All-In with Chamath, Jason, Sacks & Friedberg
00:16:37 · Will Ahmed, Jason Calacanis
And so if you can fight that trend, that... Those averages and keep your heart rate variability even just flat over years, that's actually a sign of good fitness, good longevity. And, anyway, back to the story I was telling you, the... This individual, they had a very steady heart rate variability that was within a specific range and all of a sudden it plummeted. So it was a healthy person. They had a heart rate variability of about a 100 and and then it was 50 all of a sudden for for a period of time. Again, goes in, gets all these tests. He turned out to have testicular cancer. Oh, wow. And so [00:17:11] S1: Now was that known previously that HRV would in some way be a a lead indicator for testicular cancer? [00:17:19] S0: I mean, I don't think so. You know? That's what's so fascinating about this moment in time is we've never had this much data on the human body continuously. One of the principles to founding Whoop was twenty four seven data. And it's it's why we have a fourteen day battery life now. It's why we we invented a charger that you slide on to your Whoop. [00:17:39] S1: I was about to say these are super important. And when you first told me about this on on this week in startups, I was like, wow. I need a... I wanna see how many steps I've done and the time. It makes no sense. And you were like, trust me, Jake. I'll... [00:17:51] S0: You know,
00:17:51 · Will Ahmed, Jason Calacanis
[00:17:51] S0: You know, [00:17:51] S1: there's a reason for this. Yeah. And now that I've used it, it turns out you're absolutely right. The reason I have had so many long streaks... The only time I really stopped using this for the past couple years is when I lost it and I had to order a new one. So it broke my streak, which was a bummer, but it is what it is. It's just a number. Because you can wear it in the shower, you can wear it in the bath, and you should. It turns out a warm bath is... When we get to journals in a minute, super, great for my recovery for whatever reason and my sleep. You never have to take it off. And you can slide on the battery charger. So you can have essentially the opposite experience that I had with my Apple Watch. The Apple Watch was annoying because it's... I'm getting alerts, and I'm getting alerts for things that I... App I installed two years ago, and it's just constantly bothering you. People think you're rude when you look at the watch, and then the battery dies every, you know, whatever. For the Apple Watch, I think I might have gotten to twenty four hours or thirty six hours, forty eight hours a couple times, but it's a terrible experience. With this, I I mean, I I've made it ten days, you know, fourteen days. No problem. And I carry... I bought three battery chargers. I just keep I keep one in my desk. I keep one on the bedside, one in my bed... My travel bag. It's like these things aren't expensive. And and that's a key to it, think, as well. Yeah? Is the the amount people wear these.
00:19:14 · Will Ahmed, Jason Calacanis
[00:19:14] S0: On average, a Whoop member wears Whoop 23.6 a day. And as a consequence, we have just over fifty billion hours of continuous physiological data. So it's one of the largest health data sets in the world. And now with the, you know, the advent of AI and what you can do with large data sets, it's just like a really exciting time to be working in this space. [00:19:40] S1: Do all the AI companies, like the frontier models, wanna buy this data from you in aggregate? Is that like a huge business for you guys eventually? How do you think about that or building your own models? [00:19:51] S0: Yeah. Today, we're building our own models, you know, foundational models within the company that are essentially training on top of this data. So one So [00:19:59] S1: you're using open source models or you're you're you're [00:20:02] S0: We're developing our own models. We've got open source models. And and what's amazing is you've got this huge physiological dataset, but then you also now have other datasets. Right? So people are uploading their blood tests or they're doing advanced labs through Whoop. They're uploading medical records. And and so as an example, we're we're now starting to see indications that through physiological signals, we might be able to predict that a biomarker is out of range without actually measuring the biomarker itself. [00:20:33] S1: Wow.
00:20:33 · Jason Calacanis, Will Ahmed
[00:20:33] S1: Wow. [00:20:33] S0: And so then, you know, you could give someone a very targeted, hey. We think you're... You know, we think you have high cortisol. You should get this one test. And and it gets... And, you know, and it gets super it gets super targeted, which is exciting. [00:20:46] S1: Which the AI would be able to figure out. The AI is gonna look at all this data and say, hey. Maybe there's a cortisol this year or here's the possibilities, which is super fascinating. So not... You you see your business in the future as this AI assistant, and you added an AI assistant, I think, to the product less than six months ago. I'm addicted to it now. Yeah. And you have these ongoing conversations. It logs the conversations. And so now it was, I'll be candid, pretty dumb at the start. It would be like, you know, you could talk to it and you'd be like, yes. You had this rate. Okay. Thanks for that rate. But now it's gotten a bit more sophisticated because now when I tell it, I'm in Japan. You didn't record my sleep. It's like, oh, okay. Here's what's happening. And it actually is understanding the product, and I I I guess you're doing some reinforcement learning. [00:21:35] S0: Yeah.
00:21:35 · Will Ahmed, Jason Calacanis
[00:21:35] S0: Yeah. [00:21:36] S1: And it's gotta be that these 4,000,000 members, including myself, are now telling it where the product breaks down. Because when you... Or or where the opportunities are to add more features, however you wanna look at it. But there's this constant struggle I was having with changing time zones. And was my, was my sleep getting recorded properly so then I can get my recovery? And then I started asking, and then it prompted me and said, hey. By the way, when are you leaving for Japan or whatever? And I was like, here's my flight. This is when I land. And it was like, you should sleep at the first half and then not go to bed when you get to Japan. Don't have caffeine at this time. So somebody in the company, I think, explicitly started doing a time zone change, or did it just learn that and infer it? What... Did you did you get specific about that, and how are you using the AI coach? [00:22:28] S0: We spun up a team on travel specifically because it's such a popular use case for a lot of our members and, you know, our origins were with athletes, but now we've got this wide audience of aspiring people. You know, I'd say that's really the... If you think about who is Whoop targeting, it's the aspiring. And that could be a great athlete, that could be someone who's fit, but it could also just be someone who wants to have more energy for their family and their job and their life Yes. Hobbies. That's right. And wants to live longer. Right? [00:22:54] S1: Yeah. Live longer.
00:22:54 · Jason Calacanis, Will Ahmed
[00:22:54] S1: Yeah. Live longer. [00:22:55] S0: But the AI, the AI coach is a big part of that, and it's a big part of our future. And, you know, we've been working on it for years now. I'm glad to to hear that you feel like it's getting good. And Yeah. About 80% [00:23:06] S1: our say it's like a seven and a half right now. [00:23:08] S0: Seven and [00:23:09] S1: half out of 10. I think there's so much potential for it. [00:23:11] S0: Yeah. As long as it's good enough for you to be engaging with it, it's just gonna keep getting better with time and we're getting better at that too. And and so, you know, one thing that's really interesting, again, back to these datasets. So physiological data we talked about, that's from the sensor. Then you've got the advanced labs and the blood testing. You've got the medical records coming in. But actually this is this is another very important piece of data which is all sorts of user feedback. Right? Like the journal itself. So about just over 50% of our members journal every day. [00:23:47] S1: It's a good point to explain what the journal is. Yeah. So this is I'm sorry that I keep explaining your features, but I'm so passionate about the product as you can hear.