeah, I mean, I think people still want to interact with people at the core. So in some ways, what not sort of anti-AI and that it's like about the human experience.

1:48:34So I think anything that undermines the human experience will always have a little bit of a knee jerk, negative reaction. And so for us, again, I think customers don't care if you're in tech, they care if you're using it. But, you know, everyday customer who uses what not doesn't care. So it's more behind the scenes stuff where we can make their lives easier. So you're a seller. Can we tell you what are the five major things happening with your business? or generate tips for you or make it so you can list products really fast. Can we get all the context from them? I think that's sort of where we're applying it.

1:49:04Is there a cohort of folks that just, I mean, there are obviously tons of people that just watch Doug DeMuro videos because they're interested in cars and they never buy a car on cars and bids or they never even buy a new car because they're just happy with their car, but they like him and they like him talking about cars. And are there users on whatnot who don't really buy stuff, but watch a lot and just see it as like some sort of meditative content. Parasocial. Maybe parasocial, but also maybe just like, it's kind of like background TV and they're just like enjoying learning about different products and understanding the market.

1:49:41Yeah, absolutely. So on like any given day, the vast majority of people who are watching do not purchase anything. And the average person watches 80 minutes. 80 minutes. 80 minutes a day. 80 minutes a day. That's so much. Hit the gong.

1:50:00That's crazy. Well, thank you so much for coming on. Thanks for having me. Really appreciate it. Really good to meet you. Great to meet you. Always welcome when you have news. Anytime you want to pop your head up and announce a new. Yeah. All right. Thank you. Yeah, I'm glad that we're here. Stay hanging. Hang out. 80 minutes a day. Wow. That is a lot of minutes. I'm going to need more minutes in the day. Yeah. Sorry, Sam Altman. Good luck with your 30 minutes a day of ChatGPT use. Whatnot's got 80. Get those numbers up. Time to get those numbers up if you're building anything except Whatnot, apparently.

1:50:34Tyler, can you find out the day that we did our Popmart deep dive? Yes, I need to know that. I also need to tell you about Fin.ai, the number one AI agent for customer service, number one in performance benchmarks, number one in competitive bake-offs, number one ranking on G2. Oh, there's this funny, funny thing. So there's Profound and like GEO is actually in the timeline. And of course, we are partnered with Profound. And so, you know, there's the conflict of interest here. Didn't we, you know, you got to get through it with a kid gloves. But the founder of Amplitude, a YC company, Spencer Skates, he posted too many VCs are funding AI companies that should be features.

1:51:2612 plus startups have raised 72.8 million in the last few months for AI visibility, basically SEO for LLMs, what Profound does. We built it in three weeks with three people and shipped it for free. And he's going back and forth with David Burton, who says, I personally demoed 12 SEO for LLM products and 11 out of 12 sucked. Did you see this? First, don't we have to talk about James's reply? Oh yes, that was crazy. Almost ratio, Spencer James James replied uh lack of focus kills and he just shared uh amplitudes uh I I'm not a fan of that I think it's not golden retriever coded it's not it was pretty hardcore it's a little it's a little too hardcore for me I think uh you gotta you gotta put on the kid clubs when you're dealing with somebody but but to me to me one thing that stood out about Spencer's original post Yeah.

1:52:19Spencer's wife, by the way, did the Series A of whatnot. No way. Yeah. Wow. Incredible. Who's that? And skates. Oh, cool. Friend. Very cool. But they said 12 startups have raised$72 million in the last. That just doesn't feel like that much for a category that could be drive like billions. Yeah. Yeah, we've talked to a bunch of these different companies. is it feels like a lot for one sort of thing all at once. It is interesting because it feels like a ton compared to how many weapons companies were being funded when Andrel started. Two-thirds of that is profound. Two-thirds of the dollars, yeah, but not of the number of startups.

1:53:03And so it is just interesting that it feels like a niche. Yeah, but yeah, for all of venture, if something's a good opportunity, expect like at least 10 funded companies. I mean, I would expect that there are at least 12 new like CRM companies out there. Of course, we are partnered with Adio, Customer Relationship Magic, the AI native CRM that builds scales and grows your company to the next level. And so, but it was just funny because David Burton is like putting all of these AEO companies in the truth zone. but Ahmed Omar says which one works? and he said profound but I didn't buy because pricey and still too black box for me but he was very interested that he responded positively this is why one of the reasons I invested in profound myself was Kareem over at RAMP was just like this is what works for us that was enough of a cosign for me Tyler did you find out when we covered Popmart?

1:54:06Yes, it was August 29th. Oh, my God. Wait, did we perfectly top ticket? August 29th? Really? Not perfectly, but August, it actually peaked on August 26th. And so this is a good learning. If we ever start talking about toys. Something cool. It's not cool anymore. Collectible toys. Just know that we are probably top ticking it. If we're talking about some trend in technology, or probably like 12 months, 12 months at least before the top. For sure. We have our next guest in the Restream waiting room. First, how'd you sleep last night, Jordy? How'd you sleep on your 8sleep? I mean, my handicap is just killing me.

1:54:51Head over to 8sleep.com. I think I got a decent. Get a pod five. I did not do too well. I got six hours of sleep. I got a 79. I think you might have beat me today. See, the fact that you can get six hours and put up a 79 when I got seven hours and 10 minutes and put up an 81, It's like, I'm going to have Mateo reset this. I think it's a skill issue. You're just not focusing on sleep. So you're waking up in the middle of the night and stuff. It's not good. Should we bring in our next guest? Let's do it. I'm excited to talk to Chris. I'm going to let you take the intro. Chris, welcome to the show.

1:55:22Hey, guys. Thanks for having me. It's great to have you. I will give some quick background. Chris, you were the Senior Advisor to the U.S. Department of State, Office of the Special Envoy for Critical and Emerging Technology from 2023 to 2025. You're a Foreign Affairs Officer for the U.S. Department of State, the Bureau of Arms Control, Verification, and Compliance from 2015 to 2023. And in 2024, you were the Deputy Senior Director for Technology and National Security on the National Security Council at the White House. And there's a bunch of other stuff that seems equally significant and important, but we can get into it.

1:55:58It's great to have you on the show. I'm sure it's been a busy year for you. But why don't you give a quick introduction on your background and anything I may have missed? Thanks a lot, guys, and really appreciate you having me. Yeah, so I was a longtime civil servant in the government and the State Department. I was detailed to Biden White House from 2021 to 24, from 22 to 24. I was on the National Security Council, and I broadly had the China technology portfolio the whole time and oversaw all the Biden administration's critical and emerging technology-related work from the White House that I went back and served in the Trump administration for a good chunk of time this year and recently left government.

1:56:40So really appreciate you guys having me, especially on today, where obviously we had a bunch of big developments kind of overnight with the president meeting with Xi and the kind of trade deal, of which technology is obviously a very major portion. Okay, so before we get into what's happened in the last 24 hours in the last week, like, can you give kind of your view on this year and how a lot of these different, you know, how the trade war has been kind of evolving, in your view? Yeah, absolutely. So, I mean, I think the administration has very much gone back and forth on where they are in China policy.

1:57:19I think that you obviously have had enormous threats of tariffs and moves to kind of really kind of separate U.S. and Chinese economies on pretty fundamental levels. On the other hand, I think you haven't actually seen a lot of implementation there. There's a continued desire to talk with the Chinese. And when it comes to some of the technology protection measures where I think the Biden administration was more focused, it's been a little bit more stop and go. I think that generally since the first Trump administration, there has been a big pivot in Washington to move away from the policies of China of what's like deep integration in order to kind of get them to change and towards a more competitive stance.

1:58:05I think the first Trump administration really kind of initiated that. It was very much built on by the Biden administration, particularly in the semiconductors and AI spaces, where I think there was a very conscious decision. I know there's a very conscious decision to really separate those supply chains and saying in those areas we have to maintain as large of a lead as possible. It doesn't mean in every single one, but in the especially critical ones like advanced chips and AI. And then the Trump administration then I think took this and initially escalated with enormous tariff threats on China across the entire economy.

1:58:39The Chinese, I think, have very much taken to heart the lessons from the first Trump administration that what you have to do is punch back and punch back hard. And that has resulted, I think, in a little bit more of a detente, actually, than we've seen over the previous few years, where I think the United States has been able to kind of slowly ratchet things up. The Trump administration has ratcheted things up a lot. So did the Chinese. And I think both sides have now kind of pulled it down a little bit. It does feel like the renegotiation of U.S.-China relationships starts in 2016 with the new Trump administration.

