Datasets:
MasaFoundation
/
bankless_Longevity__Zuzalu_3

Dataset card Data Studio Files Community
text
stringlengths
15
6.25k
Speaker E: It's quite the list.
Speaker F: Yeah, you've got your hands everywhere a little bit. And then also wins on tour, which regrow your nerve endings, which is really cool for neuropathy and pain. And a couple other companies that are in stealth mode. One of the fun ones is actually through the Methuselah Foundation. Vitalik donated 43% of the world supply of this fun meme token community called Dogelon Mars.
Speaker E: Dogelon Mars. Yeah. That's on the long tail of dog tokens for sure.
Speaker F: Yeah, it's actually, I don't know if it still is, but I think it was dogecoin, Shiba and Dojilan Mars, which is number three in the world. It's incredible fun community and yeah, it's a really interesting story for sure.
Speaker E: Yeah. Well, thank you for guiding us through all of this crazy, crazy topics. This is only like one of many, many longevity topics. So thank you for walking us through the top of the rabbit hole and I hope to continue going down it throughout this week.
Speaker F: Thank you. Thank you so much for having me.
Speaker A: Thankless nation.
Speaker B: We are here at Zuzalo and I'm talking to Michael Greer of humanity. Michael, what's up?
Speaker D: Not much. Great to be here.
Speaker B: Michael, I first met you at the 08:00 a.m. cold plunge and you have been there every single day.
Speaker D: Been best friends ever since.
Speaker B: Best friends ever since. So you're at an interesting intersection between all of the various topics that are being talked about at Zuzalu. The intersection of longevity, which we had a longevity week. I think we're going to have another one and also AI. So that's a fun little place to operate in. Tell us a little bit about how you stumbled your way into that world.
Speaker D: Yeah. And honestly, I think you can throw the public goods and the crypto into the mix there as well. Yeah, I mean, I started with, I think a lot of people in health tech start is they have someone close to them or themselves have kind of a really tragic health event. That kind of, you learn the cliche that if you don't have your health, you don't have anything. Right. And so I had people close to me find out really late about cancer. I think a lot of people, unfortunately, have had that experience and that threw me for a loop because I was in my twenties, I had this big business success, you know, this dating site that got quite large. So you're feeling like you can conquer the world and then you can't do anything for the, you know, the people sitting right next to you. And so I went down many rabbit holes, basically trying to figure out, like, early detection of. Of disease, of cancer. That led me towards genetics because of things called liquid biopsies that are now coming to fruition, which is a great space. But when I kept on going further, I went out to the valley. I took over operations at a consumer VPN. I started meeting people that started giving me this. It sounds like a very simple idea, but I didn't have it in my head before, which was, there's something better than early detection of disease. You can actually monitor the whole body going towards higher probability of disease.
Speaker B: Right.
Speaker D: More susceptible.
Speaker B: One step in front of that. Yeah, yeah.
Speaker D: Because what, what the whole. I mean, we don't need to go into the whole healthcare system, but I think most people that spend a lot of time in the healthcare system call it sick care now because we focus at that end stage where it's not too late, but it's like the person's already pretty diseased and you're trying to save them. Right.
Speaker B: And you're saving them in the most expensive part of that trajectory and also the part of the trajectory where they are suffering the most.
Speaker D: Yeah.
Speaker B: And so that is the part that the current institution of medicine is prolonging the most.
Speaker D: Yeah. And. And it's good that they're there. The problem is not that they're doing. They're giving sick care. That's great. And they do amazing job in hospitals doing that. The problem is that we don't have a lot of emphasis on much earlier.
Speaker B: Right.
Speaker D: In that. In that process. And so a lot of people have. Well, almost everybody has the intent to stay healthy or to be healthier, but there's just very few tools out there that can actually lead them in the right direction. And so, eventually, once I had that idea in my head, I don't get excited about stuff until I can really understand, like, how the whole operation works. And then I got excited. And so that's why we created humanity. And humanity basically allows you to monitor your rate of aging, which is basically your probability of disease. Is it going up or down and then basically guides you towards slowing it down.
Speaker B: Sure.
Speaker D: So across our user base, we're looking at all the users, what they're doing. Then we're finding users like you. So, you know, your kind of biological sex, your age, you know, different attributes of you, and seeing what seems, what group of actions is seeming to positively affect this endpoint. And that endpoint is this prediction of future health. And so what group of actions is actually making that prediction that you're going to be healthier in the future? Oh, those are good actions for that type of person. Let's tell everybody that that's type of person. Hey, you should do this and then see if it works.
