Fount and the Body's Magical Future
The Mad Scientist Running the Elon Musk Company Formula on Health & Performance
Welcome to the 1,357 newly Not Boring people who have joined us since last Monday! Join 108,068 smart, curious folks by subscribing here:
Today’s Not Boring is brought to you by… Composer
We all love Warren Buffet, but trying to “be greedy when others are fearful” is easier in theory than practice. In the market we’re in, removing emotion is key… but hard for us humans.
Thankfully, there’s Composer - a platform for creating and investing with rules-based trading strategies. It's basically algo-trading without Python or Excel. They have dozens of pre-made investing templates, but what's really cool is that you can edit those templates or build your own strategies from scratch, backtest them, and then invest.
You can edit their "Buy the Dips Nasdaq" template to only buy the top 20 holdings if they were up over 30% last quarter and rebalance every two weeks. Or, build your own strategy to execute paired switching trades between QQQ, USO, and Bitcoin ETFs - no code required.
Long time Not Boring long distance reader Michael Batnick recently described Composer perfectly: “Composer removes the subjectivity of buying breakouts and then having to make decisions during the heat of the moment…I’m super pumped to see a tool like this available for investors.”
I’m not only an investor in the company, but a long-time user of the product itself. Join me and sign up for free today*
You can click the link above ☝️ to read the full thing online 🎩
Fount is crafting the operating system for the human body, the way Elon would.
In February, Urban went on the excellent Lex Fridman Podcast to discuss many topics, including his optimistic views on AI:
Around 1:37:00, Lex asks Tim, “You’ve been thinking about this AI safety question, where today does your concern lie, for the near-term future and the long-term future?”
After acknowledging that humans aren’t that smart and are pretty easily manipulatable, which is scary, he said that he mostly gets excited about a lot of things, like having an AI assistant that can see into his body using wearables.
I was tempted to include the whole transcription of his answer, but I footnoted it here 1 and will use stick figures to explain instead:
In case it’s not crystal clear from that drawing, here’s the gist.
In the future, hardware like sensors in the bloodstream, infrared scanners, and Neuralinks will be able to measure what’s happening in our bodies in real-time, physically and even emotionally. We’ll all have our genomes sequenced and actually know how to use the information. We’ll be able to tell the AI our goals – more energy, lose weight, avoid cancer – and it will be able to do things like order “the perfect meal for that moment for me: for my mood, for my genome, for my blood contents” via drone, automatically.
Most humans see a scenario in which computers know more than we do and tell us what to do as dystopian. He thinks they’re wrong. He thinks it sounds incredible:
If we all had this, we would not look back and be like, “I wish I was making awful choices every day like I was in the past.” I think we’ll all be so much healthier. When we look back today, one of the things that’s going to look so primitive is the one-size-fits-all thing, like reading advice about keto. Each genome is going to have very specific, unique advice coming from AI.
I was running (NBD) when I heard that, and I nearly tripped. Tim Urban was laying out the precise thesis for a Not Boring portfolio company: Fount.
The last time we’d spoken before I listened to that podcast, Andrew Herr, the wizard CEO and founder of Fount, described practically the same end-state that Tim did: “What people really want is to quickly and easily know what they should eat, take, and do to look, feel, and perform how they want.”
Getting to that end-state has been the holy grail … or … fount ... of health and wellness for a long time. Nothing has come close to working at scale. One-size-fits-all solutions don’t work. Throwing a lot of poor quality data at the problem doesn’t work.
The problem is that knowing about what’s going on in the body doesn’t yet tell you what to do. The body is so complex that we haven’t come close to modeling it yet (see: The Laboratory for Complex Problems for the category of problem we’re dealing with). As Andrew explained, even the gut is more complex than a rocket, let alone the brain, and the interactions among all of the different pieces compound the complexity. Rocket scientists at least have physics to guide them; we haven’t figured these rules out yet for the body.
Gathering lots of poor quality data – observational studies, small clinical trials, even big trials that only look at one intervention – doesn’t capture the variability between humans or the synergies between different interventions. What’s needed is experimental data.
As we’ll dive into, the path to personalized health recommendations at scale starts with recruiting thousands of people, running dozens of the right kinds of experiments on each person over a period of months to years (but different ones for each person), and gathering data continuously throughout the process so you have rich pre- and post- data sets for each experiment. If most clinical data collection is 5 question multiple choice, this approach is War and Peace-length choose-your-own adventure.
Personalized health recommendations are the necessary first step towards the world Urban described, and towards healthier, happier people with less wasted healthcare spend, but collecting the data is so daunting that no one has even tried. Think about it:
How would you convince thousands of people to run dozens of experiments on themselves?
Who would design, manage, and update all of those experiments in real time?
How would you track data and ensure people were doing what you asked and telling the truth about what happened?
How much would it cost?
I mean, put yourself in the question: how much would someone need to pay you to run dozens of experiments on you?
$5,000 per month.
No, sorry, that’s not my answer.
$5,000 per month is how much clients pay Fount to run experiments on them and collect their data. Happily. (It drops to $1,000 per month after the three-month onboarding period.)
Because Fount’s data collection mechanism is a world-class concierge wellness and performance coaching program developed by the guy who led human performance and biotech strategy efforts for the US Military and carried out by ex-Special Ops Performance Advisors.
Fount flipped the equation on its head. Instead of applying for government grants to spend thousands per person recruiting participants and running studies on them, people are knocking down Fount’s doors and paying thousands of dollars for Fount to run experiments on them.
If that strategy sounds familiar, that’s because it’s the Tesla playbook. Make something that’s good for the world so sexy that people line up to pay up.
Like the Roadster, Fount’s price tag was a shock at first. It is … not approachable for 99.99% of people today. But it’s a best-in-class experience, and there are many people for whom the ROI is absurdly favorable.
LeBron spends over $1 million on his body and mind per year, and he’s had to put a team of people together to pull it off. Fount does that all for you, so it's not only best-in-class knowledge, it's made for real people who can't focus on their body as their full time job. Especially when it's made easy, there's no shortage of people in tech, Hollywood, and pro sports willing and able to pay $24k to improve theirs. If you’ve made a lot of money in crypto but feel burnt out, for example, Fount might be for you.