1:59:17100-Year Marathon by Michael Pillsbury is an important point. That book, it seemed like it sort of went viral on Capitol Hill. There's also the rejection of American tech companies. But what else do you think kind of led us? what was like the inciting event, if there was one, in this beginning to discuss a renegotiation in U.S.-China policy? Is it Xi Jinping becoming effectively a dictator for life, that kind of changing the perception of like, okay, well, they're not even going, they're not even trying to go for democracy. What other kind of like moments in history do you think of as the defining moments in the renegotiation of U.S.-China relations.

2:00:02Yeah, I think that's right. I think that China has kind of fundamentally, well, China has fundamentally changed under Xi. I think there's some realization here that some of the promises of kind of political change in China and market-based change are not going to happen. There have been very much market-based changes in some senses, right? They actually are much more deeply integrated into the economy, but it hasn't change the fundamental, you know, integration with the Chinese state and their companies. And from that flows kind of national security considerations. And I think, you know, Xi has obviously taken this in a much more authoritarian direction.

2:00:40And one can argue whether or not that was there all along. And actually, you know, maybe it was something that people missed in the 2000s. But I think very much in the mid-2015s, like you said, around 2015, 2015, 2016, people were starting to realize that we have to change course, and particularly, I think, in 2017, 2018. And look, the initial Trump campaign, I think, really did change the narrative on China. I think that's pretty indisputable. On certain things, I think we were starting to see a little bit of a changing narrative towards the end of the Obama administration. The last week of the Obama administration, they put out a big report, the president's Council on, PCAST, President Council on Science Technology, put out a big report on the importance of maintaining a domestic semiconductor industry, which was the first time they'd really talked about that.

2:01:28And then obviously, you know, with the CHIPS Act in the coming years, that became the major focus. But I think you have to give the first Trump administration credit for really changing the policy. And the Biden administration very much continued that. I think the approach was maybe not to have a massive confrontational policy across the economy, but to say in the areas that we care the most about, we are going to take actions that four or six years ago would have been unprecedented to separate ourselves. And we are in a very different place because of that. I think we'll see kind of where we go going forward.

2:02:08It's a little bit of a... The last couple of weeks have introduced some more complicating dynamics, I think. Yeah, one of those was the rare earth elements debate. Has America been blindsided by the importance there? Were we miscalibrated on exactly how much leverage they had? Like how much that poker chip is worth on the poker table of this trade deal? Yeah, because there was a point where people are like, wow, they're already using that play, right? Yeah, taking away the passports of the researchers and then also just banning exports. it's really crippling our ability to build with those rare earth elements.

2:02:52But is that something that we've known about for the last 10 years? Or has this been a surprise? Or is that just like one of the many chips that we expected to be played the whole time? Yeah, I mean, people in Washington clearly knew that the Chinese had dominance of rare earths. You'd have to kind of be under a rock to not know it. But I think you're totally right that no one really did anything about it. and part of that is that look it's so hard to get things done in Washington it's hard to get things done and it's hard to get things done that you need a lot of investment and a lot of coordination and massive resourcing and coordination is exactly the thing that's really difficult to get things done in Washington unless there's a crisis to really make a move yeah he did do that SPAC that's crazy Yeah, I mean, I think that's the biggest success out of all of them.

2:03:47Totally. Yeah, and I think, look, Washington actually can be good at responding very quickly in a crisis, right? Like Operation Warp Speed was very effective, for instance. But obviously it took until there was a pandemic for us to do something like that. So I'm hopeful that now people will take this very seriously. I wouldn't be surprised if we start to see legislation on Capitol Hill that's actually resource this. Because even though the administration, I think the administration is focused on it now and takes it very seriously, they're also, you know, we really need to get money, right? And exactly how that money manifests, whether it's, you know, equity investments or loans or some kind of, you know, other, you know, Defense Production Act authorities or something else.

2:04:34Like there's going to have to be some government dollars here, which certainly can be stretched a long way. but we're going to have to do that. I think just to your original question, though, the thing that's surprising, I think people were very surprised that they played this card as expansively as they played it right now. Yeah. Right? And go ahead. Yeah, no, continue, and then I want to get into kind of the last 24 hours. Yeah, the surprise on playing the card early. Yeah. Look, the fact is, this move was way disproportionate to anything that we had done. Effectively, it was saying, There was this affiliates rule that the United States put in place, which basically fixes a big loophole in export controls.

2:05:15The way that export controls work is every single company has to actually be listed on the control list to be controlled. And if the subsidiary isn't listed, then you can export to them. That's not the way that sanctions work. The way that sanctions work is if the company is listed, then every company they own is sanctioned. And you better not work with sanctioned companies because otherwise you're violating sanctions. for various arcane, you know, reasons that have to do with nothing. It seems like a really simple own goal there. Exactly, exactly. So we basically put in place, there was a regulation that said, hey, we're going to fix this really obvious loophole that is being blatantly exploited that also is just very logical.

2:05:52And the Chinese response to that was to go way over the top and say we're going to cut you off from all the rare earths and, you know, threaten your entire economy. they didn't think that was proportionate i think what they were trying to do was send a signal that we won't tolerate any additional export controls on you know any entities in china and we want any any you know additional restrictions on our access to technology we're going to go over the top it was really risky for the chinese because the u.s does have big leverage over the chinese too right there's two things that they're extremely dependent on us for on for us that you know we could cut them off from.

2:06:29The first is the US dollar, right? Their financial system is still very reliant on the dollar. Obviously, they're trying to extricate that. That would have enormous implications, but they're reliant. And the second is semiconductors. What would that look like? Sanctions on banks or something like that. I think that's something that Chinese fear quite a bit. Now, to be clear, I'm not advocating for that, and that would be a really significant move, but just it's something that would be a lever that we have. The second is semiconductors, right uh they they're not good at you know they have limited seven nanometer can't do anything beyond that and they're very all their whole industry is very reliant on chips made with us tooling and so on okay so i want to get way way more into chips but to kind of ground the next part of the conversation like catch us up to speed on the on what what you kind of what your takeaways were and what we know out of the meeting in south korea yeah so i think basically the deal was sort of a freeze for freeze.

2:07:24They agreed to freeze implementation of their massive escalation, our earth rule. We agreed to freeze our affiliates rule. So I think you think of it as largely it locks in the status quo for the duration of the deal. There was various other things that it resolves that are recently created issues. The Chinese agreed to buy soybeans, and we reserve resolve rarest access etc it did it did not you know we avoided some of the worst case outcomes uh there was no agreement to send you know our votes to dance chips to china uh there was no we didn't accept like a trojan horse package to like hey we're going to build a bunch of you know cattle and byd factories in the united states which would have been very bad for us battery and and um you know auto industry in the longer term uh but i think china really does feel emboldened right now because they kind of punched us in the face with the rarest rule and And what happened was we both de-escalated a little bit.

2:08:17We'll see what happens going forward. But the status quo, to a degree, benefits China. They're trying to buy time. They're trying to circumvent our controls. And if we're not willing to continue to engage in this technology protection effort, then every day the Chinese get a little better at working around our controls. So what what can you say about or what do you what's your read on the actual dynamic between Xi and Trump? Like, I feel like these two guys have been in the game for a long time. I have to imagine my part of me wants to believe that they enjoy this like, you know, deadly dance that they play with each other.

2:08:58But maybe they're just like deeply frustrated. Putin and Xi Jinping kind of being like, yeah, we might live forever. Yeah. You think you're going to make it to 160, bro? He's like, yeah, totally. I got some scientists that are cooking up crazy stuff. I love that this is like hot mic. Who knows if it's real? It might be a psyop, but I like to imagine that they're just, you know, throwing down. Who knows? Yeah, what do you think? I will say that was not the first time I've heard leaders of countries speak to the desire to live to 150. Here's some interesting things. You work at the White House, and I have heard that before.

2:09:32Not from them, but. Yeah, yeah, yeah. I'm thinking of what you're putting down. That's great. Yeah, but no, I mean, look, I think that that Trump's obviously coming at this from from, you know, this deal making perspective. That's not surprising. And obviously there is, you know, political dynamics here on top, both because I think there's threats to, you know, the Chinese the Chinese kind of both pose threats to hollowing out industry, but also are kind of doing acute things to cause pain to some of Trump's political base. Like I think that the soybean purchases and things like that are doing that.

