Oh, I think we're live. Hi, Max. Welcome to Venture with Grace.
Hey, Grace. Excited to be here.
To start the show, I would love for the audience to get to know you a little bit more. So, you were originally from Australia and then you moved to the US for your own startup. You went to school in Berkeley. And why don't we talk about what were some core lessons that you learned early on in your career that kind of shape you into who you are?
Oh gosh, I love that. I think the key thing is that everyone can take more risk and listen less to the dogma of the world. I think the dogma of the world is: here's what you should do, here are the steps, here's the pathway. And I think most pathways are actually false. I learned that quite young. When I was 10 to 12 years old, I used to play Minecraft, and I played on this large 20,000-person server, and I became the richest person on that 20,000-person server. It was my brother and I finding all hacky ways to do that. And in that world, it was very permissionless because it was a game. It was a game that modeled reality. Because it was a game, we weren't thinking, 'Why do we have the right to do this?' We were just like, 'No, we're just going to do it.' And I actually think that the game mapped to reality, which is: if you can do that in a game, why can't you do that in reality? Right? We had done nothing before the game. We hadn't had experiences. We hadn't gone to school. We just did the thing, right? We built businesses that were in a fantastical world, but in a very similar way to how you'd build businesses in the real world, right? There was a microeconomy there.
So I think the first thing is you can do more than you think. You can take more risk, and you don't have to internalize the dogma of everything around you. I think the second thing is that most of what's around you is a local optima. Most of the world is maybe like a four out of 10. And the implication when you realize this is that by copying what's out there, you can only get to maybe a four out of 10 or a five out of 10. So the conclusion is we cannot copy. To actually get to eight or nine out of 10, to really push the needle and push the frontier, you need to start from scratch. And I love that because it's a very empowering belief. It means actually copying is never going to get us there. We have to think for ourselves and start from scratch to truly create something better than anything else out there.
For sure. I wonder, so since I think from your career you obviously seem very entrepreneurial, you co-founded venture funds or a club for younger people to improve the collective productivity or just to learn from other young people. So why don't we talk about, for those different ventures that you have started, what are some of the biggest lessons that you've learned through your entrepreneurial journey that you kind of implement into Superpower?
Sales and learning how to sell is probably the most important, because when you're starting anything, there are so many people you have to sell to. You have to create the narrative of what you're working on. You need to inspire people. You need to communicate the mission and the vision. You need to get people excited. You need to sell to partners. You need to sell to VCs. You need to sell to founders if you are a VC. And every single one of those experiences put me in a position where I was learning. I had to do those things. I had to learn to sell. And I never really thought of it as selling. I thought of it maybe as storytelling or inspiring or motivating. But I think at the core they're synonyms, the word 'sell'.
If I look at Next Chapter, which is the global founders community, another big part of that was learning to identify brilliant people very early. We were on a mission to bring together the most talented people around the world into one place and do that from when they were 18 years old, right? And that meant identifying talent really early. And that's important because now a big part of my job is hiring people, finding the most talented people and identifying them really early before they're proven. And then if I look at Ultraviolet, the first venture fund I started, what it basically meant is from age 17 I was constantly assessing businesses, constantly meeting founders, constantly hearing about how they think about distribution, how they think about product, constantly seeing the business models and the strategic bets they were making, constantly selling to those founders to give me allocation into the round. And that hones your pattern matching, it hones your analytical ability, it hones your understanding of markets and narrative and storytelling, particularly to do that from a reasonably young age.
I wonder. So I think beyond sales, I think you're such a great marketer, from the branding of the website, of your own website, and then also how you kind of shape your company as a brand. Why don't we talk about — maybe we could first start with what is the core of Superpower? Obviously you guys are trying to basically share what makes ultra-high-net-worth people productive or predictive of their health factors, basically making that technology and the process accessible to a normal human. Why don't we talk about, at the beginning, how did you kind of create the first version of the product to what it looks like now? How did you — since you talked about sales, and I thought the marketing was also super great — how do you kind of come up with the marketing strategy behind how you are going to distribute your own product, and then coming up with the message and everything?
The first version of the product and the first version of the marketing started with a set of beliefs. I think most great things stem from a belief system at the foundation. The belief system for us was a couple of folds. The first belief is that there was a huge gap between the best of healthcare and what most people have access to, and I believe that for as long as that gap exists, someone or some company has to come along and fundamentally close it. The second belief is that the only way to close that gap is to use technology. It's going to be impossible for humans to all of a sudden close the gap. It's going to be impossible to train doctors. Instead, we need to use technology and AI to close that gap. It's going to be AI that can deliver high-quality care to everyone, not humans. The third belief is that healthcare should be preventative but also performance-enhancing. Healthcare should enhance human capability. And then probably the final belief is the way healthcare was done today, from both a product perspective and a branding perspective, was completely wrong. It was a local optima. It was three or four out of 10, right? There was so much — it could be so much better.