Shane Battier, an NCAA National Champion at Duke, NBA Champion with the Heat, NBA executive, and investor, who’s now a Fount client, told me why Fount makes sense for him:
I check on my bank statement, my kids’ grades, and even the oil levels in my car at fairly regular intervals. Why wouldn’t I check in on my own health outside my annual checkup with my doctor? Andrew Herr and Fount are the best way for me to ‘check in on’ and optimize my health in a more precise way.
The goal is much bigger than just serving the top 0.01%, though. Andrew could have done that on his own. Fount wants to help everyone improve their health and performance. To get there, the company plans to leverage the top end of the market’s willingness to pay for best-in-class service today in order to collect the data it needs to bring costs down to mass market levels over time.
When a company starts ultra high-end to fund early production before ramping production and lowering costs, people always compare it to Tesla, like I just did. Usually, the comparison is full of shit, like when any money-losing company cites Amazon’s years of losses as justification.
In this case, I don’t think that specific Tesla comparison goes far enough. Tim Urban's optimism on the Lex Fridman podcast isn't just about the technology - he believes that the way Elon Musk builds companies is key to upgrading the future. Fount is building in the same way.
Five years ago, on 4/20 (woah 🤯) 2017, Urban wrote one of his all-time greats, the six-part, 36,027 word Neuralink and the Brain’s Magical Future.
The essay is about Elon Musk’s brain-machine interface company, Neuralink, which Urban compares to a Wizard Hat. It spans an unbelievable range of complex topics – like how the brain works, how humanity compounds knowledge, and the framework behind all of Elon’s companies: the Elon Musk Company Formula.
Fount is running the Elon Musk Company Formula on health.
We’ll walk through each piece, but the result Fount is building towards is to dramatically increase the amount of human and financial capital going towards building a better future. It’s a non-obvious journey from a $5,000 per month Concierge plan to that goal, but that’s the point. If this were clear or easy, someone would have done it before. It’s a multi-trillion dollar opportunity that Andrew had just the right background to catch hiding in plain sight.
I don’t take Elon Musk comparisons lightly. Fount is an incredibly special company run by a one-of-a-kind leader, and I think it has a chance to reshape wellness and healthcare. Today, I’ll explain:
The Mad Scientist
The Elon Musk Company Formula
The Biomedical Data Gap
My Fount Experience
Increased Chance of a Good Future
To achieve its goal, Fount essentially needs to figure out how to put an enhanced Andrew Herr in everyone’s pocket. Tim Urban might call it Pocket Mad Scientist or Pocket Andrew.
“I’m good at this. I’m one of the best in the world,” Andrew told me, “but this is going to be way better than me. That’s so fucking cool.”
So let’s meet Andrew.
The Mad Scientist
There are a lot of things in the Fount plan that might raise red flags 🚩:
🚩 Building an “operating system for the human body” sounds like one of those ideas that hundreds of entrepreneurs have tried, and failed, to pull off.
🚩 Starting with a $5,000 per month product and a hands-on approach feels like no way to scale.
🚩 Trying to replace primary care doctors with ex-Special Ops and an app is crazy.
That’s why it’s important to get to know Andrew Herr first, before we get to the unbelievable stuff. Andrew is one of a very small handful of people in the entire world who could build this business.
My friend Henry Ault, the co-founder and President of Eco and the person who introduced me to Andrew, co-led Fount’s pre-seed with Founders Fund. He shared the investment memo that he wrote at the time, in which he described Andrew’s upbringing and education:
Andrew was born to a gastroenterologist father and a mother who worked for American Airlines and came from a decorated military family. Unsurprisingly, Andrew grew up fascinated by the best parts of medicine and international affairs…
When applying to college, there was only one program that combined both of these interests, while also offering daily access to opportunities for real-world experience: the Science, Technology, and International Affairs program—then an obscure major, but now the most popular — at Georgetown’s famed School of Foreign Service in Washington, DC…
Outside the classroom, Andrew worked with several US Government agencies on protecting the U.S. against terrorist and state threats by deploying advanced technologies, in the process winning a full scholarship from the U.S. Government for his research.
After his full undergrad scholarship, Andrew won another $200k scholarship from the government to continue his studies with three concurrent masters: Microbiology & Immunology, Health Physics, and Security Studies.
You can read the whole bio section from that memo here. The takeaway is this: Andrew was practically designed in a lab by the U.S. Government to build Fount and fix health and human performance at scale.
When the Pentagon’s Office of Net Assessment needed to hire a researcher “who had a deep knowledge of biology, was just as deep in national security, who thought creatively, and could also write well,” he was one of the few people in the country with the right mix of skills and knowledge. He was hired after one interview, and while there, “started his decade-long project to understand what processes in the body he could tap into to materially enhance performance and which tools he could use to do this.”
Over time, Andrew went from developing specific studies and programs to helping the Department of Defense leadership invest its multi-billion dollar R&D budget, with a focus on human performance, synthetic biology, and AI. All the while, he continued to work with Special Ops personnel – people for whom a .01% improvement in performance can be the difference between life and death - giving Andrew both the 30,000 foot view and the one inch view.
After working with the military for eight years, Andrew entered the private sector in 2017. His accolades and accomplishments during and since are absurd. More are listed in the doc, but a couple:
U.S. Army Mad Scientist distinction – twice
Adjunct professorship at Georgetown’s Security Studies Graduate Program
Johns Hopkins Emerging Leaders in Biosecurity Fellowship
Hired by the Miami Heat, USA Soccer, and billionaires as performance advisor
Wired described Andrew’s work as giving our soldiers mutant powers!
Let’s put things in perspective here. The US Military’s annual budget in 2021 was $778 billion. Behind Social Security and Health spending, it’s the largest budget in the US, and the largest discretionary budget bar none, likely in the world. And the US Government chose Andrew to lead cutting edge human performance and biotech efforts for that military, putting the performance of the world’s most expensive and advanced soldiers in his hands.