2:10:08Yeah, exactly. On the other side, I think that, you know, Xi is very much seeing himself as managing the relationship for the long term. I mean, what hovers obviously over all of this is Taiwan as well. And the desire to kind of slowly move the United States away from Taiwan, which is ultimately his like long term historical perspective and goal. So in some ways, they're coming at this from a little bit opposite. I think Xi is from the much, much longer term angle and Trump from a shorter term angle. But I don't think they necessarily have hugely different perceptions of the broader dynamics at play here.

2:10:49OK, so on the on the chip front. OK, yes. So we got the rollback. Yeah, let's move on to chip. So I don't even know, maybe it was two, three, four weeks ago at this point, but somewhere in the last month or so, they were basically like, we don't want any of your chips at all. And I didn't really know how much to read into that at the time. It seemed like the DC team wants NVIDIA chips, but the CCP said, you've got to justify why you'd want NVIDIA chips over Huawei chips. We're working hard. Let's support the Chinese industry. It wants to build up, you know, real national champions that can get to the cutting edge or leading edge.

2:11:32And at the same time, the AI industry wants to be able to make as much progress as they possibly can. And, you know, what was your read when that news broke? And then how much do you think they're still kind of gunning for Blackwell grade chips? yeah so a number one i think they are you know the chinese will take all of the u.s chips that they can get in a vacuum um like i i i think that there's some negotiating ploys here uh and i think i think they were holding the h20s kind of hostage because they wanted to to turn the tables on us and get us to give them blackwells which which obviously didn't happen there is a possibility that there's the information filtering up to Xi is saying, actually, we can do this ourselves.

2:12:24You shouldn't take these. And that would be like Huawei being like, no, give us, just keep the money coming back to us so that we're close and we'll be a lot closer in a few years if we can take those dollars, reinvest them, et cetera. That's right. But the problem is there's just no way that China can meet their domestic demand with Chinese chips alone. I mean, if we can have all sorts of debates about Chinese production numbers, and I'm happy to engage in those, but fundamentally, the United States cannot meet demand for our own AI ecosystem with all the semiconductor production in the world and all the best companies.

2:13:02So it's obviously impossible that China, who has a much more nascent ecosystem, is heavily restricted on tooling. And I think all the estimates of their seven nanometer production capacity, not to mention, obviously, they don't have any 5, 4, 3, 2 production capacity, is somewhere in the realm of 1 25th to 1 100th of the United States. There's just no way that they can fill their domestic demand with Chinese chips. So any effort to say we're going to not take any U.S. chips is going to inherently hobble their AI industry. I mean, I guess at some point, if you're very close, the marginal value of that would be such that you're willing to take it.

2:13:39But I think, you know, just given the dynamics we're broadly seeing with USAI companies, I'd be very surprised if they were actually, when the cards were down, were going to reject NVIDIA chips. Is Jensen being in South Korea at the same time just a coincidence? Or did he, do you think, you know, he sort of planned around that? Do you think anything with Jensen is a coincidence these days? No, I don't think so. Yeah, yeah, yeah. No, it makes sense. I mean, he's basically... He's the CEO of the biggest company in the world. He should be there regardless. Yeah, yeah. He's running... NVIDIA is basically a country.

2:14:19I mean, their market cap is bigger than... Their market cap is equivalent to twice Canada's GDP. Something like that. It's greater than Germany's GDP. Remarkable. Not that you can really compare those. It's kind of Apple stories. You wrote a piece that you published last night around why Chinese car investments are a national security risk. Would love for you to kind of like give us a preview on that front. Yeah, I mean, the point we were making, there was a lot of rumors in advance of the summit that another thing that the Chinese were dangling, which did not come to fruition, was saying, hey, appealing to the dealmaker element of Trump and saying, hey, we'll give you a bunch of investment in the U.S.

2:15:01in exchange for whatever. I mean, as part of the deal. But that kind of very obviously is a Trojan horse. We very much restrict Chinese investment in critical industries in the U.S. via the CFIUS process right now. And obviously, particularly given the rare earth controls, you would think that that would, I think it has shown that we really can't be dependent on China for critical supply chains. So if we were to have, you know, BYD factories, like I was saying before, like BYD factories or cattle factories in the U.S., that's just as bad as relying on Chinese graphite for batteries. So when you say Chinese investment in the U.S., you mean them setting up local factories, which is different from making equity.

2:15:47Yeah, yeah, because China wants us to make as many equity investments in their companies as possible because that gives, I mean, think about DJI is a good example there's american venture capitalists that were in dji they have uh huge returns just like sitting there they now have an incentive to say hey we shouldn't we shouldn't ban dji like it's not a problem that china can fly you know millions of drones over over all of our cities uh just just worry about something else we saw this with i mean we saw this with bite we saw this with bite dance too right yeah yeah no that's right and there's a whole separate conversation on kind of outbound investment and what we should think about with U.S.

2:16:28investors investing in China. And then with inbound investment, the question is, what are the risks that the Chinese, and not passive investment, that's very different, but active investments and particularly setting up large facilities in key supply chains, what are the risks here? And the trend we've seen over the course of multiple administrations has been to take those threats much more seriously. um so it's it's you know it's i think that's a positive thing i think that the other thing the administration is going to have to really think about is uh you know first of all what are the investments we need to make in those supply chains if we're not going to accept that and we're going to diversify not just from like chinese batteries but also like all the way down the chinese battery supply chain um what does that mean but also we have to think about kind of what are the sectors where and this is where we got to the car uh was what happened with cars and we talked about in the piece, what are the sectors where just having large-scale deployment of kind of connected Chinese products in that area poses a risk?

2:17:25And I think there was a recognition in the Biden administration that actually, you know, electric vehicles in general, having, you know, 20%, 30 % of the cars on the road were Chinese and could theoretically be like, you know, switched off by Beijing. Yeah, I mean, they could cripple our transportation infrastructure overnight, right? That is something that was unacceptable. So we've effectively banned Chinese cars, but did it before there was big market penetration. So I think we have to kind of think about what are the other, like, I think there needs to be some pretty hard conversations on robotics, right?

2:17:56Yeah. So what do you think about, uh, you need to read just selling humanoids on walmart.com. We were thinking about buying one here because, because I've, I, I, I've been, uh, very vocal over the last year, as long as we've been doing the show that, you know, I think it's super disappointing that we let, you know, like GoPro should have been, the dominant American drone company. And there's a longer list of companies that could have made great products in that space. Maybe people would have to pay slightly more, but we just let DJI flood. And from what we know, DJI was just willing to lose money for a very, very, very, very long period of time in order to just take market share.

2:18:36And so when I think about humanoids and, you know, I don't, I'm not going to put a a unitree robot in my home. Uh, but if they're the cheapest and, you know, if they're kind of the leading edge right now, there's a scenario where we just let them flood the United States with humanoid robots and maybe they're adding value and home or in factories and different settings. But then it's just obviously like you just watch any sci-fi movie and you can understand like the potential, like national security risk of allowing robots that can be tele-operated to, you know, just proliferate all over the country.

2:19:13And so it just feels like hopefully that comes up at some point in some of these discussions of just, hey, maybe we make the call to not allow this to happen like it did with drones. Yeah. I think we just have to have kind of class-based regulations on these things. It says anything that fits these parameters, you know, a robot in your home that's capable, that has It can pick up a knife. It has equipment and audio, yeah, and has mapping technology and maybe has some kind of appendage. I don't know. You could figure out a way to do it. It just cannot be sold in the United States if it's from a Chinese company or Chinese, Russian, North Korean, right, like the broader set of actors.

2:19:54The only category of items that we've done that for as a class is vehicles. But I think that needs to be – we need to think about it. we have the tools to do it and we need to think about what are the areas where that pose kind of similar risks. And you're worried that we'll wait until there's a robot uprising to do, to do something like this. I mean, you said earlier, like the U S is really good at reacting in a, in a crisis. And I just hope I like at some point, like if I'm China and you're letting me sell humanoid robots into your country at scale and you're a political adversary, like I'm laughing, I'm laughing at you.

2:20:34Like, it just seems like totally unserious. Yeah, I think it is easier to do these types of things before there are huge market impacts. So part of the reason that the cars move was doable was because, first of all, like industry was very supportive, right? We had Detroit banging down the door being like, this is we're going to get killed by the Chinese. But in the same way they are with tariffs. But second of all, the market penetration of all the Chinese car companies was very close to zero still, partly because we had such significant tariffs, although those were going to become muted in effect.