Speaker E: Right?
Speaker B: Yeah. There's a lot of magic behind the.
Speaker E: Scenes at the app that I definitely want to dive into, and we'll see.
Speaker B: If we can get some visuals up on the screen as well. But first, I still want to zoom out and just, like, can you give us, like, a high level, just mission statement for you and humanity? And also tell us just a little bit more about the is ought gap between the current institution of health and where you think it would be more optimum if the institution of health progressed towards.
Speaker D: Yep, I'm gonna. Sometimes I'm gonna stop you and ask for definitions of some of these things. Sure. Is ought is a.
Speaker B: Is odd gap. Yeah. So there is the is odd gap.
Speaker D: Is, like, currently have and have nots or the.
Speaker B: Currently. The world of medicine is one way.
Speaker D: Okay.
Speaker B: And I think you, as a result, ought to be some other way.
Speaker D: Cool.
Speaker B: So where is it and what it be? And then, like, your own personal mission statement with what you're building here at humanity.
Speaker D: Cool. Yeah. So we try to boil down our mission to real simplicity. We want to give a billion years of health back to humans by 2030.
Speaker B: That's funny. So at bankless. Sorry, just a slide. Cuts. We talk about we want to help a billion people go bankless. That's been our line.
Speaker D: A billion. It's a nice, round number.
Speaker B: Yeah, it's the largest number that humans can still reason about.
Speaker D: Exactly. But, I mean, so the consumer VPN that I was running, we got to 900 million users. So, like, the beauty of the Internet is that you can. You can actually affect that many people. Right. So that's the mission. We want to get back those billion years and very specifically healthy years.
Speaker B: Right.
Speaker D: So those fully functional, or, you know, mostly fully functional, healthy years that we, almost all of us enjoy living. And so that's. That's the mission and kind of what we have right now. I would say the biggest thing, if I was to cut straight to the chase, we don't measure our health. So those that are real seekers and kind of out there, like early adopters trying to biohack and all that kind of stuff, very few of them even measure their health. They're content consumers, and there's a lot of great content out there, and you're trying different things, but you're not monitoring yourself over time to see if the group of things that you're doing is actually moving you towards what most of us are going for, which is we want to be healthier for longer. Right. And so what the system doesn't have right now is measuring. The cool thing, though, is we've just entered maybe the last five years, I would say we've entered a time where almost everybody has a device on them which can monitor their health. So, like, our mainstay of monitoring your health as a user is your movement pattern. And so if you keep your phone in your pocket or near you throughout the day, we'll be able to at least give you some prediction of your future health. If you have a wearable, we can feed in the heart rate.
Speaker B: Right. There's so many out there, and you're saying that even the very blunt tool of your phone in your pocket still can provide sufficient data. But then we're many years into the Fitbit revolution. Now we've got the apple watches, there's the whoop band, there's the oura ring, there's the sleep eight mattress that actually tracks you when you're sleeping. So there's even other non wearable devices that are inputting data. And so I think this is the industry, the sector that humanity is really tapping into. There's all this data out there, and now we can start to apply it.
Speaker D: Exactly. And apply it in, almost leave the kind of clinical trial system and these kind of, like, small research study things behind in a way, because the secondary problem. So that was the big is odd. I. The secondary problem is that we. Because I think in the past, it was easier to have one common marketing message. So if you're a public health official, and it makes sense, like, if you're a public health official, like, you can't go out and say, hey, there's 20 instructions to follow, you just say, hey, sugar is bad. And you just hope that message can get as far as it will go. Right? And the problem is, each of us is different, and the combination of actions actually affects the outcome. And so to give you an example, so, like, if you're a certain type of person, which I am, lucky for me, don't be jealous. I can eat as much chocolate cake. It's not going to spike my sugar. I've experimented multiple times with continuous glucose monitors. It just won't do it. I'll have to try a couple more times to make sure that the results are sound. But the normal person would say, hey, you know, that's unhealthy. You shouldn't be doing that. Right. And so each of us needs to start measuring ourselves so that we can actually know, you know, what's actually working and what's not. And so the getting away from, you know, how do you, how do you get away from that real need, which was these simple messages, to actually applying a much more personalized thing. Unfortunately, I don't. I don't want to say everything's in your pocket, but, like, again, we have this great delivery service that we unfortunately stare at a lot during the day.
Speaker B: Right.