While the military is rigid in certain areas, it doesn’t subscribe to any orthodoxy when it comes to the performance of its soldiers. It’s the best-funded human performance lab in the world, and it offers an additional benefit beyond the money: soldiers who, unlike normal trial participants, want to comply with any programs or experiments that could enhance them.
Andrew’s experience improving human performance from within the government’s third-highest spend category gave him the skills necessary to try to dramatically improve results and lower costs in its highest: healthcare.
In 2019, he launched Fount to rethink health from the ground up. He’s raised money from funds including Founders Fund, Not Boring Capital, Allen & Co., and Elysian Park (the LA Dodgers’ fund), to tackle it. And – to use a very on-the-nose baseball pun – he’s in like the second inning.
Fount as it stands today is just a way to fund the really big goal.
The Elon Musk Company Formula
Elon Musk hasn’t touched healthcare yet, but I imagine that if he did, it’s unlikely he’d try to fix it from inside the healthcare system.
In the Neuralink post, Urban wrote about Musk’s new company thought process:
His initial thinking about a new company always starts on the right and works its way left. He decides that some specific change in the world will increase the likelihood of humanity having the best possible future.
He knows that large-scale world change happens quickest when the whole world—the Human Colossus—is working on it. And he knows that the Human Colossus will work toward a goal if (and only if) there’s an economic forcing function in place—if it’s a good business decision to spend resources innovating toward that goal.
Fixing healthcare from the individual level up fits the bill.
There’s certainly an economic forcing function in place. Healthcare in the US is a $4 trillion market. 18% of GDP is attributable to healthcare. That is not a good use of resources.
As we’ve covered a few times – when discussing Oscar and Cityblock, for example – one of the reasons that healthcare is so expensive is the misalignment of incentives. Insurers typically pay a certain amount for various treatments and procedures – fee-for-service – so providers are incentivized to perform as many services as possible. The system ends up spending most of the money when it’s too late – when people have already gotten chronic conditions or have ended up in the hospital.
Where we are today is the worst of both worlds: we spend more and people are less healthy than if we focused resources on making everyone healthier upfront.
Fixing that checks the second box: it will increase the likelihood of humanity having the best possible future.
The scenario that Tim Urban described earlier – sensors, Neuralinks, and individualized perfectly healthy meals – would obviously be amazing on the individual level, but it would be great at the societal level, too. The result would likely be healthier, happier, higher-performing people who can contribute more to society and a dramatic reduction in healthcare spend, which could go to more productive uses.
That’s the world that Fount wants to make happen.
Starting from the top right, Fount wants to increase the chance of a good future for humans, like Elon’s companies.
The big goal is to help humans upgrade simultaneously in two ways:
To be able to do more, to have more productive energy, focus, creativity, endurance, and beyond.
To have greater agency over their thinking and emotions – to better understand and be able to influence anxiety, stress, excitement, and joy – so they can channel that productivity.
Fount believes that optimizing humans in both these ways will help humans build an amazing future.
But to get to that future, Fount needs to light a “Match That Ignites an Industry” via a Company Innovation, which is funded by a Sustainable Business Model.
In every case, according to Urban, there’s something preventing all of that from happening:
Often, before a booming industry starts booming, it’s like a pile of logs—it has all the ingredients of a fire and it’s ready to go—but there’s no match. There’s some technological shortcoming that’s preventing the industry from taking off.
In the case of individualized healthcare, that shortcoming is the biomedical data gap. That’s the Starting Point.
The Biomedical Data Gap
Andrew loves clean data. He uses words like “exquisite” to describe data. Problem is, we don’t have enough exquisite data.
To get to the future that both Urban and Andrew describe – people quickly and easily know what they should eat, take, and do to look, feel, and perform how they want – requires highly accurate predictive models. The hard part isn’t the models themselves. We know how to build the right models. The challenge is the data.
According to Andrew, “all of the biomedical data sets available today, or that anyone is working on, have one or more fatal flaws that prevent us from building these models” (I’m just going to quote him since he’s the scientist):
They are observational, so they miss most of the important covariates and can’t ascribe causality. This works in marketing, but in health/wellness it’s a disaster. This approach is what gave us innovations like the ‘sugar is ok, but fat is evil' mantra that persisted for decades and gave millions of people diabetes.
The vast majority of experimental trials have tiny sample sizes. Anyone who’s run a clinical trial can tell you that n = 20 or 30 people isn’t going to tell you much. I ran one with 400+ subjects, and even that was too small.
Even in bigger trials, researchers collect sparse data about their subjects and the response. They only want to show an intervention “works,” and more data is costly and will potentially show unwanted effects. So, when an intervention ‘works’ (on average), it may only have helped something like 60% of participants, did nothing for 20%, and hurt 20%, and there is no way to predict whether you’re in the 20% or the 60%.
And even well-designed trials almost never try more than one intervention at the same time, so you have no way to understand synergistic effects, good or bad, that may result when combining diets, supplements, sleep tools, etc.
According to Andrew – and Fount COO Clayton Kim, who led machine learning engineering at Wayfair prior to joining Fount – you need a seemingly crazy data set to build highly accurate predictive models that could deliver personalized recommendations, experiments, adjustments, and plans. Specifically, you’d need to run “dozens of experiments per person, across thousands of people, with rich pre- and post- data for each experiment.”
You’d need to get comprehensive blood tests, understand habits like sleep, exercise, and diet, orient around goals, and then design experiments across a variety of categories – nutrition, supplements, meditation and cognitive tools, sleep, light, temperature, exercise, and recovery – track results with wearables, more blood tests, and conversations, and then adjust and run more experiments.
If you could pull that off, though, you’d have that exquisite data you’d need to build the predictive models. According to Andrew, experimental data isn’t just the best data, it’s the only data that works for the goal. When you increase the number of experiments you do per person, that’s when the magic really starts to happen. This is not a problem where you can just throw a ton of OK data at the problem and make it up with good models. Visually, it looks something like this:
Even with observational data from infinite people, or sets of small-n or single intervention trials from the same, you can’t get enough predictive power. But with clean, iterative, exquisite experimental data from a relatively small number of people, you can.
So that’s Fount’s proximate goal: to collect experimental data from enough people to build useful predictive models.
