a16z Podcast - a16z Podcast: On Wearables, Quantified Self, and Biohacking

Episode Date: May 1, 2017

It’s the end of the beginning — not the beginning of the end — for wearables, argue the guests in this episode of the a16z Podcast. Especially as we move from the first, to the next, generation of wearable devices: not just activity trackers and watches but VR/AR gear, “hearables”, continuous glucose monitors, and more. The quantified self movement then takes these empirical tracking- and data-gathering tools to better reason about what works and doesn’t work in our bodies to help us solve problems and live better lives. Yet the act of gathering data isn’t the hard part… it’s linking them to insights and outcomes. Because we really do have very little data about what works at a collective let alone an individual level. With a new age of biohacking upon us — where people can apply engineering principles to manipulate what we take into our bodies (inputs) to tune how we perform (outputs) — can we finally embrace these tools? What will it take to make something that’s mainly a niche activity/community (quantified self was formally started a decade ago!) into something more mainstream for all? (Hint: it involves cookie recipes.) And finally, what are the societal implications of all this, from avoiding data dystopias to embracing the consumerization of government projects too? Joining us to explore these questions and more (in conversation with Sonal Chokshi), we have: neuroscientist and data scientist Rachel Kalmar, currently a fellow at The Berkman Klein Center for Internet and Society at Harvard University; co-founder of The Quantified Self blog and community Gary Wolf; and Geoffrey Woo, co-founder and CEO at Nootrobox (an a16z company).

Hi everyone. Welcome to the A6 and Z podcast. I'm Sonal. Today we have a really interesting topic, which is all about wearables, quantified self, biohacking, and all three of those are actually related concepts. And joining us to have this conversation, we have Rachel Kalmar, and she is currently at the Berkman Klein Center at Harvard University. And I actually met Rachel a few years ago when she was at Misfit and I was at Wired and Misfit was a wearables company. And the thing that struck me most, Rachel, was that you had like six watches. on each hand or like some number. I think my peak was 38 devices that I wore every day. I'm down to like seven now. Which is a very reasonable amount. And then we have Jeffrey Wu, who's the co-founder and CEO of NeutroBox. Last but not least, we have Gary Wolfe, the founder of the quantified self movement. Welcome, guys.

I think the first thing is, like, why do you guys care about this topic, I guess? Yeah, I care because our humanity, our, are, our, our, are, our, our, are, our, our, are system is the most important asset that we have. And by system, you mean a body? Our body, yeah. I think we all want to live longer. I mean, I think this is a very human instinct, Cortez, finding the fountain of youth. And I think what's interesting with quantified self, measuring biometrics, biohacking is that we finally

now have the tools and the sensors that actually quantify and measure the inputs and outputs of the human system. So let's apply engineering perspective to the human biology. That's an interesting way you're framing it. And Gary, I'm curious to hear your thoughts. when we coined the phrase quantified self. How did you think of it? Now in daily life, we can apply some techniques

that used to be really professional techniques that engineers and scientists would use, but were very difficult for most people to use. These have been packaged into software. Even just storing the data is a lot easier. So much easier to track what's going on. So all of this empirical toolkit is now available to us when we want to think about something.

And yet we actually don't know very much about what works and what doesn't work, even in improving the daily aspects of our lives, never mind kind of the impossible dreams that are in front of us. And so by a daily aspect, you mean something as concrete as how we sleep? Right, exactly, exactly. How we sleep, how we feel. Go to any bookstore and you'll see shelf after shelf of ideas about things that are supposed to dramatically improve our lives.

Most of these ideas are tried without any attempt to carefully measure whether they work for us or not. And you may think that that sort of failure to follow up and failure to reason carefully about our actions typifies just new age stuff or self-help stuff. But let me tell you, even in the world of professional science, the drugs that we take, there's really a lack of ability to carefully think about how these affect us as individuals.

Even if we have good evidence at the group level, often what we care about most is, did it work for me and also what price did I pay in terms of, say, side effects? And all of these areas are now sort of dramatically accessible to us in a way that they never have been before. Right. So I'm hearing this theme of being able to link inputs to outputs and for the first time, because of all the information and tools we have, that we have this ability to do this. Rachel, you haven't shared your framing yet for how you think about this world and arrive to it because when I think of censorification, wearables are a big part of that and you're a leader in thinking about wearables. You got your PhD at Stanford neuroscience. Yeah, I spent

12 years looking at noisy signals from the brain and trying to understand how the brain produces is behavior. The thing about neuroscience is that most of it is done in a lab environment right now. But what we really care about is behavior in the real world. And up until, you know, the last like five, ten years, it was really difficult to get data about behavior in the real world. And now all of the tools that have been coming online provide us a way to be able to actually collect data about our daily patterns. and what we do, and that not only will it be helpful for understanding how the brain works, but it is really important for medicine as well.