Speaker D: And that thing can give you very personalized stuff. And we. And we actually are unhappy when it, you know, we talk about how personalized it is. These. These ads know too much about me.
Speaker B: Right.
Speaker D: That same thing. Knowing everything about you can give you specific health knowledge. Right. So.
Speaker B: Right. We can start to turn it around and start to use it for our benefit. And I think this is where just perhaps the AI conversation starts. Because the idea is when so many people have their phones, everyone has their phones. And even not as many people have phones as has wearables, but a large number of people have wearables. We have a ton.
Speaker D: And in their closet, too. And now they're taking them out to use with the app.
Speaker B: Right, right. Yeah. So, like, I've had an Apple Watch and I've loosely worn it here and there, but since coming here to Zoozalu, doing the morning cold plunges, going on the runs, having this app to actually tell me what's going on with that data, all of a sudden, my rigorousness about how frequently I'm wearing this wearable has, like, tripled.
Speaker D: I'm going to cut this part of it. I'm just going to send that to Apple. Feature us more, please.
Speaker B: Yeah. Official iOS partnership app.
Speaker D: Right? We're selling watches.
Speaker B: Okay. So we have all this data. We have way more data than I think we know what to do with. And maybe that's even, like, one of the big problem statements. We have more data than we actually can apply to actually know what to do with it when it comes to our health behavior. But I think this is where perhaps the AI conversation starts and just the big data conversation starts. So can you walk us down that rabbit hole?
Speaker D: Yeah. And I think what you're touching on is also, I think we've had a few years. I think all this stuff just puts it in the mainstream. So I don't think any of this was, like, a wrong direction, but we had a few years of like, hey, we have a bunch of data on you. And here's some cool graphs, right? And, and I think people started to get a little bit of like, okay, well, that was cool. But graphic didn't really help me. And they're like, okay, you know, you can only look at your HRV for so long, and then you're just like, I get it.
Speaker B: That's your heart rate variability, by the way.
Speaker D: And so I think, sorry, what was the question?
Speaker B: So now that we have other tools, not just wearables, not just data, but we have AI and lots of data, what can we do with this?
Speaker D: Yeah, so I think that's where the cool, the cool thing that happens is when you have enough of this data, no matter if data is noisy, no matter if data is interspersed, and you're not getting it every day. The great thing about AI is it actually can just still, it looks at all that data and it can find all the relationships. I mean, deep learning is about finding the kind of relationships and associations between multivariate systems and just taking that as knowledge. And then in many cases, it's applied in different ways. You can fold all the proteins in the world, or you can, you know, it's that deep knowledge, that deep learning that gets that knowledge of how these variables are associated. And so that's, that's kind of the, the byproduct of deep learning. But it's actually the beautiful thing that now we're applying to health, which is you can take that knowledge and actually then feed it that you can create an algorithm off of that and then feed it back to the user so they can feed in their data and it will actually give them guidance. The other thing, I don't know if we want to get too deep into it, but we can go off on a tangent. I think what we're working on also at humanity and kind of plays into the public good that's happening behind us. A lot of great talks is we want to find a way, and there is a way we want to actually build out a proof of concept of that. You can keep the data private, stays in the same place. The owner of the data is the user. When they want to delete it off of that place that they put it, they have full reign to do that. But you're also able to, the route word going is creating synthetic data. So we're basically deep learning, learns all the relationships, and then you create fake users that have all those relationships. You can open up that data to, then train models on, and everybody can do that. We have 150,000 users, but there's health tech companies that have 100 million users, and there's the NHS in the UK and Montenegro health system. They can keep their user and patient data completely private, not anonymize and send it off somewhere, keep it where it is completely private, but still train models on it and allow researchers to do that, allow anybody to do it. I think that's now that we have blood in the water with AI, with chat, GPT just made it completely mainstream right now, people are like, what else can we do with data? So I think now is the time really that we need to push to open up this data, but not the real user data, but actually just the learnings from it.
Speaker B: Right. And I really want to drill down on kind of the problem of all this data that we have for anyone with a wearable or an apple will. I'm sure Android has this, too. But if you open up your health app, you can just scroll and scroll and scroll through all the things that it's measuring.
Speaker D: Right, sorry, Apple Health app.
Speaker B: Yeah, they're taking some credit back. So, like, it measures things like your gait imbalance, right? Which leg that you put more weight on while you walk. It measures heart rate variability, which is very useful. It measures how many steps you take. Like, it measures everything that it can get its hands on. And what the cool thing about the humanity app is that it actually just boils that down into just like, are you living longer or are you living shorter? And that's really the power of applying AI here. Just consolidate all that data down to one output. But hopefully, can you just illustrate the magnitude of the data problem and how AI fixes that?