How many people is enough? 5,000 people. With experimental data from 5,000 people, Fount thinks it can deliver incredible results for 70% of the population across the areas people really want: energy, focus, mood, sleep, stress, gut health, losing fat, gaining muscle, athletic performance, and longevity.
Fount believes it can collect that data set in the next two years.
And as mentioned above, not only are they collecting the data, they’re doing it with a program that is so focused on the experience and benefits that their customers happily pay them to do it. That’s Fount’s Company Innovation.
Which brings us to my experience with Fount.
My Fount Experience
When Henry first introduced me to Andrew in late July last year, it couldn’t have come at a better time.
After a year of living with Puja’s parents during COVID and the first ten months of Dev’s life, we’d just moved back to Brooklyn at the start of the month. It was just the two of us and a baby for the first time.
Plus, the world was reopening – after a year or so of a very light social calendar, weekends weren’t entirely free to write anymore.
Plus plus, the writing schedule was getting intense. There was a period in June during which I wrote five pieces – on Cityblock Health, Zero Knowledge Proofs, Cometeer, Scale, and NexHealth – in 14 days.
Plus plus plus, I had just launched Not Boring Capital and was preparing to write my first LP Update, and I was stressing that my LPs would think I’d made way too many investments in the first quarter of the fund’s life. Not to mention the work of making and supporting all those investments.
I’d always been pretty even-keeled and low-stress, but I was starting to break.
When Andrew and I spoke in mid-July, we talked about the business (I committed to invest after our first call) but also about my stress and energy levels. We came up with the deal: Andrew would put me through the Fount program, and if it worked, I’d write about my experience.
In August, we scheduled a full blood test to get a baseline. I went in to get my blood drawn at a center, but in many cities like LA, SF, and Miami, they’ll send a phlebotomist to your house. The process was smooth despite that – Thea on the Fount team set it all up and gave me instructions for the whole process. I didn’t have to pay; it was part of the program, and not just because I was a special client - once you sign up with Fount, everything is included.
My first hint that this was going to be more comprehensive than I’d expected was this: they took 14 vials of blood and a urine sample. That’s when I officially knew that Theranos was bullshit. They also sent a full set of wearables:
Around the same time, Andrew and I set up an hour-long intake call on which I told him my goals. The biggest were:
While I wasn’t paying for the program directly, I wanted to keep the ROI in mind as if I were a normal client, so I set this high-level goal: “If I could write even one more deep dive a year, Fount would pay for itself.”
Andrew took that, my wearable data, and my lab results to craft a v1.0 personalized program – or set of early experiments to run – and in September, I had a Program onboarding call with Andrew and Guy, the ex-Navy Special Operator who shared the role of Performance Advisor for me.
To kick off, we went through the results of my blood tests. Hearing Andrew explain the results, and how they tied into the fatigue and stress I was experiencing, was like hearing Miles Davis play the trumpet. He’s a master.
And he didn’t pull any punches. While I came into the program with pretty good lab numbers – my cholesterol numbers, thyroid, liver, and kidneys were all in the optimal range – I learned some surprising things, too:
I was anemic, which I never knew.
My fasting glucose was high at 96.
My Vitamin B12 was functionally low.
My Omega-3 levels were very low.
My testosterone was a little low.
Based on my interview, I had signs of elevated inflammation.
Given those numbers, Andrew told me, it was no surprise I was feeling so tired and stressed. As he and Guy wrote in the intro to my v1.0 Program in the app:
Then came the important part: the program itself. Andrew and Guy gave me a comprehensive, personalized program across six categories:
Nutrition & GI. Personalized diet (which I only did an OK job following tbh) and supplement regimen.
Resilience. Supplements to support my immune system.
Cognitive. Meditation routine and guided meditations, plus the amino acid l-theanine to support calm energy.
Sleep. They threw everything at sleep: light filtering glasses, Hue color changing lights, nighttime shower and meditation, nasal rinse, recommendations on when to stop eating and drinking, and sleep supplements.
Exercise & Recovery. The Fount program comes with WRKOUT training sessions included. They’re live video sessions with a top-notch trainer – shout out to Marc! – as many times a week as you’d like.
Supplements. A meta-category that wove through and supported the others.
To give you a sense for how detailed and personalized things get, take how Andrew talked me through whether I should eat breakfast or not. Peep the footnote2 for Andrew’s full explanation, but the long and short of it is this:
If you Google “should I eat breakfast,” you’ll find that the answer is definitely yes, definitely no, and also maybe, but we now have an understanding of why eating in the morning can help some people and may hurt others…
When we looked at your labs, your wearable data, and your interview data, it looked like you had both high inflammation and sensitivity to stress hormones, so we experimented… We tested what happened when we supported you via supplements to keep your inflammatory levels in the optimal range and then added breakfast to support your brain, and voila, much more energy and focus when you wanted it!
Ultimately, one size does not fit all; it doesn’t even always fit the same person! And that was just one of dozens of experiments.
Everything in the program is supported in the app, where I could check in each day to see what I needed to do – when to meditate, what to eat, what supplements to take when, etc… – stay in touch with my team for any questions, and take notes on what I’m feeling, all in one place.
The whole program is designed to make it as easy as possible to comply.
If you drink and want to keep drinking, that’s more than OK. They’ll work to mitigate the effects.
Only want to work out twice a week? Fine.
I asked for blue blocker glasses that were less intense so I could use them while I worked, and they sent them the next day.
When I went to a friend’s bachelor party in Las Vegas, Andrew sent me his famous FlyKitt, which kills jet lag, and a hangover fighting kit. Hand to god, it worked.
When I tweaked my hip running, Andrew hooked me up with a top physical therapist who gave me three stretches I’d never tried and fixed it in a session.
To show the extent they’re willing to go to for customer experience and convenience, they built their own custom supplement packaging facility. They know that no one is going to take supplements from 7 different bottles every day, so they send personalized pre-packed supplements by time of day. Here’s them preparing my supplements using their fancy machine – they’re literally personalized with my name and list out the supplements for each meal (you can even watch a video here):
Throughout the program, I’ve had video check-ins every 1-2 weeks with Guy and Andrew. These check-ins are crucial for gathering data that they might not be able to get through tests. Every time we spoke, Andrew asked “Anything else you’ve noticed?” three or four times, hoping to uncover little connections. I was legit shocked when I’d describe something little that I’d barely noticed until he asked, and he’d say something like, “I thought so, that’s probably x, y, and z happening.”