And if you look at the combination of these beliefs, that is our product and it is our marketing. We stand for something aspirational, and we stand for human performance and prevention as a consumer brand. And you see that in the name Superpower. You see that in the visual identity, using orange where every other company uses blue. You see that in the website and the distinctiveness there. You see that in the language we choose. It comes through in the product. The fundamental starting point was: how do we democratize high-end concierge medicine? You see that in the way that we've continued to bring the price point down. We started at $10,000. Today we're $499 a year. We want to continue going cheaper. You see that in the bets we make, the people we hire. We have hired technologists a lot of the time rather than doctors, and some people will critique that. They'll be like, 'What do technologists know about healthcare?' But the thing is, healthcare today is not being delivered right. It's established that healthcare is not working, and it's also established that we've tried to solve that by just hiring doctors and throwing them at the problem, and it hasn't worked. And if we go back to that fundamental belief, it is that technology is required to democratize high-quality care. And as a result, it changes how we hire. So if I look at our product and our marketing, it fundamentally starts with these beliefs that we have that others might not have, that allow us to make decisions that are unique and distinctive and opinionated, and that's the starting point.
I wonder, since I feel like a lot of people tried to do these kind of testing projects, but I guess not many people have successfully continuously sold the second act of the product, or product two, you know, for health measurement. There were companies in the past that were maybe easy to sell to the customer once, but later on, after people got certain results, they didn't find it very sticky — it's really hard for people to continuously purchase a product. So after, let's say, they got the first test, what are some ways that you kind of expand into multiple different products down the line? And obviously right now you guys can test a lot of different sub-sectors that you measure, but how do you think about the longevity of a product line?
We don't think about that. We think about how do we solve problems for our members. Because the second we solve problems for our members, the second we deliver health outcomes, they stay with us. After someone has done their first blood test, their response is, 'There's so much stuff wrong. Help me, right? Fix this stuff for me.' That's their response. So our job then is to help someone. It's to fix their problems, to help them achieve the health goals they have. And if we do that effectively, cost effectively, someone comes back to us. So what we fundamentally think about is how do we solve problems that our members have time and time and time again. And if we do that, then we earn the right to exist and continue to be around and have people come back to us.
I wonder, so for people to measure these different indicators, I think you mentioned this on a different podcast, that compared to traditional testing, what you guys have is various different data coming from different segments. So for example, you wear a ring, or you wear an Apple Watch or something, and then you collectively combine the data with your traditional health data. And I wonder, is there any surprise finding that you kind of found from combining these different metrics, and how do you evaluate it to create a cohesive health message for people?
So the problem with how healthcare has always been performed is it looks at siloed data streams. What's possible in a technology-led paradigm is you can look at all of that data in concert, and you can start to map between blood and gut microbiome and toxins and sleep and what someone's reporting. And we see that a lot — we will see, okay, yeah, your sleep is disturbed, but that kind of shows up in high cortisol, or that shows up in a mold allergy or high exposure to mold in your room at night, or it shows up in a high respiratory rate in your room, which suggests maybe you're not getting enough oxygen and there's high carbon dioxide in the air. So the idea is when we look across all of these data points, we start to be able to draw connections that allow us to get to the root of what is going on. And that's the fundamental goal: how do we get to the root of what is going on, because then we can take remedial steps.
I wonder, since this is such a strong biotech and data problem to solve, how do you construct a team that's connecting both biotech and data? And then since we briefly chatted about talent, you used to run this cohort of talented young people — how do you identify the undervalued talent? And how do you first assemble your own core team to build the initial set of product, and then from the product angle, how do you hire the best talent for something like this?
We assembled the initial team with some people we know, right? It's really important a lot of the time to start with bringing on the best people you know and working with them. You have shared context. It's derisked. Most founders should know a handful of talented people. And then from there we try to find the best people in the world to solve a problem. We say: here's the problem, here's the set of things that we're looking for, here's a very clear understanding and definition of the kind of person we're looking for. And then we say, okay, let's find the top five people in the world and let's hire one of them.
Mhm. Since you're aggregating all these different medical records and blood work and wearables and even genetics, what kind of infrastructure or UX choices do you feel are harder to get right?
What do you mean by that?