Speaker D: Yeah, and I don't want to be overly positive, but there is the other side of the coin, which is basically these platforms, these hardware companies, or if they consider themselves hardware companies, their sensors are getting more and more prevalent. More and more people have them and they're getting better and better and they're monitoring more and more stuff. The secondary that you touched on is they're also. And I think this is a place for them to play as well. They're also creating processed values like, you know, gait instability, whereas I might not have a person on my team that necessarily is going to be an expert in that.
Speaker B: Right.
Speaker D: But they create this process, this feature that they take out of that data that then we could use. Like, I mean, if you're a sports app, you get gait instability. You're like, okay, you're injured. Here's go out on this route. Right? And I think that's so there's an ecosystem where there's easier and easier ways and cheaper and cheaper ways for people to get this data on themselves. Then there's an ecosystem that can live within the same player where it's like, how can we process this? So you're not just sending raw accelerometer and gyroscope data, and then you got to just learn how to deal with it, because it just makes it easier for people like us, like humanity, to then take that, and then we're like, hey, well, actually, now that it's processed, it's a few kilobytes, and we now can do this very cool thing with it.
Speaker B: Okay, so when you apply AI to all of this data, can we just unpack a little bit more, like, what that means? Because in this day and age, we just say, and then we do AI.
Speaker D: That's where the variable reward is. You press the button and magic happens. What magic am I going to see next?
Speaker B: So how does humanity actually apply AI to produce meaningful results for its users?
Speaker D: Yep. And so we're doing something, I wouldn't say simple, but it's, you know, it's kind of like really good machine learning on it, which is basically saying, what? So we see all these actions that people are taking, and so we take all these values from your. From your wearables, from your. From your phone. We take values from your input. You can say what your mood is. If you're not monitoring your sleep, you can put in your sleep. So we take some manual input. And so then we just take these all as these are actions. We're not making any judgment on them whatsoever. So person did these actions, and then we have this prediction that happens every single day using your movement pattern and your heart rate pattern, if you haven't. And we're just saying, okay, this prediction, is it predicting that you're going to be healthier in the future slightly or less healthy in the future? Slightly. So that's the end point. So you can think of it in traffic navigation. The endpoint they're trying to look at is time to destination, right? So they're saying this path takes five minutes, that path takes seven. So we're basically saying, okay, is the prediction of your future health better or worse? And then if it was better, you almost just, like, label back into that group of actions and say, this seems to be a generally net positive group of actions for this type of person.
Speaker B: Right?
Speaker D: So that's where the stratification is the most important. And so you just do that repetitively every single day, and you can take time periods, you can start to understand if there's a lag between an intervention. So maybe if you do a cold plunge, we'll see your prediction of health better in three days and not immediately.
Speaker B: Right.
Speaker D: So you start to learn all this from the data, but in the end, it's, it really is exactly like traffic navigation. You're not trying to do a study on side streets versus highways. You're basically like, hey, these cars were heading on all these different paths and all these edges put together made a better end point. Got you there in five minutes instead of seven.
Speaker B: So, okay, so let's talk about some of the healthy behaviors that do actually move the needle. So we have all of this data, but most of the data is probably trash. And some of the data is probably really, really good.
Speaker D: You always got to be careful with that because I think the genetic space played this out very well. They're like, these, these are gene encoding areas, and then there's a bunch of randomness that, you know, throw that stuff out and they're like, oh, shit. That actually.
Speaker B: Well, no, no, some data is going to be more valuable than other data.
Speaker D: Correct?
Speaker B: In the moment, yes, in the moment. Right. And so, like, maybe we could actually just, like, zoom out and put humanity aside. We'll, we'll open up the app in a second. But just like, talk about what are the big behaviors that do move the needle for people that can actually be measured by things like our fitness trackers. But just like, what are healthy behaviors that most people aren't engaging in that the humanity app or just longevity efforts would tell a user to engage with?
Speaker D: Yep. So I think there's a bunch and you need to be monitoring yourself and the combination is important, so I'll keep repeating that throughout. But the combination and the one that probably people, people could understand that they're different, and so they pick up on that quicker. Like, okay, it works for him, but it's not going to work for me. But the combination of things makes a big difference. And that means, like, if you didn't get enough sleep, I'm just making this up. Like, if you didn't get enough sleep and then you do a high intensity workout, it might actually be net negative. If you got a bunch of sleep and you did high intensity, it might be very positive. Right.