I also had monthly blood tests to understand what was happening in my body and adjust the program accordingly. In the early tests, my labs actually weren’t moving as much as we would have liked in a couple of areas and I was still a little groggy in the morning. So we re-worked the plan – new supplements, diet tweaks, cut Snus, put on the blue light glasses earlier – and kept checking in. The labs and energy levels came right along.
When work was cutting into my sleep time the week of ConstitutionDAO, they designed an optimal napping program (turns out shorter can be better!) and Guy gave me his pushup and squat program to use for a pickup between meetings. Then they’d check in again to see if it worked, and keep or tweak pieces accordingly.
This is crucial – when I wrote about experimental data above, this is what I’m talking about. Andrew would form hypotheses, test interventions, measure the results, tweak, and try again. They capture it all on the backend. I happily gave Fount that data because it made me healthier and higher-performing. The combination of feeling better subjectively and having objective data to show the benefits is very powerful.
Note: Fount is not in the business of selling my or any client’s data. It’s their competitive advantage.
So did it work?
The proof is in the labs. As of my last lab tests, I’ve improved in every area we identified in the beginning:
My fasting blood sugar improved from 96 to 86 md/dl
My Red Blood Cell count went from an anemic 4.11 back to a healthy 4.61 million/ml, and my hemoglobin went from 12.8 to 14.3 g/dl
My Vitamin B12 went from a functionally deficient 339 to a normal 501 pg/ml
My omega-3 levels doubled from a deficient 3.1% to a good 6.2%, on the way to optimal
Even my testosterone went up from 572 to 625 ng/dl
I gave the numbers first because just saying, “And now I feel great!” doesn’t really mean too much. But I’m legitimately more energized and less stressed than I was before the Fount program. I went from Sleep Scores in the 70s before the program to consistent 80s and 90s now. I wrote more than 25,000 words in the past week alone. I have a gauntlet month writing-wise ahead – something like eight pieces in four weeks – and I’m not worried about it.
All-in-all, even if I’d paid full price, the program would have returned itself many times over already, and it’s ongoing. Now that I’m through the three-month Core program, I’m staying on the Support program: I still have supplements coming, less regular blood tests, monthly check-ins, and the ability to message Guy and Andrew whenever something comes up. It feels like how a billionaire would manage their health.
It’s a win-win: I get a billionaire-level performance coach and program on-call, and they get to continue to collect my data to make this level of optimization available to everyone.
The question is: how can Fount scale that magic?
Fount has a Sustainable Business Model today: at $5,000 per month for the three-month Core program and $1,000 per month for the ongoing Support program, a long waitlist, growing corporate sales, and software-level margins, high-end health and performance advisory is a very good business.
But remember the Elon Musk Company Formula. The current business is just the Sustainable Business Model that funds the Company Innovation.
So Fount needs to scale beyond its current approach in order to improve humans’ health and performance, lower the incidence of chronic illnesses, and allow people to achieve more. It’s doing that in two stages:
Scaling the data collection and experimentation machine (Company Innovation)
Using that data to build predictive models of how the body and brain work in order to optimize them at scale (Match)
Fount could get to Stage II without optimizing Stage I, it would just take longer. But there’s an urgency to Fount’s mission: the faster they collect exquisite data from 5,000 people (and beyond) and build models with it, the faster they can start improving lives at scale.
So let’s dive into Stage I.
One thing you might notice from that last section is that it’s very hands-on. It doesn’t seem particularly scalable for the CEO to get on every check-in call with a schlub like me. So Andrew has figured out how to replicate himself, by training elite coaches, so that Fount can build the dataset to automate Andrew and enhance Automated Andrew with better and better predictive models.
In the beginning, Andrew needed to be involved in everything. There was no software. There was no team. There was no data. There was no shared knowledge. It was all in his head.
Slowly, as he grew the team, built software, tuned ops, refined an internal training program, and shared knowledge, he lowered the percentage of situations that call for his or others’ expertise. Now, the Fount Performance Advisors can handle any standard situation and most that used to feel like edge cases. They only call in Andrew for the toughest edge cases:
Those edge cases are things that aren’t obvious; things that maybe no one has ever solved. This is when he needs to play House.
Andrew gave me an example of how this process works. It’s not one-size-fits-all; it’s “run experiments with each client until something works, then add the learnings to the toolkit.”
Fount had a client who was taking his company public. Because of a tough childhood, the client’s sympathetic nervous system (fight or flight response) is turned up so high that if he uses any stimulants like caffeine, he can’t sleep. But he needs to focus. So the Fount team tried a few things:
They tried stimulants that usually lower stress levels and don’t normally impact sleep, but they did in this client.
Then they tried various types of meditation. It didn’t work, because the problem was in the body, not just the mind.
Then they tried breathwork, which helped the client’s sleep, but not his focus.
Then Andrew paused and took a few days to reconceptualize the problem.
Andrew knew that (in a very reductionist sense) focus requires activation of the brain’s reward and drive pathways. Staying reductionist, these equate roughly to the roles of dopamine and adrenaline. So he asked - what if we increase dopamine without increasing adrenaline? What if he gave the client a mix of dopamine precursors with the breathwork?
That was the solution. The client told Andrew it gave him “95% of the effects of Adderall, but I can sleep,” and he successfully took his company public.
Figuring out the right reconceptualization of the problem was really hard and took Andrew’s decade-plus worth of experience and knowledge to figure out. But now, it’s just a tool in the Fount toolkit that all the Performance Advisors know how to use. It doesn’t work for everyone, even for everyone who has a similar setup to that client, but for the people it works for, it’s magic. And Andrew told me, with that look that I get when I brag about Dev, that his Performance Advisors now solve edge cases he hadn’t run across before without him.