2:21:09So if you have an area where it's still pretty nascent, it's doable. And humanoids are perfect. Like the second, once you have a million Americans that have spent 20 grand on a humanoid and they're like, I'm not like, I want to buy the next version. I like this one. You know, it's not like we've now seen the Chinese can build hit consumer products. There was a long time where people didn't believe that was true. But clearly with TikTok and DJI, like it's no longer a debate. And DJI is a great example because, you know, DJI, it's a much more complicated to do this for UAVs right now because DJI is everywhere.

2:21:44Also, all sorts of like state and local and municipal governments, even the federal government still uses DJI drones. They have such dominance that even if we did do it, we would have to do some kind of mitigation steps even to implement it if we did it today, which maybe we should. But it's massively complicated, whereas if you do it nascently, you could just kind of cut off the appendage. Yeah, anybody that disagrees about the national security risk here and the potential downsides, like you can't believe that, like if you believe that Israel was able to run like a decade long operation to get pagers like sold into the right group of people that they could then carry out this, you know, coordinated action.

2:22:28and then you think it's safe to just let robots from an adversarial nation flood our country, like, I don't know what to do. I think you're just, you know, blissfully. And they won't let us develop the domestic robot industry unless we do something like that too, right? Like, they will consistently undercut us and we'll never, our companies will never get mature enough where they'll be able to benefit from the economies of scale where they can even remotely compete with the Chinese. Instead, they're going to be 100 % off. We were having this debate before you, not debate really, but before you got on, it's like, I believe that China can simultaneously buy as many leading edge chips, Blackwells, whatever we'll give them, and simultaneously provide the long -term capital to allow their own industry to catch up on whatever timescale it needs to catch up on.

2:23:18What is the tolerance level? Basically, is it possible for us to run the Chinese playbook against them? We subsidize Blackwells, flood their markets with Blackwells, and then they can't go down the learning curve with SMIC and SME and Huawei. And then at the last second when they're super dependent, we pull it back when we need to right before, I guess, ASI hits or something like that. That's the basic thesis debate. I mean, I think two things with that. One is that history would show that's never something that the Chinese are going to accept, right? They've never been okay being reliant on any Western technology of any significance.

2:23:57They also, you know, we've been very clear about our intention to keep this dominant now. I think you can't put that toothpaste back in the tube. Like, they're going to go. We saw they were going all out in indigenization, you know, back in 2020. It's even more the case now. Their constraints are not desire or cash. they are actual access to equipment and tools and we're lucky that the SME and not just lithography machines but also etching deposition machines like anything all the kind of really advanced equipment is among the most complicated technology in the world But we can assume that they're actively that they have thousands or tens of thousands of people that are tasked globally with just collecting as much information as they possibly can to reproduce it locally Yeah, the only reason this works with chips is because this is the most complicated equipment with the most complicated supply chains on earth, right?

2:24:48Like an UV lithography machine is literally the most complicated machine ever made by man. So therefore it's possible the Chinese can actually indigenize every single thing on earth except this. Like that is a logically consistent outcome. The other thing I would just say with respect to your thesis is if that was the play, the more effective way to do that would be to cut them off from everything completely but then tell them you can have it via U.S. cloud because then you can cut it off at any moment's notice. You could debate if that's a good idea or not, but like chips are going to be a lagging effect, right?

2:25:23It's going to be two to three years until all the GPUs kind of go bust or - Or get depreciated or you need to move on in your architecture. It's interesting. What is the most current debate on the ground in Washington around like Blackwell specifically? So there is the question of whether or not we should ship them. But actually, the more pertinent question is, is there's a piece of legislation called the Gain AI Act that's moving on the Hill that actually the Senate passed as part of the NDAA, that must pass annual defense bill and is being debated with the House right now. And it's sponsored by Senator Banks.

2:26:02And basically, it says I was a Republican from from Indiana. And basically it says, look, we're not going to weigh in on the question of whether or not we should ship chips to China. That's a separate question. But regardless of where you are on that, American companies should get priority access before Chinese companies. And it basically requires, before shipping any chips to China that are controlled and authorized, it requires American buyers to get right of first refusal. Which I think is a, from my perspective, is a pretty logical approach. which whether or not you support or oppose AI chip exports to China, saying that America should have them first is logical.

2:26:44It also does one other thing, the most recent text, which exempts trusted U.S. exports to trusted U.S. persons to anywhere other than China from license requirements. The reason this is important is actually there's a lot of pending deals in the Middle East right now from U.S. data centers that are trying to make billions of dollars of deals in the Middle East. Those are subject to license requirements. there's a lot of negotiations on whether or not those can go, and they've been held up for a long time. But this would say as long as the government certifies them as a trusted U.S. person and they're subject to a couple standard restrictions, like they keep 50 % of their computing power in the U.S., then they can go without U.S.

2:27:22government approval. So really basically trying to prioritize U.S. customers in any decisions vis-a-vis Chinese companies and also U.S. companies vis-a-vis any decisions on the international AI build-out. Do you think Jensen will be happy with that? I think NVIDIA has been pretty public of their opposition to the bill. But look, my response would be— And is that just they want—they don't—they just want the flexibility to do deals effectively? Because it's not exactly a secret that NVIDIA, like, they have a Shanghai office and R &D center. Like they're trying to, I've joked that Jensen is, you know, the juggler, just sort of like juggling a bunch of different, you know, national interests always.

2:28:13And I guess that's his duty to the shareholders. But what is their specific, do you know what their specific opposition is? I think, I mean, any regulation is generally opposed. And I think they're concerned it would kind of hamstring their operations or result in kind of bureaucratic complexities that actually would prohibit them from selling anything to China, even if it kind of was prohibited. I guess I think that this is a good faith effort to work to develop a system that would allow that if that's the administration's policy position to do so. I think my response would generally be like if – the question is really is there a supply shortage of GPUs or not, right?

2:28:55Because if there is, I think it's pretty clear with Blackwell that there's more demand than there is supply. Then I think that U.S. companies probably it is not surprising to say that the U.S. government for the U.S. government to say that U.S. companies should get priority over these over foreign companies for these chips. If there's not, then the bill, in theory, would have no impact on their on their ability to sell them to others. Now, certainly, I think there's there's tweaks to the language and things that people would be. I know that Senator Banks and his staff are happy to work with industry on, but that's the intent here.

2:29:30I'd hope that we could come to a resolution there. Makes sense. Really enjoyed this conversation. Come on again soon. We'll reach out. I'm sure there'll be much more to talk about in the near future. This is fantastic. Thank you so much, Chris. Thanks, guys. Appreciate you having me. Look forward to talking more in the future. Cheers. See you later. Bye. Let me tell you about public.com, investing for those that take it seriously. They got multi-asset investing. They got industry-leading yields. They are trusted by millions, folks. And our next guest is already in the restring waiting room. Let's bring in Max Junestrand from Legora into the TBP at Ultradome.

2:30:06Boom. Max, how are you doing? Welcome. Welcome to the show. I'm great. Good to see you. How are you? Congratulations on all the progress. Introduce yourself, the company, and then give us the news. Of course. I'm Max Junestrand, CEO, co-founder of Legora. We're an AI-powered workspace for lawyers. We just closed our Series C, founded the business back in spring 2023. How much did you raise? John's ready. John's ready. $150 million on a one-day basis. You can't even hear the valuation. I got all of these pictures back from the gym this morning from Stockholm. They had bought a huge inflatable unicorn.

2:30:46Maybe we can send you some pictures. There we go. It literally takes up half the office. Amazing. If you added it to the seating calendar, so you can book it. Are you a YC company? Yeah, we were in YC winter 2024. Winter 2024? Wow. Yeah, winter 2024. That's insane. So it's 18 months. This must be one of the fastest growing YC companies of all time. I think so. That's actually crazy. 18 months in C to Series C. Wow. That's impressive. So the VCs, they'd said, we need 10 % max. and you're like best I can do is 8.3. I'll tell you actually, this one was easier to negotiate than the first one I did with Benchmark.