Speaker B: And probably also the order of those two things also matters, right?
Speaker D: Yep. Yeah, exactly. Timing. Yeah, yeah. So order, you can see so well with nutrition, right? So if you talk about, like, blood glucose and it's spiking, and that. That being generally bad, at least, you know, according to what we know so far, is, you know, if you go for a walk, this is why some of these common public health messages get out there is because for a lot of people, if you go for a walk right after a meal, it drastically lowers any spike that you might have gotten, no matter what you ate. I really. And so, yeah, the order and the combination. And so what we. Sorry, I lost a question again, just.
Speaker B: What are the behaviors that really move the needle?
Speaker D: Yeah, people always want to know. So I'll list a few, I think, from our data. I'll give it straight from our data we're actually seeing for some cohorts, and you may or may not be in that cohort, so please don't take this as a prescriptive thing. In some cohorts, moderate intensity activity seems very. Has very little impact, really. And so this kind of, like, stay in the, you know, just go for a moderate jog for some people. From our data so far, with 150,000 people, seems to not really move the needle. And so. But the low, low intensity, which basically means, like, walking around and possibly kind of low intensity yoga, that sort of stuff, is fairly impactful for most of the strata, and high intensity has quite a variation in its impact. But is every single strata needs some of it interesting, but the moderate for, I'd say, like, we have pretty large, different cohorts now, but, like, let's say three of those cohorts, I could just see the data in my head right now. Like, three of them, it just doesn't show up.
Speaker B: So. And you said moderate activity doesn't move the needle. Can you define what that means to move the needle? Like, what does that mean?
Speaker D: Is it impactful, in this multivariate analysis, this group of actions that people are taking in a day, does it seem to be impactful on reducing their probability of future disease to that endpoint that we're trying to affect?
Speaker B: And that's just based on scientific study that we are looking at from indicators of health?
Speaker D: No, that's directly from the data, is saying, yeah, our prediction that we're getting from your movement pattern and your heart rate pattern on a daily basis, that prediction doesn't seem to be moved, impacted by, in a couple of the strata, by the moderate. And so that means. And you don't want to then immediately take that and then jump to like, oh, well, that's. That's what we thought we'd find.
Speaker B: Right.
Speaker D: But it is exciting now, I'll admit it. It's quite exciting to, like, see the real data now, because I think in the longevity space, it makes a lot of sense. But again, this is my own personal Michael Gere kind of conjecture is there are certain things that you trigger in the body when you do high intensity exercise, one of the main ones being a hypoxic response, meaning your cells think they're running out of oxygen, and so they start cleaning up, breaking down misfolded proteins, like doing good stuff that makes your body stronger. Just getting nutrients to your body throughout the day, which is you walking around and not sitting at your desk, at your computer, is, you know, it makes a lot of sense from a systems like physics. You know, you're going to more consistently deliver more nutrients to all your cells throughout the day if you're moving and your heart's pumping it just slightly higher and getting the blood around right. And so that makes sense. Moderate, you can't necessarily find a lot of kind of in the very high level longevity kind of framework be like, oh, it doesn't really do much more than the low intensity because, yes, you're moving blood around, but you probably were moving blood around fine, walking around, and it's not reaching the point where you're getting a hypoxic response, where your cells think they're running out of oxygen. And so it makes sense that for some strata, you don't see much going on.
Speaker B: Right. And we would only really be able to know this if we have a lot of data and the models to be able to create these relationships, assuming. Right?
Speaker C: Yep.
Speaker B: And so, like, one question I have is, like, going back to, like, what does it mean to move the needle? There's one. There's one perhaps way of making a health app, which is going through all of the literature and all of the studies and all the doctors that say, hey, this is good.
Speaker D: It's kind of the way that most health apps are made, actually.
Speaker B: Right. And then there's another way of doing this, which is just giving models, AI models, a bunch of data, and then also cross referencing that with how healthy these people are, how long they live, and actually have the AI models determine what is good.
Speaker D: Exactly.
Speaker B: And then that sounds like kind of more of the approach of the humanity app, where you don't want to actually have, like, inputs as to what is good or what is bad. You just want to input a bunch of data and have all of that data create relationships with itself, and then have the AI model say, like, oh, well, this person's doing all these things and they are living longer and being more healthy, but not actually telling the models what health is. Allowing the models to determine what health is.
Speaker D: Yeah.