The Twenty Minute VC (20VC): Venture Capital | Startup Funding | The Pitch
00:26:06 · Harry Stebbings, Matteo Franceschetti
[00:26:06] S1: The Humanoid Games, I think. [00:26:08] S0: Yeah, it's insane. The way I look at that is, okay, you look at this thing, it looks a bit funny, they look a bit clunky, but they start doing these things. Knowing how fast AI is moving, if next year, those same games will be mind blowing. That is what worries me. What I saw today, right, if you just look at that and you say, oh, hello, look at the thing, these robots, you know, they are not really good at walking. Even if they did the records in the 100 meters, there was a robot that ran faster than Usain Bolt. [00:26:37] S1: So they know how to I did a video on this. There was robot that beat the high jump. [00:26:43] S0: Yes, I saw that. And as they were fighting, you know, the fighters. Tennis. Yeah. [00:26:49] S1: Like, real Tennis time was really cool. It's insane. Yeah. But in one year from now, we will be blown away at speed of AI. Did you see the paper last night, or essay or post last night from someone at OpenAI, Sam promoted it, that basically said, we actually need to put a slight pause on this. It is getting too far too fast. [00:27:10] S0: Yeah. I think the part that is worrying everybody is the part of self learning from these models, and so they are getting to the point where they train themselves. And that is when we start losing control. There is a point, even in our [00:27:22] S1: model Do know that we're already there?
00:27:22 · Harry Stebbings, Matteo Franceschetti
[00:27:22] S1: model Do know that we're already there? [00:27:23] S0: I think we are. Then I think there is a big delta between what we see as users, and probably what the frontier model companies have in house. And I think they have already seen it behind the scene. You have seen some of the things that are happening already in terms of cybersecurity, in these civilizations, so that there were three attempts of civilization, I think, in an OpenAI model where agents were building their own group. [00:27:49] S1: I'm gonna go there. You sell in China. You you spend a lot of time in China. A lot of your investors, including Keith Raboy, who's a friend of mine, fucking hates China. How do you reflect on that? Is it as evil and CCP driven as it seems? Are we in a deglobalized world that should hate China? How do you think about that? [00:28:08] S0: Yeah. I don't have a specific opinion on that, so I cannot really comment. I can just tell you, for us, it's key market, it's a big opportunity, and we decided to sell there, and it's doing really well for us. [00:28:20] S1: Is it very different selling in China? [00:28:22] S0: Yes, because you cannot sell on eightsleep.com. Literally, you don't sell on a website. They don't use website. And so for us, it was literally learning everything from scratch. [00:28:31] S1: What do you mean they don't sell on a website? [00:28:33] S0: Chinese don't buy on websites. There is no e commerce as we know it. They only buy on their super apps. So on WeChat, Rednote, and equivalent. We had to learn all that.
00:28:33 · Matteo Franceschetti, Harry Stebbings
[00:28:33] S0: Chinese don't buy on websites. There is no e commerce as we know it. They only buy on their super apps. So on WeChat, Rednote, and equivalent. We had to learn all that. [00:28:44] S1: What the fuck? So where do they land? Like, you send a referral link to your friend? Where's the referral link from? I'm confused. [00:28:51] S0: Yeah. Imagine that everything, the equivalent, to oversimplify, is like if you were buying everything on WhatsApp. They don't go on any website. Website are, at the most, just literally like if you just want to scroll and see some picture and read a bit more about the product, but all the purchases, all their credit cards are on Rednote or kind of Amazon equivalent and WhatsApp equivalent, or TikTok, obviously. [00:29:15] S1: Does that impact how you think about how we should think about security for these AI assistants? Everyone now is like, Oh, would never put my credit card in Instinct. Or, What if it gets hacked?