2:31:33When I did the one with Benchmark, we sat down, we took an Excel, Chathan put in some numbers and I went, there's no way dude. And he goes, I've never done a deal where I didn't get 20%. And then we started going down on the like point decimals. And I think we landed on something like 18.51%. It was extremely... You got to back off just a little tiny bit. It was good. He broke his rule for you. Okay, so how are you... We've talked to a bunch of... Yeah, let me give some context here because a good friend of mine is a lawyer and the firm he's at, he's on their tech kind of advisory, I don't know, whatever, big law, and he's been evaluating products and about a year ago he like was not impressed at all and when i saw him uh i saw him last week and he was telling me that uh he now is seeing ai tools do um he called it like a hundred thousand dollars worth of like diligence in like 10 minutes and he was like huge wake-up call he was like thinking about um like he didn't he didn't even think his firm would be able to evolve to that even though he even even though and this is a big you know big firm but he uh he was like clients are going to realize like what's the game has changed like very quickly and not every law firm is going to be is going to react uh quickly enough and so um yeah kind of want to get yeah wanted to get the update on like just how much has changed kind of over the last year because there's product innovation that's happening and then there's just like the models are getting better and they're both feeding into each other yeah well i i kind of view it like this here's the model capabilities here's the product capabilities and here's the people capabilities in terms of using those products and getting the most out of it so there's a huge sort of gap still to be filled in terms of what people are using the products for i mean back in 2024 when we were part of YC, I mean, Legora was a glorified GPT wrapper with sort of good rag, good citations, you know, solved compliancy and data issues.

2:33:50But now I think your friend is right. The products have moved from being a conversational, simple AI into something that solves much more sophisticated tasks. And it looks very differently across transactional work or in litigation or in in-house work. If you take due diligence as an example, I mean, it's effectively the exercise of looking through a large number of documents and understanding what is within those documents. And I think LLMs were better at reading than writing initially, which is quite an important distinction. So we came at this whole problem very much from the review angle. I think some companies, I mean, look, we were not the only ones who realized that legal and LLM was going to be a nice marriage they came at it from the thing like research angle or drafting angle which was much much harder to build a good company on in the early days yeah and i think now what's happening is i mean i don't know how many apps you use on your phone i use maybe five and what's happening is um the lawyers don't want to learn 20 different point solutions we're bringing everything into one suite and then with things like mcp we can actually also allow the really sophisticated law firm law firms to bring in a corpus of their own bespoke tools and capabilities, which I think gives a much richer environment.

2:35:10I have a very painful lesson around diligence, which is why I care. I'm very excited about this technology. I worked on a deal last year before we started the show, and we were in diligence on this company. And there was a single document that killed the deal and the lawyers had already racked up like like a hundred like a hundred thousand dollar bill before we before we before we found it um and it ultimately meant that you know that all the parties uh walked away but we were still sitting there with like this massive bill and it was just like unbelievably painful because it wasn't even complicated it was just the business was incredibly old and there was just this one deal that they had done like a decade ago that meant that it was just like kind of dead in the water.

2:35:58Yeah, I mean, what the tech allows is that you can also do a DD on the cell side before you even put the documents into the data room and you say, hey, let's go open it up to the other side, right? Like it's funny that it's called a data room, right? Because it used to be a physical data room with a bunch of boxes. You would go into that room, you would open it up, you would start reading all the documents. Then we got control F and you could start control F-ing for different change of control clauses. And then came the early versions of the sort of AI systems where actually the seed for what became Leguara was founded.

2:36:29Back with the early BERT models, if you remember those. And you had to fine tune the system to find specific provisions. And now what's cool is you basically just ask the query in natural language and the system understands it. So you can even go in there and say on a scale of 1 to 10, how much of a smoking gun is this document? And the AI will just respond and provide its reasoning. So I think in the comment that you made, will some law firms disappear? Absolutely. I think law firms will go through a huge wave of consolidation, frankly, because they've never had tech and sort of real scalability as a weapon.

2:37:07And I mean, you have the big four in auditing, and tech has been very good for auditing. It's very good with numbers, very good in finance. Tech has never been good at language until large language models. And so now I think you'll see increasingly firms leverage these things to not only sort of out deliver their competitors, but also to out price them, you know, so is the is the solution for law firms, you think, transitioning to kind of value based pricing flat, flat, you know, more like project fees? because if you compare this to the creative world, the best logo designer in the world is not going to charge an hourly rate because they might be able to make that logo in 30 minutes, right?

2:37:49And so if they charge... You can't charge a$100 ,000 hourly rate, but companies will pay$100 ,000 for a great logo. So is that... Which is correct. I think that's correct for a lot of work. I don't think it's correct for all the work. So I'll give you an example. When we negotiated this term sheet, we were working with Goodwin and our partner Craig there. And when I need his expertise to help me negotiate the term sheet, he's worth a lot more than his hourly rate, which is arguably already high, but he's worth a lot more than that in that specific case. But when his team is reviewing simple documents, maybe that's not worth as much if AI can do that.

2:38:31So I almost view it as if AI can do a specific piece of task, it will do it. The only question is, like, where does that get transacted? And I think you'll see value-based pricing in taking a portion of the deals. Like, the best law firm in the world, arguably, is Wachtell. And they don't do hourly rates at all. They just say, this is the price to win. And you pay them. And they get a portion of that. But I'm also actually seeing another way where really big companies, like the big tech companies, are going to law firms and saying, hey, we're going to pay you$200 million. over the next two years, and we want you to do all of this work.

2:39:11Can you do it? And the law firm is then forced to say, are we going to turn down$200 million or are we going to figure out a way to do that? And of course, then tech will be part of that answer. So I actually - Yeah, it's about creating the right incentive to adopt the tech versus like currently some firms are going to resist it and being like, I actually like doing diligence the old-fashioned way where we could get three people to spend, you know all these hours uh this was super fun i'm super excited for you guys i have a feeling we'll have you on a lot more in the future just remarkable progress which we had more which we had more time i love it's a cyc company winning and i rarely see one winning this hard this is incredible this hard this fast i gotta give a shout out like we since we started in europe we just launched in the u.s back in mark so since then we've killed a team to almost 50 people and And the growth that's coming in the U.S.

2:40:03is frankly why we're doing the round. Congratulations. Thank you so much, guys. Thank you so much. We'll talk to you soon. Have a good one. Cheers. Bye. Before we bring in our next guest, let me tell you about AdQuick.com. Out-of-home advertising made easy and measurable. Say goodbye to the headaches of out-of-home advertising. Only AdQuick combines technology, out-of-home expertise, and data to enable efficient, seamless ad buying across the globe. Our next guest is Christina from Vanta. She's been on the show multiple times. She's back. We're getting an update. How are you, Christina? Great to see you.

2:40:29Thank you for supporting us for this entire year. It's been fantastic. You've been on a tear. What's new in your world? Well, I got a hat in the mail from you all, and it's great. Thank you. Well, you should have gotten a jacket, too, but the team may have sniped it. Oh, no. The jacket never showed up. The jacket never showed up, but really into the hat, so thanks for that. I'm glad. I'm glad. Things are going well. So a little bit, a lot of AI going on with everybody, as you're probably talking about. Getting excited for Vanticon, the security and compliance event of the season. Let's go. And just did some surveying of security folks about how they're feeling about AI.

2:41:08Yeah, we were talking to the CEO of CrowdStrike yesterday. Very interesting nuance. And there's basically like AI-generated viruses now where the actual code that hacks the system is generated on the fly by a generative AI model. And can do reasoning. Yeah. So your malware can think like a human now. I'm in danger. I'm in danger is what I feel like. But how are people feeling? What are you learning from the customer base, from the folks that you're talking to on the ground? Yeah. So we talked and focused on the ground, and then we also surveyed thousands of security leaders around the world. And the result's actually super, super encouraging on the AI front.

2:41:49And then I actually even guessed. So we found like four in five security leaders are using or plan to use AI agents to prevent against those sort of attacks. So the solution to the attack, to the AI agent attacks, is good guy agents. It's in fact more AI. Yeah, of course. Yeah, long tokens. Yeah. So there's that piece. And then we found like kind of a similar percentage, but 70 % of folks are letting those agents give input on security strategy. Yeah. And it's like giving it to a person who's considering it. Sure, sure, sure. We're in like default let the agent talk to your folks and influence them versus like default skepticism.

2:42:29And again, this is among security folks. So I was like pretty impressed with this. Do you have a sense of like how token intensive some of these processes are? Because I imagine that like one of the things you could do is just the benefit of AI is that it doesn't mind working 24-7 for months on end. So you could just say, actually go through every single line of my code base one by one. And then just, you know, see, does this does the same like there could be vulnerability here and just generate tons and tons of, you know, ideas. And then we'll have a human filter through what it thinks could be vulnerabilities.