We can take this data that right now is more in the health and fitness space, look at it in a longitudinal sense, and tie it together with clinically relevant outcomes. That's going to allow us to have better predictive models of health and disease. Imagine that we're driving cars and we only let ourselves open our eyes every minute. And that's essentially like the snapshot of information we get when we go to the doctor. Right? We get our panels and then the doctor's like, hey, course correct with this XYZ, right? That's like us, you know, jerking the steering wheel.

I heard another podcast about a health feedback loop for humanity. We have more signals that we use to recommend movies to us than we do for how to take care of our bodies. Exactly. A lot of these continuous measuring allows us to be constantly correcting. We should be manipulating our health and well. on this on a continuous basis, not a snapshot basis. That's what I think about with neotropics and biohacking. We have interventions entering as inputs into the human system. And, you know, we have sort of noisy signal coming out, whether that's working or not. Well, to really solve and answer

these questions on a rigorous level, we have to actually close a loop here. Finally, with concepts from quantified self, with wearables, with sensors, we can actually finally close the loop. You talk about we, but it's really, in this podcast, this is the three of you guys, how do you think about this as moving beyond early adopters? What do you think the obstacles and challenges are to actually making it more mainstream? I mean, Gary, the quantified self-aspect of this like 20 years ago. And Rachel, you've been studying this like 12 years, I think, including your time at Stanford. What are your thoughts for making it not just something that like really motivated people do? I think about this as like cell phones or cell phone cameras as they became cheaper and more

and more people have them. And cell phone cameras also used to be pretty poor quality photos. But at the end of the day, the best camera is the one that you have. And so people kept using their cell phone cameras and they got better and better. And now most of us don't own regular cameras. We just use our cell phone. The devices that we wear are on a similar trajectory, where right now we're in the poor quality photo phase.

Over time, they're going to be better. resolution and they're going to be used by more people. But I also think in passive ways that don't necessarily require engagement, a lot of the tracking will be invisible, which is different from the quantified self. There will always be room for people who want to be engaged and people who will have this data collection happening in the background. That's actually a really great analogy. I mean, even more of an original analogy, Homebrew Computing Club computer hackers, right. These are garage tinkers and now everyone has a supercomputer in their pocket. So echoing ratio, I think a lot of these things will become a lot more passive. I've been wearing

continuous glucose monitors designed for diabetics. Are you diabetic? I don't mean to I'm not diabetic. Okay. But I use it to understand my blood glucose levels and how it responds to fasting, how it responds to ketogenic diets versus, you know, a more carb-heavy meal. Because I think we all intuitively understand, hey, certain routine or regimen, or sometimes you feel better or worse. And I think reflecting why we care about collecting the same in the first place, let's actually get quantified to all the possible inputs that could contribute to a mental state of productivity or sloth. I mean, it's like high resolution. Right now, I eat something heavy for lunch and I'm like, oh, crap, I'm very tired. It's intuitive.

We understand that. Right. But I don't know specifically in what, yeah, exactly. But Gary, you've got the longest vantage point of us in terms of what worked then and now. Like, do you see much difference in the community genuinely for the community that you engage with? I do. I'm interested in quantified self as a cultural phenomenon. And cultural change has a certain time scale that seems fast in retrospect, but slow if you, for instance, have to deliver returns to your investors, you know, in a few years. And I think quantified self

has two meanings. On the one hand, it's used as shorthand for a market, and the market is, you know, wearables and fitness devices. On the other hand, the quantified self is an actual community of people doing real things and interacting with each other, developing methods, and sharing knowledge. The market is much bigger, and Fitbit alone, you know, has well over 20 million active users. So this is a lot of people. It's not, you know, just a very small, you know, group of advanced users. You know, Right. However, quantified self as a community is much, much smaller than that. And that makes sense to me because to be motivated to use these tools to make a new discovery, you have to have