With each edge case Andrew & team solve, they identify new tools and get faster and faster at figuring out which magic will work for which person… and then run tests, collect data, and keep improving the models so clients get results faster and faster. Each time, the amount of things Andrew is needed for shrinks. Each red sliver that goes from Andrew’s brain into software becomes another reinforcement in Fount’s moat.
Put another way, each time Andrew or someone on the Fount team solves a new challenge, it increases the gap between what Fount knows and what a human expert knows, and increases the gap between Fount’s ability to figure the next one out and a normal expert’s. It reminds me of that wild Google stat:
And of 7 Powers author Hamilton Helmer’s hypothesis on Google’s Power on Invest Like the Best (full transcript here): it’s a Network Economy, because so many searches are unique that the fact that people turn to Google more often means they’re better at returning results for the edge cases and long tails, which creates a more magical experience for each user, which means more people search on Google, which means better solutions to edge cases, and so on.
That’s just search. Fount is trying to do the same thing for the human body.
In addition to the edge case solutions, Fount has increasingly optimized its ability to collect the best data from clients efficiently. Today, this is what they collect from each client:
115 lab data points every month
30 minute video check-ins weekly
20+ wearable data points daily
1-5 app-based prompts answers daily
5 Factor Personality Data
Soon, they’ll be adding genetics, epigenetic markers, full-body MRI, body composition, medical records data, and more.
Importantly, they’re building up a library of not just what works for which types of people, but also when and in what order. They’re developing an understanding of softer things, too, like how to sequence interventions to increase adherence. For example, they typically start with “easier wins” like nutrition and supplements before building up enough extra energy and buy-in from clients to get them to, say, meditate regularly (if that’s what they need to reach their goals).
Ultimately, in Scaling Stage I, Fount is simultaneously building up the team, operational muscles, experimental data, software, and institutional knowledge to get ready to really scale.
That’s Fount 2.0.
The Fount Concierge program that I went through is the gold standard. It’s hard to imagine a more comprehensive, white glove experience than the one Fount currently offers. The product is essentially access to a world-class team dedicated to running personalized experiments and providing tools and interventions to help you meet whatever physical or mental goals you set.
But that only scales to a point. $5,000 per month is a non-starter for 99.99% of people. And remember, the goal here isn’t to build a profitable small business, it’s to increase human and financial capital available to build a better future. For that, Fount needs to scale access to Andrew and his team and drive down costs.
Luckily, the world of software has figured out a solution: APIs.
In APIs All the Way Down, I wrote:
When a company chooses to plug in a third-party API, it’s essentially deciding to hire that entire company to handle a whole function within its business. Imagine copying in some code and getting the Collison brothers to run your Finance team.
At this point, no company in the world could afford to hire the Collisons to run their Finance team, but every company can afford to plug in Stripe’s APIs and get access to a suite of financial products designed by the Collisons and the people they’ve hired, enhanced by the huge volume of data across Stripe’s customer base. Stripe has thousands of smart people and trillions of data points with which to solve problems, cover edge cases, and continuously learn and evolve narrowly-focused products.
That’s a key point: every company would now be better off using Stripe’s APIs than hiring Patrick and John to code payment acceptance and identity solutions for them.
While Fount says that it’s building an operating system for the human body, what it’s really building in the short-term is a set of health & performance APIs that allow you to access the collective knowledge of the Fount team and toolkit they’re building on top of their unique data set.
After Fount uses its concierge offering to collect the experimental data across 5,000 or so people (which it will obviously keep expanding from there), it can build predictive models that would allow it to turn the core Fount program into a dramatically lower-cost software product: Fount 2.0.
In under two years, Fount expects to have that product live. That would mean that for maybe fifty dollars a month, about what the average American spends on supplements monthly, people will be able to access an experience like working with Andrew and Guy, right in the app, supercharged by a growing corpus of experimental data and predictive models.
That doesn’t mean that we’re quite at the end-state that Tim Urban described – the perfect, personalized, health-optimizing meal delivered at just the right time – but that software will be able to help you run the experiments to find what’s optimal for you. As Andrew described it:
Considering how complex the body is, it’s very hard to make software that can just give people the perfect answer for their health or performance off the bat. Instead, with our initial data set we will be able to design a software system that can give people the optimal path of experimentation to get to the right answer.
Start with the goals – you want to achieve X & Y – and we’ll tell you which blood tests to get done, surveys to take, and wearables to use, so we can establish a baseline and understand your unique characteristics. Then, we can design 5-10 experiments that get you to the right program and results within 30 days, and continue to adjust based on data and your feedback.
From a business model perspective, Fount evisions charging a low monthly fee for the software and digital program so that it’s accessible for millions of people – which is better for the world and for Fount’s moat – then layering on upgrades for supplements, coaching, trainers, additional blood tests, more wearables, and even meal delivery.
My biggest feedback on the program was that instead of just telling me what to eat, they should actually just send every meal that I need to my door, leaving me no room to stray. I didn’t eat as well as I should have despite knowing what to eat, because I’m human and I’m lazy. Andrew said that’s coming this year – either via partnership or even by hiring chefs in Fount’s biggest markets.
While many meal kit businesses have struggled – largely because retention is difficult – I can see a world in which Fount actually builds an incredibly sticky meal kit business because Fount will know better than anyone what you should be eating when. It could even algorithmically update meal kits to subtly run experiments without any mental load on clients. They already vary supplements week-to-week, and sometimes within the week, to do this. A meal kit competitor would need something more than a free first box to pull me away from the company that’s designing meals to my specific biomarkers.
Fount’s opportunity in meal kits hints at the biggest hidden secret in this business: the many, many things that Fount can do to improve human health and performance – both on its own and by igniting the industry – once it’s collected enough exquisite data.
Increased Chance of a Good Future
In all of Elon Musk’s companies, the Sustainable Business Model funds the Company Innovation, which it uses to light a Match That Ignites An Industry.
Urban gave examples of such matches pre-Elon:
These are BIG fucking matches. For Elon’s companies, the three matches are:
Tesla: Affordable Super Popular Electric Car
SpaceX: Dramatic Reduction in the Cost of Space Travel
Neuralink: Breakthroughs in Bandwidth and Implantation
I shared my idea that Fount was running the Elon Musk Company Playbook with Andrew and asked him to fill in the blanks.