2:43:00Yeah. So, OK, I saw your interview with Alex and Berkus yesterday. yesterday. And I think we're like seeing something similar where it's just like paralyzing. And we were like, oh, okay, go off, run, take all the tokens, like see what you come back with. And maybe it's something. And then I just did two things at once and maybe it's nothing. And then I'm like, I kind of just did the one thing and I have to move to the next task, but like, I'm in, I'm in the spot I am. And so we are seeing like a ton of paralyzational work here. Yes. Yeah. What else is empowering people on the AI side in security compliance?

2:43:35I can imagine that when we're talking about the technology, we talk about, oh, the future of the agents will just be able to do it. But then when we talk to people on the ground, they're like, I'm just using ChatGPT to do my homework, or I'm using it to get educated. And I feel like there's a whole world where we go and we rebuild the compliance training processes in an AI native way. In a not awful way. Yeah. It's a lot of times it's pretty awful. I've been through a few of those where you're clicking a button and watching a video and you need to move the mouse. Is there progress there? How fast is that rolling out?

2:44:12What does the future of compliance training look like? Yeah. Okay. Well, compliance training in particular, I think a ton and to your point, like adaptive and compelling and not the like um watch a video that feels you know at best stuck in the 90s or probably worse things you could say about it right yeah it's like where does this work um yeah we see folks doing like the phishing email was like always the classic thing but now it is just they're they're really good because it's also like oh let me scan your email inbox and be like what do you actually reply to not just like i'll send you the i don't know tvpn newsletter and like try or fake that and try to get you to click on it.

2:44:47Like it's super personalized. So there's some cool stuff there. To your other question, my model of this, kind of an overall, but I definitely see it in security and compliance to steal from Tyler Cowen, is like Centaurs. And so it's like AI plus person. And we're at that stage. Yeah. Right? And we'll be at that stage. And maybe, you know, AI will take it all the way, like the way it did with chess. Yeah. But even with chess, there was like a multi-decade, you know, AI plus person stage. and we're seeing a lot of that with our customers. Yeah, that's fantastic. We were asking the CrowdStrike founder about, basically, was he surprised that we haven't seen a Gen.AI-induced high-profile security incident yet?

2:45:33I don't know. Like a SolarWinds. I'm sure, yeah. Supply chain attack. Yeah, what's your, you know, So how concerned are you about that sort of like, you know, industry shaking? That was not a generative AI driven problem. It was just DNS configuration on a database. Okay. So actually building off of that, I think, I don't know, probably pretty soon, but I don't think it'll be some like crazy complicated attack. Like a kind of, you know, it'll be like, oh, you left your database unlocked. It's like the classic Equifax thing, you know, and it's like maybe AI spotted that. Sure. but like you kind of did the equivalent of like, well, you left your front door unlocked.

2:46:11Yeah, yeah, yeah. Yeah, so just another lesson, just like go back to the basics, drill the basics, human in the loop. There's definitely the centaur era is going to be here for a long time, I believe. I completely agree with you. Last question. Wish we had more time today. Like how are you and your team thinking about headcount planning over the next two years? Oh, yeah, that's a great question. It's like all I'm doing right now. In like an AI sense? Yeah, well, just not even in an AI sense, but just like it's a new world. It's a new world. We see public company CEOs doing layoffs and saying, we're doing these layoffs because of AI, and maybe some of it's because of AI, but others is just it's really good.

2:46:58Alex Karp told us he's not going to really grow headcount, and he thinks he can massively grow revenue without actually scaling up the talent pool that much. Government contracts are amazing. The way we're thinking about it is like someone or like AI won't take your job, but someone using AI might like back to the SEMR thing. And so have a bunch of stuff. And that's like our framing inside to Fanta and folks work at Fanta, whatever. And we have a bunch of stuff both to like encourage and teach and be like this. And then this will be a core expectation of your job. And so like that kind of model going forward and filtering on it and hiring and, you know.

2:47:36So you're telling the team like you should have three, three, at least three bots join every hangout that you do that are all summarizing, all summarizing the meeting notes and competing to summarize the meeting notes the best. I saw this too. I, yeah, I'm, I don't know. I'm a note taker and like notes make me pay attention. So I can't use those meeting recorders because then I stopped paying attention and I'm just like, I know this. Not everybody is like this, but yeah. Yeah, a lot of it depends on the actual use case. If you're in some sort of meeting where you need a full transcript, that's probably different.

2:48:10But certain types of meetings, it's going to make more sense. Vanta should make a really menacing agent that joins and is watching out. Make sure there's no funny business going on. Yeah, yeah, yeah. Say some words that will pop up at you. Well, thank you so much for taking the time to join the show today. We will talk to you soon. Great to catch up. Have a great rest of your day. Awesome, thanks for having me. Bye. Talk soon. See ya. Before our next guest hops on, let me tell you about Bezel. Your Bezel concierge is available now to source you any watch on the planet. Seriously, any watch. Okay, quick fire drill for the team.

2:48:44Is Elan from Nivon in the waiting room yet, Ben? We have - Because I know we have a couple more. Fireworks AI in the Restroom waiting room right now. Okay, but team, think about ordering here because Elan - We have a special situation because Nivan is going public. Today. And the IPO's timing might be... Yeah, I just want to be cognizant of his timing. Okay. So we will have the production team working on that. But let's bring in Lynn. Let's bring in Lynn from Fireworks AI into the TV panel room. Welcome to the show, Lynn. How are you doing? Hi. Thanks for having me. Thanks so much for taking the time.

2:49:23Big day. Would love to get your introduction to the company, how you're framing your business relative to everything else that's going on in AI. It's such an exciting time. And then we'd love to go into the news. Yeah, definitely. So Fireworks, we are an AI inference platform with application-specific inference. So here's how we think about inference differently. That we do not think about inference as we buy in compliances at home. For example, we buy a refrigerator, use that for 10 years, and then upgrade to the next one. We believe model should be continuously becoming smarter and smarter and learning and adapt into the application.

2:50:08So every single application developer, they should have their own model. They shouldn't use off-the-shelf any model other people use and become a model API wrapper. Instead, the model should be part of their product design. And the model is like we raised child. It will continue to learn and be smarter and be more specialized in solving particular problem really, really well. So that's kind of the inference platform we provide with smarter model, faster model, and a much more cost-efficient model for every single application. So, I mean, you're competing with – am I crazy? Are you competing with AWS, Azure?

2:50:47Like these are the biggest hyperscalers. hyperscalers, they are somewhat model agnostic, but serve up inference as APIs. Lynn's competing with everyone. Is this the craziest? She's coming for everybody. But I mean, take me a click deeper on the actual strategy. Is it just that there's so much demand that you can go and satisfy the extra demand? Or do you think you're building something that will just be completely differentiated from the hyperscalers forever? Right. So currently we are running very high traffic. We are processing more than 180 requests per second. That is an order of magnitude of, it's same in the ballpark of Google search traffic volume.

2:51:28We process more than 10 trillion tokens per day. Okay. Basically, hit that goal. People's earning call. That's similar to their...

2:51:40That is actually the same as their token processed from Gemini. Wow. So, we process a large volume of data. And that just means Open Model has the future. And there are so many startups and enterprise, they reinvent the new user experience. and they want to have full control of the model as their IP and have the model co-develop, co-evolve, co-adapt into their product. And we see that as a future. And that's how we differentiate from product focus point of view. And we do not provide this inference one-size-fits-all. We provide inference one-size-fits-all specific tailored towards application.

2:52:23Are you GPU rich? Are you GPU poor? Is this part of your strategy to have GPUs available when your companies need them? Is that a differentiator? Yeah. I would like to say we design our product for GPU poor because, interestingly, we are at the beginning of S-curve of huge AI Navy product explosion. But product market fit today doesn't mean a viable business. in the sense that many of the, we talk, we work with so many companies, they just couldn't scale to millions of developers or billions of consumers. They're going to scale into bankruptcy. Fundamental, infrastructure is so expensive. Even if they build the best product, it has so much value, they just couldn't make a viable business.

2:53:14Scaling into bankruptcy is an iconic line. It's a great line, Lynn. I'm not going to name, there's one company that really comes to mind. I won't name them. Give us the news from today. On the fundraising side. Get that gong ready again. What happened? We raised our CFC. Yes, give us some more details. Who came in? How much? What is the valuation? What else are you sharing about the round? Yeah, we raised$250 million. Co-led by Lightspeed and Index. Boom. Love it. With participation by Eventik and Sequoia. Fantastic. You can just call it a quarter billion now. You're getting the really big number range.