a question that is really driving you. And fortunately, we're not compelled to really, really think hard about something that's really, really bothering us all the time, right? That would be that would be no good if that were the general condition that humans lived in. But none of us get away with a whole life without having something that we have to think about in which the off-the-shelf solutions aren't working. You know, we might have tried one or two things that seemed kind of obvious, but something didn't work and we have questions in our mind. So we use our brain and we use our tools that are available to us to observe and to make changes. In that dimension, the dimension of how do we reason about ourselves and make discoveries about the things

that are really important to us, even if you imagine, just for the sake of a kind of quick little heuristic, that only a tenth of a percent of the people who are actively using some wearable technology are engaged in a process that you might describe as a quantified self-process of reasoning with their data, that still tens of thousands of people, even today, growing ultimately into hundreds of thousands of people. And for cultural change, think about something like Wikipedia or other novel forms, new forms of knowledge making, that one-tenth of a percent at any given time is generating material that's really useful to many other people. Yeah, sort of a multiplicative effect. Yeah, I think seeing this cultural dynamic is really what's important in understanding how quantified self will change the general culture over.

time. Well, that's actually really interesting because didn't Neutropic sort of spawn on Reddit as a community? That is sort of the equivalent of sort of this very engaged base as creating knowledge for people in public. But I have a pushback, which is I buy the argument that things that are early trends, they start small and they grow big. I buy the argument that a highly engaged group of people can help drive that. I also think people are very lazy fundamentally and very resistant to change. And sometimes change does happen upon them passively to quote a word, earlier, but people wear wearable. And they don't do anything with that data because it's very easy to optimize for one half of this. And you kind of treat the act of doing it as the act itself.

It's as if you've already done it. So I think that that's completely right. And that's why we're seeing kind of a bit tapering in the wearables area. Like it's not the beginning of the end of wearables right now. I think that we're at the end of the beginning. So the way that I like to think about this is I like to compare all of these devices to mills that are turning into flour. And instead of flour, they're producing data. And some people say, hey, I love to bake. I have data. Like now I know exactly how I'm going to answer these questions about myself.

This is great. Most people say, okay, I have flour, but I don't bake. I want a cookie. Where's my cookie? And the challenge is that there aren't cookies yet. And that's not the fault of any one company. It's just a reflection of where the field is. We're still in the early days.

We have the crappy cell phone cameras of the early 2000s. The quantified cell community, people are swapping recipes and figuring out how to make cookies, but we haven't figured out kind of the mass cookie recipe that works for everybody. I think that what a cookie will look like will be apps and services built on top of this data that don't require people to look at the data. Predictive models like, hey, you're at risk for a heart attack because we've seen in all of the populations like you

that preceding heart attacks, this is what data looked like. That will be one path when we can tie these things to clinically relevant outcomes. And I think another path will be closing feedback loops automatically. So without the person in the loop, the mootering that would monitor how much experience, exposure I've had to light. And instead of going through the seasonal effective disorder,

kind of a funk where you wonder what's wrong with you. You could just have this mood ring, talk to the lighting in your house. They could automatically compensate and you wouldn't have to go through that cycle every fall. You're talking about closing the feedback loop and that's one of the cookies with predictive models and other cookie. What about design? I mean, design plays across all of this. Do you guys have any thoughts on the design aspect? thing that I've been thinking about a lot is that wearables are interesting, but could the sensors be in the environment around you? So, for example, could you imagine, you know, a futuristic clinic gym where, you know, instead of going to the urinal and you're peeing away all this data,

you're collecting all your, you know, your feces, your urine, you're having your microbiome sequence then and there. A biofund, which is a distinct fund from this conversation, one of the things that they said was that they think the bathroom is actually the ultimate diagnostic of the future. We're extruding out this data. Yeah, exactly, that it's the ultimate input output, you know, literally, vehicle. Just speaking of vehicles, the vehicle, like your car is another great place for monitoring. You're a captive audience. You can have sensors in a steering wheel.

You can have cameras that are pointed towards your face because you're facing in the same direction. There are a number of car manufacturers who are adding at these kinds of, sensors in the car, whether for use in driving, looking at drowsiness, et cetera, I guess in the same way as the bathroom is piggybacking on something that you already do, sitting in a car is something that you already do. Well, there's a best practice in design, which is to get deeply involved with the users of the products, there are often some really unexpected discoveries that you make when you bring yourself close to that situation, a range of questions that. that isn't very well described by the optimization culture of Silicon Valley.