So since I feel like you're collecting so many different metrics, how do you figure out the connectivity between them? And then also for your own company, internally, how do you decide which factor we should value more? Should it be my Apple Watch or should it be my medical record? And how do you bridge the connectivity to this, since you mentioned maybe your sleep is bad because there's something in your air? Is there an actual doctor looking at everybody's current metrics and then telling people how it works, or how do you evaluate what metrics should be valued more than others?
The point of what we're doing is to not build algorithms, to not manually put weights on things. The problem with doing that is that it's very non-scalable, and it's very hard to programmatically define the weights that you apply to different things. Lots of people have tried to create algorithm-based medicine, decision tree medicine. That's been a problem for 60 years. No one has really done it effectively. The key unlock of a large language model is you do not have to do that. The key unlock is the large language model actually, somewhere in its magic, has a sense of how to weight things and arrive at the correct conclusion. Now we validate any AI output with humans. But again the catch is where these models are getting to: they are going to far exceed what is possible for humans, and we're not going to fully understand the way they weight everything. We'll be able to probably see their reasoning, and a clinician will be able to validate the output, but we're not going to understand the exact algorithm that is used to reach the final conclusion. And that's kind of the point. So we don't weight things. Instead, this is the magic of AI.
Do you want to showcase your product? Since I feel like maybe a lot of people have heard about it but they haven't become a subscriber yet, because they may not have seen the product itself. I don't know if you want to show a demo. We have a present button here. Let me know if you want to.
Okay. Share screen. Cool. Can you see my screen?
Yeah. So there are a few parts of what we do today. The idea is: how do we create the best health membership in the world, right? And there are a few parts to our health membership today. One is we start the year with a 100-plus blood biomarker test. So we'll send a nurse to someone's home and they'll collect over 100-plus blood biomarkers, and you can add on any of this additional testing as well: gut microbiome analysis, environmental toxin analysis, Grail or cancer screening, and we'll pull all that data into a page that looks a little bit like this. And we'll visualize someone's results over time and show them how they track and trend. And we'll also pull in someone's past medical records, and then we'll finally integrate wearables. And it's important to have all of these different data points, because in a world where the AI delivers care, this defines full context on the person. Google doesn't have full context, Apple doesn't have full context, ChatGPT doesn't have full context. Our goal is to have that context on someone.
The second part of what we do is, now that we have this data, how do we connect the dots to get to the root of what is going on? The unfortunate reality is most people have a 10-minute doctor consult a couple of times a year with the doctor. That's it. 10 minutes. The doctor cannot look across thousands of data points to get to the root of what is going on. The AI we've built says, 'Here's thousands of data points. Let's actually connect the dots across them to see exactly what this person should do.' And a lot of people get this report, and their response is, 'Wow, this is game-changing. No one has told me this. This might even fix a problem I have.'
And then the final part is rather than just saying to someone, 'Good luck, go and fix the problem' — that's what the doctor's office does — we will instead say, 'Let's fix it with you. Let's help you. Let's show you the exact next step.' So follow-up testing can be done in one place. At any given point in time, if you need to, we'll show you here's the exact next step. You should never have to think about the exact next step. For example, I've got a blood test booked later this week, and then I've got the option to book in for an advisory call as well.
And then the final piece is: let's make it really easy to keep taking action. If you need access to a supplement, a clean home item, a healthy grocery item, a pharmaceutical, we've brought all of our favorite ones into one place, 20% cheaper than Amazon, and we show someone: here's why we chose the product, here's specifically why we chose this one, this is why we think it's the best in the world. And here are our curation standards, the levels of evidence we vet products against, and all the different databases we cross-check — ConsumerLab, Examine.com, Labdoor, etc.
And then the final part is: if you have a health question, you have two options. One is you can send a message to your AI, or two is you can ask your care team, right? Which means pulling out your phone and sending a text message to them, and you can also kick that off here. So I'll send, and then my medical team of three people, my doctor and my longevity clinician, will get this message and they'll respond and tell me what to do. So that's the membership today, which is: test everything, 100-plus labs twice a year, all of your data in one place — medical records, wearables, a protocol that connects the dots — and then access to care, access to your care team, access to the ecosystem, access to products, access to follow-up testing. And all of that today is 42 a month, or $499 a year.
I wonder, so when you first designed it — I guess when I see there is a longevity coach and then there is a primary doctor and then there is another person — how does it come with my health network? So for example, is the doctor my current primary doctor? And then if they work at like a Kaiser or something, are they connected to Superhuman — I mean Superpower? I'm so sorry, I feel like the email system is very much on my mind. But how do you think about integrating Superpower into the existing health network, and how does it work? Do you guys work with pharmaceutical companies to promote their medicine on top of it? Or do you work with a specific healthcare group to integrate their doctors onto the platform? How does it work from a business function?