2:53:57It brought us about our valuation to$4 billion. $4 billion. Wow. So really excited for all the customers, partners, and investors. And quickly, how old is the company? We're three years old. Wow. That is impressive. Congratulations. Well, thank you for everything you're doing to power the AI revolution. Helping people avoid scaling into bankruptcy. That should be the company motto. That's honestly one of the greatest lines. We're going to be using that line quite a lot. So thank you for that. And it's great to meet you. Congrats to the whole team on the milestone. Thank you so much. Have a good day.

2:54:32Before our next guest hops on, let me tell you about Wander. Find your happy place. Book a Wander with inspiring views. Hotel-grade amenities, dreamy beds, top-tier cleaning, and 24-7 concierge service. It's a vacation home, but better. Our next guest is from Navon. We have Alon Twig, the CTO and co-founder of the world. There he is. Welcome to the show Get that gong ready Congratulations Let's ring the gong preemptively Preemptive hit I love it Now we can have a conversation Champagne Cheers, cheers We have Happy Dad Nobody's joined Nobody's joined Drinking Champagne Ever on the show If you're joining Drinking Champagne I'm going to have a Happy Dad From the Nelk Boys Go do that From my friend John I had a really long day So that's my last thing before the big party starts.

2:55:21Well, congratulations. Amazing. And you're live from the NASDAQ. How's the day been? And yeah, maybe kind of back up a little bit and actually introduce yourself. I think people can probably tell what you're up to. Yeah, sure. So my name is Ilan Twig. I clearly like champagnes. I'm the CTO and co-founder of Navan. Navan is a travel and expense. management company. We started there more than a decade ago. Today we IPO'd. Super pumped. Congratulations. It was a beautiful day. Yes, exactly. My entire family, by the way, I think I broke the record of the number of family members that joined me to the ceremony.

2:56:04That's amazing. It's so funny because the last time we were at an IPO, it was for Klarna, Sebastian, who's a friend the show and he for him it was it was just like another day he was like oh yeah i didn't even bring my family i just kind of stopped by ipo the company come back but so every founder every how many family members like processes these things differently it's always interesting to hear how people treat it how many roughly like 30 30 no it was just for my family it was 17 17 17 family members at the IPO. Yeah. Congratulations. Really amazing milestone. And by the way, if I may say one last thing about the guests.

2:56:47So my parents were here, both of them. Sure. My dad turns 90 in two weeks. Wow. And he was here with me on the stage. He was like happy. That's amazing. I think most of my excitement was watching them. Totally. Being so excited. Yeah. Not understanding anything because they don't really speak English, but they kind of got the gist of what's happening and it was super exciting. Yeah. That's incredible. Take us back through, I mean, I'm sure when you're looking back into like the depths of COVID, which probably it feels a long time ago, but you were in a business that was, I mean, I just imagine you aged more rapidly than maybe Brian Johnson would like during that period.

2:57:30Walk us through kind of some of the key moments getting to this point. Yeah. Maybe I should first tell you that before COVID, I had black hair. It was all black. Oh, wow. It looks good on you. It looks good on you. You look more like a serious executive now. Yeah, yeah. I can fool people around you that I'm serious. Yes. You know, you run a company that the business model is we make money when people use the platform. There was no monthly fees. There was no way you don't pay. No long lock-in, no enterprise contract. No enterprise contract. And why? Because both my co-founder Ariel Cohen and I, we had this philosophy that if the company generates money on usage, then the company will always be focused on making sure that people would want to use the product.

2:58:20because it's a great experience. And it worked well for us for the first five years. And then when COVID hit, no one was traveling and no one was using our platform, not because it was an inferior product. It was because there was a virus and no one could travel. And so we found ourselves with the best product out there with zero usage and zero revenue. And it took six weeks from like, I don't remember the number, maybe 100 million to zero. Six weeks. So then what? Yeah, it was a very interesting point in time. But I can tell you that, you know, five years later, here I am with a glass of champagne.

2:58:59My parents were here. We were celebrating the IPO with a company that has way more products, a much more resilient business model. And trust me, this time it's pandemic proof. So we are not going to go to zero. We'll go down, but not to zero if it happens again. So we learn, you know, that's why we have gray hair. Battle scars. And yeah, it was definitely a challenging time, but I think that, you know, judging by results, I think that it was managed well. Obviously, the entire company, the entire management team, the board, we were all united around the decisions and here we are. How are you? We were talking to Brian Chesky at Airbnb last week and he was sort of mapping out his vision for the future of travel.

2:59:44this idea that as we become more online, we also maybe want to become more offline as well. And maybe we're traveling more as well as spending more time just online. How do you internally in the company think about the future of just business travel, pleasure travel, just like the overall travel market? Right. So first I should tell you, you know, I'm the CTO and I love technology. I mean, this is my thing. I really, really love technology. When ChatGPT came out, the first thing that I did, I bought like a$30 ,000 worth of a GPU. That's what I did. No way. And I built a PC. My co-founder thought that he should fire me.

3:00:27It's like, you have a PC under your desk? I said, yeah. And I paid for it. You don't worry about it. I paid for it. And then I started deploying LLMs on top of it, unsensored LLMs. I dove deep into that. And I really wanted to know the nature of these creatures, how they behave, because if I were to deploy anything to production, based on these creatures, I call them, I need to know who they are, how they behave. Will they lie, for example? Do they know what a lie is? Will they use it? No, they lie, yeah, but forget the information. They will also lie for other reasons, intentionally, like they would not want to lose, for example, and they would lie in order not to lose.

3:01:05Lots of discoveries. but ultimately it led us to, once we really understand the technology, how to deploy this thing to production and let it run completely unsupervised and do stuff for us. And I think the important thing is completely unsupervised. The first thing that we've done is we deployed Ava. Ava is our virtual travel agent and she communicates with travelers as things happen, when they need to book something or there's disruptions and they need to push their travel by one day, cancel this leg, argue about the refund, et cetera, et cetera. And she would do all of these things for them.

3:01:43And it is all based on an internal framework that we built. We call it Navan Cognition. I really like the name Cognition. I didn't invent it. It was a professor from Colombia that we worked with. And why Cognition? Because it turns LLMs into a highly functioning cognitive system. So Ava today is completely unsupervised. She manages thousands of chats every day with a very, very high customer satisfaction that matches the human level. And she offsets more than 50 % of the incoming chats. She does not deflect. She solves. So I think our approach to travel is we look for ways to leverage technology to make travel easier so that people can focus on being there rather than getting there.

3:02:30getting there is just a noise that you have to it's a penalty that you have to pay but the focus for corporate travel is to be there and to be fully there and there's people outside trying to make me laugh from outside and I'm a little bit I'll try it I just have one suggestion make an agent one of your baby creature and call it Twiggy, please oh that's a good nickname Microsoft has Clippy Nivan needs Twiggy you already went through one you went through a rebrand you were Trip Actions now you're Nivan I want you to rename Ava to Twiggy I love it I'll do my best I'm going to dream for that cheers cheers thank you so much for coming on the show on such a big day congratulations such a monumental moment and please extend our congratulations to the entire family all 17 all 17 and the whole team it's a beautiful moment America loves an IPO I hope you have a great rest of the day Thank you so much I'm going to party tonight Thank you so much Thank you Love your podcast What a legend That was one of the most legendary That was tough to follow We have Taranjeet coming in next I think that's how I pronounce it Memzero He's in the restream waiting room How are you doing?

3:03:59Great to see you. Thank you so much for hopping on the show. Really tough one to follow. Yeah, the last guest at IPO, he was drinking champagne, he was drinking. Raised a billy. It's a crazy woman. But I'm excited for this. But thank you so much for hopping on the show. We're very excited to talk to you. Yeah, thank you for having me. Great to be here. Of course. Talk to me about the business. Introduce yourself in the business. And then I want to go really deep into how memory will actually transfer and what the equilibrium of the market will be. But we can go into all the technical or like market structure stuff after we kind of get the general news, where the business is, the shape of what you're building.