And so I would say that it would be a useful exercise to stand a little bit closer to the reality of many, many, many other people and listen to how they describe the questions that are really on their minds. And just to give you an example of that, many people are deeply involved in caring for other people. And in doing that caregiving and caretaking, they often deprioritize some things like how much time they spend at the gym and how closely they monitor what they eat. And yet they take those values, family values and community values, to actually be superior. So in their minds, they make a choice that is a conscious choice and a justified

choice in which their physical health is sometimes sacrificed. In terms of designing products or designing services, I do think it's important to start pretty far upstream with an attitude of kind of open-mindedness and respect to the variety of circumstances in which users of those products may find themselves. I come from the world of ethnography and ethnographic methods are critical because you're not relying on user studies of what people say they want or do, but you're actually observing them in their natural behaviors and their natural habitat and then then designing for the individual as part of this larger ecosystem to your point. Are they a caregiver? Are they caretaker, mother, father, aging parents. It could be any, you know, permutation. Thinking about

technology and context is pretty critical for adoption. I mean, that's how the GUI came about, that you're really embedded in this perspective of how people are actually using it. Jeffrey, you referenced the Homebrew Club earlier, but the reality is it's only when people started thinking about, okay, how do we make this like for kids to be able to use computers? I think it has to be super passive for 90% of people to use it. So things like continuous glucose monitoring, you don't need to worry about it. You pull dead off naturally. Or to Rachel's point earlier, building it on existing things, like grafting it onto cars, grafting it onto the bathrooms, gyms, right, exactly. And then the only thing you can make active, you have to make it feel good at the end of the day.

You have to have some sort of closed reward loop. A lot of people that like exercising, it feels good to do that 15-minute run or when you're fasting. It feels good to, you know, get into ketosis, make sure that they actually have some sort of like gamified software loop. Yeah, gamification mechanic, right. I'm feeling good as its own reward, but it doesn't hurt to have, like my favorite thing is a to do list and being able to check it off. It's like the most satisfying thing in the world. It's a very minor thing, but it's a perfect example of how you can design for. And I make one quick comment. If you think about your idea about censorifying the environment, not just like the wearables, the first thing that went through my head is what is a baby monitor

in a baby's room, but that is literally putting visual sensors. It's actually a way of demystifying this, like, scary idea, this augmented vision into a child's room. Yeah. I think it's a good. That's a good example. This is a good moment to just do a quick terminology check. Gary, you define quantified self in terms of the market and the community, but like what is quantified self itself? Well, one of the ways to think about quantified self is as a variation on the words personal

computing. Quantified really is a synonym for computing. By quantified, we don't just mean it has a numeral, but we mean that it's a, uh, observations are structured in such a way that you can use some of the techniques of science and technology to work with them. So that's quantified. And then the self is a synonym for personal. It's a usefully ambiguous word that concentrates all of the most important and often implicit things that we care about. But the difference between personal computing and quantified self is that personal computing is a category of computing. Computing is the noun.

and personal is the adjective. Quantified self is really a way of thinking about the self. And the computing or the technology that's in the term is a way of thinking that's connected to lots of other ways people have thought about themselves throughout history. That's so interesting because it's the inversion of personal computing, but much more than that. Jeffrey, you referenced some other quick keywords?

Sure. So neutropics are compounds that have cognitive-enhancing aspects in colloquial use, neutropics or anything that enhances cognitive performance. That's not like a pharmaceutical drug. Well, I think neutropics is orthogonal to legality. There are, you know, scheduled drugs, supplements, and then drugs, prescribed drugs. Think of them as orthogonal pieces. Okay, so that's neutropics.

You said ketones a few times. So our bodies produce energy, ATP, using either glucose through a process called glycolysis or what's typically sort of a backup turbocharged state, using ketosis, which develops ketones. So ketones are compounds that are broken down from fat, and our cells can use that to create ATP. And it's very interesting because a lot of emerging data around ketones being more efficient for mitochondria to produce energy. You actually get more power per unit carbon using ketones as opposed to glucose. So that's been very interesting from a biohacking perspective of how to use ketones potentially as a more efficient fuel for

various performance. And how do you define biohacking? Biohacking to me is approaching the human system from an engineer's perspective. So being very thoughtful or algorithmic about inputs to maximize certain outputs. Rachel, how would you define wearables? I'd say that generally they're electronics that you wear, although I have an antique Tiffany pedometer from the late 1800s, which is completely mechanical. So I'd say that wearables or anything that you wear, that can help track or interact with the world.