Right now we do not do full-stack primary care. We recommend someone still have a primary care doctor. That said, 60% of our members don't, because they realize that primary care doctors are a little bit stuck in the past, don't have much time, are hard to get into, and are expensive. In time we might integrate more with the system. We're building out the ability to refer out to the system. But otherwise the experience is ours, our owned experience. We don't act as a marketing function for pharmaceuticals. We don't act as a marketing function for supplements. We hardly make money from either. The goal is to just reduce prices of those things, right? If someone's already buying them, we say, 'How do we put you on the best one, the highest quality one?' because there's a lot of garbage out there. And how do we put you on the lowest cost one? Because there's a lot of price markups out there. And that means not working with pharma companies. It means working without them, because there's a lot of perverse incentives in drugs and drug pricing in the US.
Is my entire result purely generated by AI, or how does it work? So for example, can you walk me through — let's say I do my... So basically when I go on your website, I order the kit, and then you send me a kit, and then I poke my own finger, I put my blood in somewhere and I ship it out, and then after I ship it out I just wait and then get a result from the website, and then basically like what you showed me, it would just tell me what I need to do from that point. So basically scheduling an appointment with someone, or take these supplements and stuff like that. Is that correct from the whole process?
The point of the AI is to do what very few human doctors can do. So we work with a lot of the best doctors around the country, and if you want to get in to see them, it's going to cost you 100-plus thousand dollars a year. That's how much it's going to cost. Our job is: how do we provide that quality of care for a couple hundred dollars, right? That's the goal. And the AI, the point of it is not efficiency or cost reduction. It maybe does that a little bit. The point of it is: how do you deliver the highest quality care? How do you do that? Just obsessing about quality, and that's the point of the AI. Now, your medical team is always going to look over it. But the idea is your medical team looks over it and they're like, 'This is amazing. This is so much better than I could have done. This has drawn conclusions that are true. I understand them. I understand the mechanism why they're true, but I maybe wouldn't have thought about them,' particularly given how little time doctors have. So that's the point of the AI.
I wonder, how do you tailor that to individuals? Since I feel like one of the key reasons why these doctors may be charging that much is because they're working with you specifically, or working with another, you know.
Yeah, but they're not that personalized. The challenge is they have limited time and limited cognition and limited storage. The benefit of an AI is that it knows everything about you. We give you surveys. It remembers every one of those data points. It remembers everything you've ordered. It has in its head any data point on you — your personal history, your family history, all of your past medical records. Some people have like 5,000 pages of medical records. Think your doctor's going to spend time reading through 5,000? No, they don't. So the implication is that these systems can be highly personalized, right? Way more personalized than humans using templates or applying standard checklist-based medicine can be.
I wonder, so what goes into these different algorithms? Since you mentioned some people have 5,000 pages of medical records, I'm sure not the whole 5,000 pages are —
Everything goes in. Everything goes in. More data and more context is always more helpful.
Okay, so is there any particular — let's say when you are shaping your data team or the AI team — what are some primary benchmarks that are the most important for them? Basically, is there any data that they have to have for this to —
Well, data that's more recent tends to be more helpful. Data on what is happening for you — your current medications, your current symptoms — is helpful. Blood biomarker data is very helpful, particularly because we're measuring 100-plus blood tests and most people have never done a blood test that comprehensive. And that allows us to discover all sorts of things that someone might not have seen before. I discovered I had pre-diabetes two years ago and I've since reversed that. I discovered I almost had mercury toxicity two years ago and I've since reversed that. So the blood test data I think is particularly helpful.
I have another question from a business angle. So since you mentioned the pricing is going down, why should people purchase now versus, you know, maybe two months from now, maybe it becomes $300 since there's more people buying it?
Well, that's kind of like — people want to be early adopters, right? Tesla in 5 years is going to be less than Tesla today. Tesla next year is going to be less than Tesla today, but people can wait around if they want. That's fine. There'll always be the people who choose not to buy the Tesla today because they want to wait a few years, and then there are people who say, 'Well, this is $499 today, this should be like 5x the price. I want access to testing and healthcare, so I'm going to buy it today.' So, personal choice. And I think everyone should and can make their own choice.
I wonder, when you're testing the different 100 biomarkers, why is a regular hospital not doing it? Is there like — yeah, I think also, is it the same amount of blood that you need for these 100, versus do you need less or more blood for this test?