3:04:38Thank you. So I'm Taranjeet. I'm the co-founder and CEO of Memzero. Memzero is building memory for AI agents. Right now, everybody is trying to create an AI app, AI agent. But there is a fundamental issue. all the apps are done because they are stateless. They don't evolve with you as humans evolve. We are trying to fix that. We are the, so far, we are the most dominant player in the market with like 41 ,000 GitHub stars, 14 million downloads. And we recently announced our 24 million funding led by Basis Set Ventures with Kindred, P50, IC, and GitHub. Total agent domination. So, I mean, I got to just jump straight into the actual market dynamic because I have to imagine that right now OpenAI is saying where you're going to log in with OpenAI and we're going to store the memory because we want you locked to us as much as possible.

3:05:36And for everything else, if we want to build it within the app, we want to do one-click checkout in the app. And if we're going to send you out, you're going to OAuth with ChatGPT. And so are you the Android to ChatGPT's iPhone? Or is there a way that you can play together? Is there a data privacy thing where I can go to Sam Altman and say, hey, look, I know I'm daily driving the ChatGPT app. But you got to let me take my data with me. my personalization data with me because I want to go out into the open LLM ecosystem and interact with things all over the place? Yeah, I think that's a very good question.

3:06:15And there are two ways to look at it. First, it's good that big labs are adding memory in their consumer offering. They're actually educating the market that memory is needed. And they also realize that as models commoditized, memory is the next mode. And it's a matter of time that memory becomes available as an API. but you know developers who are building all sorts of cool ai apps are using multiple llms so in that scenario because memory is not just read only it's also right you would want to decouple it from llms out there so that's like take one now the fundamental thing for why you know this is happening is you know we are going through a technological shift so far we have been interacting with click swipe and scroll but for the first time we are we and everyone on this planet is getting the ability to talk to software you can chat with it you can say a few words it will say it back to you so you as human i as human we are accumulating rich personal context now right now it's a handful of ai agents that we use on a day-to-day basis and we use like chat gpt cloud and you know other apps are there but five years down the line it's going to be a lot of apps in our life and imagine if i as a user or you as a user have to explain every app who you are that's going to be frustrating for you.

3:07:29So in that scenario, we strongly envision and we strongly will make this plaid for AI kind of, you know, functionality out there within you as a user are empowered to carry, to own and carry your AI memories across any app or device that you interact with. Okay, so if you're the plaid for memory and in the future, I am using, I mean, this is a real story. I use Gemini all the time for VO3. I also, if I'm interacting with a YouTube video, I'll always go to Gemini for that. Gemini is faster in a bunch of different places, but I use the voice mode on ChatGPT. I also use Claude every once in a while.

3:08:12I have little memories all over the place. And if I'm in the Platte analogy, that's similar to having a Chase account over here, Morgan Stanley account over here, an E-Trade account over there, a B of A account, a credit card, all these things. and I want to bring those together. Is that something that you imagine I'll be able to do when someone comes out with a completely new product that they want to customize? I could go over, we were talking to Adobe or something, and then I could say, yeah, I'm going to auth with Memo and then I'm going to allow you to access my memories or my memory in all of the other apps and port all of that in.

3:08:51Is that something that they would let you do? see that's uh that's the future that we feel is highly possible they won't i mean like big labs won't let you do because for them memory is one of the modes right yeah but if you consider from other side of users then like you know the world is not going to be just one app the world is going to be multiple apps right it's like open i will be there and jenna is there claude is right in that scenario once so right now it's like a technological shift so you and i can relate but not everyone on the planet right but soon they will relate and because of this new interface of voice like i will feel like why do i have to tell this software again and again why is this software so dumb in that scenario i think like users should be empowered to own and carry their memories so far it has not been the case because it's just been like clicks data you know your session history but now you're sharing a lot of personal rich context you should own it you should be saying that okay i give my health care memories to this app i give my financial memory to this app and they are in sync i think that's a future that benefits everyone at least everyone on the planet yeah is cold email underrated i i actually changed my life uh because of cold emails like i wrote like 200 cold emails to everyone almost everyone in silicon valley back in 2003 i still write like a lot in this round also I still wrote like cold emails to angel investors uh I think uh it's the most valuable asset uh for everyone out there you should be writing more cold emails everyone should be writing more cold emails I love it uh handwritten or AI generated what's it take no I I still write like everything personally I just take AI help to fix you know some grammar or some structural issues, but should be personally written.

3:10:41Yeah, the AI cold email. Hey, saw you're in New York. Hope you like, I bet you like Central Park. Yeah, do you like Starbucks?

3:10:52Anyways, congratulations on the fundraise. It's great to meet you and hopefully get to use Memzero out there in the wild very soon. Yeah, thank you for having me. Fantastic. We'll talk to you soon. Have a good one. I have the perfect post to end the show on Please hit us That is that Tyler's app is broken Oh, taking shots of Tyler He's doing his best Did you want to do the virality equals taste one? No, I wanted to do Maytob says I keep getting ads for cocaine on Facebook Oh yes Just using Linktree as a cloaking page Kind of surprised this works And to that, I would say, how do you think who's going to pay for the CapEx, right?

3:11:41Oh, okay. I see you're bringing it back to where we started the show with meta earnings. If you see this, you got to assume, you know, whoever's running that ad, advertising cocaine, they got to pay a pretty penny. They got to pay in their Facebook ads manager. This is so funny. There's a reply here from this guy, Zach. What data does Meta have on you? And Metab says, I was in L.A. for a day, I guess. This was a thing on Facebook of people advertising drugs. And it is a game of whack-a-mole back and forth trying to catch people because the lingo will change. And the way people are kind of dog whistling to each other that they have something illegal that they're selling.

3:12:30it's a huge game of cat and mouse that the meta team is kind of constantly working on. And you see those same kind of like breathless, panicked news articles about like, is Facebook like a hotbed for, I guess, cocaine in this case, or something else and a variety of different adult content or whatever. And it's always this like dance of like, how much are How much are they going to invest in trust and safety? How many people are going to be in the loop? How much AI are they going to throw at it? How are they going to have false positives? Because what if you're just posting a meme about Tony's, not Tony Soprano.

3:13:10Tony Montana. Tony Montana. Yeah. Our favorite song that we used in our launch video. If we post that, there's obviously a reference to illicit substances in there. Do you want that to be flagged? No, obviously not. And so it's a constant dance. But very funny. that someone actually slipped this through the Facebook's ad review team. I have a funny story about the Facebook ad review team. There was a time when you could straight up, like there were companies that would basically go and find the individual ad manager who would approve, who would have admin access to the Facebook side of the ad manager.

3:13:55and the company would go and put up an ad and try and get it approved and it would get shut down. And the ad would be like, Harvard scientists created this limitless pill that's used by Johnny Depp. It was always Johnny Depp and it was always Harvard scientists. And it would be a picture of Johnny Depp looking great and then Harvard scientists. And it would link you to some scammy website that would sell you basically some fake nootropic or fake supplement. And these ads were against terms of service. Like it's straight up snake oil. Like it's not even like, there's no thesis of like, oh, it's actually good for you.

3:14:30It's just completely fake. But they realized that it would convert and it would be very high margin. And so what they would do is they would go to the Facebook ad manager people and they would try and bribe them and say, hey, if you go in and you approve my ad, I'll put a hundred million dollars of ad spend behind it and generate$200 million in revenue. And if I can get that ad approved, if you as just the individual ad manager can go and tweak the back end. And so there's like a constant war even inside the organization to make sure that nothing illicit is happening. I heard something interesting recently that there's a health influencer and people make AI videos of them just promoting random supplements on YouTube.

3:15:14So they'll make an AI video. They'll post the video and in the description, they'll put links out to products. and YouTube does not, it will not systematically identify the content themselves. It's just they say that we support AI video. Sure. And so this influencer is paying another AI service to consistently monitor videos and then report them automatically. Yep. But they have to spend$10 ,000 a month. It's 10 grand a month. on the service because it's super, it's super intensive from a, and, and this influencer like needs to, uh, so you can imagine somebody's, you know, I know exactly what you're talking about.

3:15:57Yeah. There's a, yeah, it's a, it's a constant war. It's a constant war. The solution to bad AI is good. Yes. Good, good AI agents with guns, I guess. Give your Neo robot a Glock. Uh, it needs to happen. Anyway, thank you so much for tuning in today. Thank you for listening to the show. we will see you tomorrow it's Halloween we have a special show for you please tune in leave us five stars on Spotify and Apple Podcasts tomorrow is going to be one for the books spooky it could be silly have a great have a good one and I can't wait goodbye have a great evening