00:29:15 · Harry Stebbings, Matteo Franceschetti
[00:29:15] S1: Does that impact how you think about how we should think about security for these AI assistants? Everyone now is like, Oh, would never put my credit card in Instinct. Or, What if it gets hacked? [00:29:25] S0: Yeah. I think the solution for that to me is, for example, my Grock bot is buying things for me already, but is using Amazon on the browser. And so Amazon has my credit card. And so there I feel confident because I'm not sharing my credit card with AI yet. I think in AI, we are where we were with credit cards on spending and buying things on the internet twenty years ago, where people, particularly European and Italian, I was still living in Italy, they literally didn't want to put their credit card on any website because they thought it was a scam. And sometimes there was a scam back then. I think we are at the same point with giving our credit card or information to these bots. A bit... Sometimes you feel uncomfortable, but I think in a year or two from now, our bots, particularly bots from frontier models, from X, OpenAI, and Cloud, you will start trusting them soon. [00:30:17] S1: What seems crazy today that you think will be very common in five years? I totally remember I wasn't allowed eBay when I was young because my parents were like, no way are you putting credit card details online. I remember when it was like, you think you're gonna find a partner online? These apps are ridiculous.
00:30:17 · Harry Stebbings, Matteo Franceschetti
[00:30:17] S1: What seems crazy today that you think will be very common in five years? I totally remember I wasn't allowed eBay when I was young because my parents were like, no way are you putting credit card details online. I remember when it was like, you think you're gonna find a partner online? These apps are ridiculous. [00:30:34] S0: Now 60% of relationships are online, and I think we live online transaction wise. Probably your bot buying a cybercap that will start making money for you. So you will buy a cybercap that makes money for you with a bot, and the bot runs the cybercap. I would be pretty sick. [00:30:52] S1: I mean, it's not that far off happening now. [00:30:55] S0: No, and as it will be the health part. That is another area where I hope that cancer can be solved within ten years. I think companies like Eight Sleep will be able to save your life, save the life of millions of people. [00:31:07] S1: Does the Eight Sleep product actually change that much with the greatest of respect in a completely AI driven world? [00:31:12] S0: 100%. We start seeing the likelihood you develop diabetes in our data, hypertension. We are becoming a medical device for sleep apnea. So your bed will become the most advanced medical platform you ever had. Guy, said about retail earlier. [00:31:28] S1: I bet retail has been a question for you for years, and I bet there's been different board conversations had. How do you think about when's the right time? Because you're still not... Still d to c. How do you think about when's the right time?
00:31:40 · Matteo Franceschetti
[00:31:40] S0: Well, let me step back for a second. One thing I'm learning as an entrepreneur is there is a time for things. Right, similar to what I was telling you about the out of home campaign. Nailing the timing of what you do, so when you do what, is actually really important for CEOs. Because it's not that it's binary that, I don't know, retail doesn't work, or out of home doesn't work. No, it works, but there are some second or lower level conditions that are required to make that project work. And so I think for Eight Sleep, that is happening now, and so you should imagine retail to become a meaningful part of our channels in the next month, years. But before, it would have been too early. Same for China, right? Launching in China, one thing I spoke to a lot of Chinese founders, and what they told me is sometimes when they start their consumer products, they actually start selling outside China first. And I say, Why? You have such a big market and you know the market. And they say, Because the Chinese market, they love to buy products that are already well recognized in the Western world. And so you want to build the product and the brand first in the rest of the world, and then you come back to China. Otherwise, in China, you would be one of many knockoffs, and it would just be a price war. And so I didn't know, we just decided to launch China, and so we were just doing it, but I didn't understand that for us it was the right time to get into China because now we're already selling in 35 countries, we are well recognized, we sponsor Charles Leclerc, now work with Ferrari through Charles. That is a very well recognized brand. Many athletes use us. Elon spoke about us. Mark Thackenber spoke about us. Elon is really big in China, right? And so you have a level of trust where you command a certain price. But if we were doing the same thing five years ago, we would be destroyed by a knockoff that is $500 cheaper.