He said that for Fount, the Company Innovation is the way to collect rich experimental data from thousands of people, by packaging data collection as a white glove concierge program. The Match That Ignites an Industry will be predictive models of how the body and brain work.
Anticipating my question, Andrew jumped in: “But are models really a 'match' like integrated circuits? Yes.”
We've moved from systems that can be modeled from the top down using computing and physics to systems that are so complex and varied that we have to start from phenomena and use those to derive models - which will then help us to understand the systems and build tools for them.
You can build a rocket and understand essentially all the physics around it today, but we can’t even model one human effectively, let alone understand the general rules that govern humans with real fidelity. But it’s worse than that - we can’t even model one part of the system effectively because just one part of the human system, take the gut for example, is more complex than a rocket. Let alone the brain. Or at least they are until we build the data from which to understand them and unravel this complexity.
Now, this is obviously going to be really fucking hard to pull off. There’s a disclaimer that I should put at the top of every Not Boring to keep from repeating myself: the odds are that Fount won’t pull this off. The companies that are ultimately most impactful are probably the ones with the odds stacked highest against them at the start. Fount’s is an absurdly ambitious goal. If it pulls it off and builds this model, it will be worth tens or hundreds of billions of dollars; the current valuation reflects a steep discount rate on that future.
But Andrew and the Fount team have the best chance of anyone I’ve come across to build the predictive model of the body and the brain. If it does, the opportunities are enormous, for Fount itself, and for humanity.
First, as we’ve discussed, will be Fount 2.0: Andrew, in app form, supercharged by data and models.
By bringing the price down by two orders of magnitude, Fount will be able to improve the health and performance of many, many more people and grow the business substantially. Over 20 million Americans already spend more than $100 per month on supplements. Even Equinox, at a much higher price point, has 350,000 members, and they’re limited by physical proximity and space. More importantly, though, many more members will mean two things:
Expert Vortex. API-first companies tend to pull in many of the people who care the most about, and are the best in the world at, the specific thing the API-first company addresses. That makes the company even better at solving all of the hardest problems in the space, and even more of a no-brainer. If there are ten Andrews in the world, or maybe 10 true experts in each micro area of health and performance, Fount should be able to pull a bunch of them in to solve an increasing number of edge cases and ensure that Fount stays cutting edge in every area it helps clients.
More Exquisite Data. If Fount can build the predictive model that delivers great results for 70% of the population across ten major categories like energy and muscle with experimental data from just 5,000 people, getting experimental data from hundreds of thousands of people should help it continuously improve the model, serve more people, and address more goals. They will discover brand new approaches no one expected. And they will get better at ordering interventions and making the program as easy to comply with as possible so that more people stick with it, even if they don’t have a coach.
From there, things start getting really crazy.
Fount has an opportunity to build the platform that others – entrepreneurs, pharmaceutical companies, and more – can plug into to work together towards the goal. The core asset that Fount is building up is that rich, exquisite experimental data, and that data might be incredibly valuable to companies beyond Fount, even when they aren’t selling the data.
What might that look like? Just a few examples:
Clinical Trial Recruitment. It’s hard and expensive for pharma companies to find the right people for clinical trials, and they’re often willing to pay thousands of dollars per person to find participants. Fount can pre-qualify people based on their physiology and goals and match them with trials that might help them. Fount customers win because they get preferential access to new therapeutics, and the pharma company gets the people they need and a massive amount of data to understand who each participant is, if the Fount customer chooses to share.
Develop Candidate Pharmaceuticals. If Fount has the data, why not develop its own pharmaceuticals where lifestyle isn’t enough? It might be able to use its insights into the physiological effects of supplements and nutrition to develop drugs on its own, and its ability to predict what will work for whom and to understand side effects better will give it a leg up in clinical trials.
Platform for Tech and Tools. Fount will have a good idea who will be interested in and respond best to new wearables and supplements, so it can either purchase and resell the best tech, tools, and supplements, or partner with the best products to give them targeted access.
In all of these cases, it’s important to note, the outcomes are win-win-win. Clients win via access to the trials, drugs, tech, tools, and supplements best suited to help meet their goals. Partners win via more targeted and data-driven access to the right clients. And Fount wins by providing a better experience and outcomes to clients, strengthening its moats, and taking a cut of economic activity.
Most importantly, it wins by accelerating the future it wants to come true.
There’s a vision of this future in which Fount goes around the medical system entirely – and inspires other entrepreneurs to do the same – and replaces a big chunk of the primary care market.
But the one that Andrew’s more excited about is a future in which doctors use Fount to supercharge how they help patients.
The average doctor gets less than 10 hours of training on nutrition. Working with Fount to optimize patients’ lifestyle choices would make doctors more effective, and partnering with doctors would give Fount more experts to feed its brain and another channel through which to reach more clients. At some point, Andrew thinks that Fount might hire its own doctors and build out an insurance program that runs off of Fount 2.0. It’s the ultimate value-based care – using predictive models to make people healthier and underwrite insurance policies. It makes Fount’s market even bigger and increases the chance of a good future.
Whichever way it goes, at this point, we can see how all of the pieces of the Fount plan fit together, working back from “Increased Chance of a Good Future.”
We started by talking about how Elon works backwards from an increased chance of a good future. When you flip it around and start from where Fount is today, the path gets more clear.
The Starting Point is that healthcare is broken – we spend too much money on managing diseases versus preventing or curing them, and we rarely optimize for performance. We’re wasting human and financial capital.
Today, Fount is in the “Sustainable Business Model” phase. Just like rich people have always been happy to spend on nice cars, they’ve also always been happy to spend on things that make them look, feel, and perform better. At $5,000 per month to start, Fount generates strong margins on real, and rapidly growing, revenue.
Fount 1.0 funds the Company Innovation, directly and with VC dollars. Fount’s innovation is that it will be the first organization – company, government, or university – to figure out how to collect the experimental data needed to build predictive models. The company’s best estimate is that that will take 5,000 people. You can be one of them.
That data will help Fount build predictive models of the brain and body that deliver real, measurable results to large numbers of people via software. If it pulls it off, the models will be the Match That Ignites an Industry.