Hospitals aren't doing it because insurance isn't paying for it. Insurance isn't paying for it because they don't care about prevention. Insurance doesn't make money if you prevent disease. They make money if you get disease. Insurance also takes a very short-term time horizon because you typically are with your insurer for only a few years. Many of these biomarkers are useful for prevention and early intervention, right? The ones that are reimbursed at the hospital are the ones that show: are you sick right now? Is there a problem right now? And more blood is needed for more biomarkers, right? We will draw several vials, whereas if you go to the doctor they might just draw two vials.
I wonder, how do you think about the —
Future of health — obviously, I think this is like such a great product for the people who are like the early health adopters. But I think, like, to challenge the entire healthcare business model, that takes a lot of like guts and also like strategies to do so. Like, what are some ways that you feel like the future of healthcare is going to work? Do you feel like the current existing system of like the insurance — you know, you pay insurance or your company pay insurance, pay these like big healthcare providers — or like how will it kind of shape out in the future? And I guess, like, how do you think about like — I wouldn't say challenge, but I guess like work with the current existing healthcare system, because it seems like it's been there for a long time.
Uh, we like, we want to be a new front door. The front door today is not someone's doctor; it's Google. There are a billion Google searches every day related to health. 10% of ChatGPT volume is health. They're really bad front doors, right? We want a better starting place, a better way for people to explore and understand health. We want people to be able to access the things that doctors are telling them, no, you cannot access. And that's the front door. Now, we're not going to replace hospitals. They still have a place. Hopefully people don't need to go to hospital if they're taking preventative steps. But we want to be the first port of call, the first place people turn to, the place people go for prevention and performance, and then when someone is sick or has to enter the sick care system, that's when we refer out.
I wonder, is there any legal protocols around this? So for example, from collecting people's data perspective to, you know, like providing medical care, prescription or like suggestions — is there any kind of like protocols or like I guess like legal prevention or something like around this sector?
Uh yeah, there's lots of laws that try to ensure patient safety and privacy, and the important thing for us is we maximize patient safety and we maximize patient privacy.
Um, how do you kind of like do that? Is there like a, I guess, test that you guys have to go through, or like how does it kind of like work in practice?
Yeah, so safety — everything is reviewed by doctors and clinicians across every step. Any blood biomarker that comes back, if it's out of range, our medical team gets notified. If it's really out of range, we get notified in an urgent way, we escalate it immediately. On the privacy side, we're fully HIPAA compliant. We are audited by Vanta constantly. We have a security team. So we take that really seriously.
I wonder, so like when you started building out the first version of the product, like what is it? I guess at the beginning, who were the people who are like paying 10K versus like the people right now paying for like $499?
Initially, it was people who really cared about health and they wanted better preventative, performance-enhancing healthcare. They tended to have a little bit more money. They tended to already be following like a Peter Attia or an Andrew Huberman, and they wanted something better, right? They saw this product, they're like, 'This is my dream product. I've heard about concierge medicine or I work with a concierge doctor. This looks like it does all of that plus a little bit more with technology.' So they were the initial audience. Today, it's widely distributed across geographies, ages, income earners, etc. Slightly more skewed towards California, New York, as you'd expect, but otherwise very surprisingly evenly distributed.
From a distribution perspective, like how do you think of like getting the core user? Or like at the beginning, obviously like those are high-income people who are following like Peter Attia and Andrew Huberman. I guess at the beginning, where do you find this group of people? Do they all hang out in one place, or like do you go to, I don't know, Soho House or something to like distribute the product there? Or like at the beginning, how do you capture them versus like now, thinking about distribution?
Uh, we would just do a lot of viral marketing, largely on Twitter and LinkedIn, and our early customers would post about us. Lots of people would talk about us. We did viral marketing drops, and that helped us in the early days.
Um, what about like later on? And also like, how do you kind of like — what is a good conversion rate? From like, let's say you have like one post that went viral on Twitter or on LinkedIn, like what does I guess like the conversion kind of look like for you at the beginning? Um, conversion from what, like a post? How do you kind of track if let's say we went viral, and then like what is like a good conversion rate? Is that like, I don't know, like uh...
We don't really track it like right now. Like in the early days, the point is just do a lot of different things and make sure that the numbers are going up. We implicitly have an understanding of it because we'll see, oh, I made a post and now we have 3,000 more people on the waitlist. But we are not trying to perform attribution on a personal post or a post of someone in our network and waitlist or customer growth. We are starting to now because we're getting into the world of like paid marketing now, but that's a novel and new phenomena for us.
Um, when I guess like for the doctors who are on your platform, like do you integrate them as like full-time employees, or like are you working with a network of doctors to kind of like go through the product and then the result with them? How do you think about like from the supplier perspective, like you know, beyond the AI team there's like I guess like the doctor or like the human side of it?