It will take a revolution in the way the industry approaches health and a wave of companies building powerful solutions on top of good data and models to get to that future. Many companies will build on top of Fount’s platform, others will compete.
If Fount is able to successfully ignite the industry, it could mean achieving the Goal of putting a big chunk of the $4 trillion we’d normally spend on healthcare annually in the US alone to more productive uses, helping people reduce stress and tension and increase self-understanding, and giving people the tools to maximize their potential (while still enjoying life to the fullest).
The Result of the Goal, if everything goes just right, will be to optimize people so that we can build an amazing future together.
Fount’s approach might just be crazy enough to work. If it does, we might get the future that Tim Urban predicted, with sensors, AI, and drones teaming up to deliver us the things we need to effortlessly stay healthy and perform our best. Fount is getting a headstart on building the predictive models that AI will use to deliver the right solutions at the right times.
Andrew pointed out that Neuralink data would actually be a great input to Fount models, and might even need something like Fount to really work in humans.
“If you want to overclock the brain, like Neuralink does,” Andrew pointed out, “It needs to be incredibly well-supported physiologically and can’t have interference from a poorly-functioning body. You need to optimize the whole.”
All this, starting from a $5,000 per month Concierge performance program and ending in an increased chance of a good future for humanity… mind-blowing.
Try Fount For Yourself
Experiencing Fount for yourself beats reading about it. If you have health and performance goals that you want to hit this year, it’s worth talking to the Fount team.
To apply for the Fount Not Boring Scholarship, fill out this form:
To skip the waitlist and talk to the Fount team about a personalized program:
Thanks for reading, and see you on Thursday,
*Investing in securities involves risks, including the risk of loss. Composer Technologies Inc., SEC Registered RIA.
The Whoop is going to seem really hilariously old school in twenty years, like “Remember when we used to wear this big bracelet,” right? It’s gonna turn into little sensors in our blood probably or even infrared, it’s going to be collecting 100 times more data than it collects now. More nuanced data, more specific to our body. That’s the hardware side.
And then the software’s gonna be – and I have not done my deep dive, this is all speculation – the software’s gonna get really good. This is the AI component.
Think about if hardware were able to collect… first of all, the hardware knows your whole genome, and we know a lot more about what a genome sequence means. Because you can collect your genome now, we just don’t have much to do with that information. So you have your genome, you’ve got what’s in your blood at any given moment, all the levels of everything… It knows how fast your heart is beating, and it knows exactly the amount of exercise, it knows your muscle mass and weight and all that. Maybe even knows your emotions…
Now the AI can go to town. Think about when the AI gets really good at this, and it knows your genome. I want the AI to just tell me what to do.
OK how about this? Now imagine attaching that to a meal service, and the meal service has all the million ingredients and supplements and vitamins and everything, and I tell the AI my broad goals: I want to gain muscle, or I want to maintain my weight but I want to have more energy, or whatever, I just want to be very healthy. Everyone wants the same ten basic things: you want to avoid cancer, various things, you want to age slower. Now the AI has my goals, and a little thing pops up and says “Beep beep, in 15 minutes you’re going to eat.” Cuz it knows that’s the right time for my body to eat. 15 minutes later, a little slot opens in my wall where a drone has come from the food factory and dropped the perfect meal for that moment for me: for my mood, for my genome, for my blood contents. Because it knows my goals – so you know it knows I want to feel energy at this time, or I want to wind down here – it knows you way better than you know yourself, better than any human could ever know you. You still have some choice – a little thing pops up and says, “Coffee?” because it knows that I can have coffee for the next 15 minutes only because at that point, it knows how long it stays in my system, it knows what my sleep is like when I have it too late, it knows I have to wake up at this time tomorrow because it knows my calendar.
So this is something I think humans are wrong about. Most people will hear this and think, “That sounds awful! It sounds dystopian!” No it doesn’t! It sounds incredible, and if we all had this, we would not look back and be like, “I wish I was making awful choices every day like I was in the past.” And these aren’t important decisions. Your important attention and focus can go on to your kids, and work, and helping any other people… things that matter.
I think we’ll all be so much healthier. When we look back today, one of the things that’s going to look so primitive is the one-size-fits-all thing, like reading advice about keto. Each genome is going to have very specific, unique advice coming from AI.
If you Google “should I eat breakfast,” you’ll find that the answer is definitely yes, definitely no, and also maybe, but we now have an understanding of why eating in the morning can help some people and may hurt others.
If your challenges are primarily due to inflammation - inappropriate activation of your immune system - then not eating in the morning (intermittent fasting or time restricted eating) can turn down your immune system as your body saves energy for other purposes. That might sound bad, but when it’s activated at your own body, less is better. In return, you’ll have more energy throughout the day and a better mood, since inflammatory signals increase anxiety.
But, if your issues are primarily driven by how sensitive your brain is to fuel availability or stress hormone levels, then skipping breakfast can be a bad thing. If you haven’t eaten, then you will have less fuel in your blood stream for your brain, and this happens right when your stress hormone levels are naturally elevated, since you get a natural pulse of cortisol in the morning when you wake up. Cortisol makes it more difficult for the brain cells in your prefrontal cortex, the higher thinking part of your brain, to bring fuel in, so low fuel + high cortisol can really affect your ability to focus, analyze, control your emotions, and generally do the things that you want from your brain. If you get hangry, then this will further raise your cortisol and exacerbate the cycle, so eating can be very helpful to make sure your brain doesn’t run out of steam later in the day.
But it doesn’t stop there. What if you have both challenges, which isn’t that uncommon? Then you may feel better stopping or starting breakfast, but be stuck in a local optimum, or you’ll feel better in some ways, but worse in others. Then, we need to solve one or both challenges - e.g. support healthy levels of inflammation - so you can eat breakfast to support your brain, or vice versa.
So back to your case. When we looked at your labs, your wearable data, and your interview data, it looked like you had both high inflammation and sensitivity to stress hormones, so we experimented… We tested what happened when we supported you via supplements to keep your inflammatory levels in the optimal range and then added breakfast to support your brain, and voila, much more energy and focus when you wanted it!