We have fitness trackers and activity trackers as one category of wearables. Another category that's becoming more popular now are curables. Use our headphones that often have sensors in them. I recently got a Kickstarter pair of curables called B-B-I. They do a lot of the biosensing that is done on your wrist elsewhere. And they also have an AI coach.

And so the goal of hearables is to move towards like augmented reality for your auditory system. So like in the movie for so hearables is another category of wearables. And then headsets like emotive and things like that. Where does that fit in? The mess of wearable obviously, but like is that just like its own category? So I think anything that you wear on your body that collects data about you or allows you to interact with other devices. could be a wearable. It's interesting to see the different categories of wearables. It's not just activity trackers. Obviously, virtual and augmented reality is another category of wearable. Google Glass

paved the way for wearing wearables on your face. I have the new Snap Spectacles, and I just collected a lot of interesting eye-level videos on my recent Cryptoom. Yeah. I mean, I wear Ringley, which full disclosure is an A6 and Z company, but I actually knew her before, which is, it's basically a ring that does notifications. They also have bracelets with a now a fitness tracker built in. But my number one thing is that I care fundamentally about the fashion is wearing a piece of jewelry to me. And that to me is what makes like a wearable more ubiquitous, which is why I love that you mentioned and referenced snap because we don't have to think about these big clunky ugly devices. Also the 1800 Tiffany Podomer. That's, that's cool. I know. I want to see that.

I know. I'm wearing in this podcast, I'm wearing a ring that's an antique watch. We should actually share photos right after. People wear things for all sorts of reasons. But if we wore things just for utility, we'd probably be walking around wearing our workout clothes and our running shoes all of the time. And so this goes back to what we were saying before of understanding your users is really important.

If you can't get somebody to wear the device that you build, none of the sophisticated algorithms that you have matter at all. I totally agree. I mean, the broader things, theme of this conversation too, which is about changing culture. People have been wearing totemic objects for ages that have no function except maybe signaling for how many calorie shells you have on your necklace. And I think just like, I think reflecting back to maybe why it's a slow adoption.

I think I think Gary's point is spot on. Culture takes a long time to shift. I think one interesting notion is that we're probably the sickest cohort of humans in the history of humanity. If you look at obesity, Alzheimer's diabetes, I think one of the latest statistics, up to 75% Americans are predicted to be obese by 2050, which is insane. Yeah. So you're saying a majority of our country will be extremely, will be overweight in just like a few years. Yeah. And I think that just from a historical perspective, going to the gym is only a recent phenomenon

in the last 20, 30 years. Physical labor was just a part of our livelihoods for most of human civilization. And now we have to make an intervention of exercise back into. Well, we're contriving because of change and, you know, the way we work and live today. Like the fact that we have standing desks. Right. And I think if we constantly eat, maybe, you know, intermittent fasting is an interesting shared intervention.

We're all going to be tracking our blood glucose or insulin responses. I think that could be something that brings quantified self from more of a niche activity to be a standard protocol for people to live by. It's a natural evolution of etiquette and culture as innovation happened. Rachel, you mentioned noisy signals. Do we want to do a quick short definition on that? If you knew the ground truth of, say, how many steps you've actually walked today, that would be great, but that's not possible. And so all of your different devices have different algorithms that they're running.

And so you have multiple noisy views of some underlying round truth. And so one way I like to think about this, especially with respect to healthcare, I like to think of going to a doctor's visit as going to have a studio portrait taken, like a professional portrait where the lighting is right and they know how to pose you and a lot of the digital health tools that we have are much more like selfies taken on on an old cell phone which are not really great quality photos but they can taken together with the studio quality portraits they add more information and if you take all of the noisy snapshots of your life it will give you a better picture of what your life is actually like than if you only had high-quality studio portraits. You have the best analogies. Are there any parting thoughts on what's next or what you want

listeners to leave with after hearing your viewpoint on this podcast? Well, I would just say that if you have something that's bothering you, a question that's hard to answer, you could try some empirical approaches. What are the most interesting things that you're tracking now? Like, what's most actionable for you? The thing I'm most excited about is I did just get a freestyle Libre Pro arrived in the mail, and so I'm looking forward to learning about my glucose levels. What is that? Is that a continuous glucose monitor? Oh, CGM, I'm right. Continuous glucose. Okay, gotcha. Very cool. Yeah, I think that continuous biometric data is going to be inevitable. Data can be used for good or for bad. There's dystopic versions of how to use this data,

but the status quo as is is pretty dystopic, where everyone's obese. So I'm just excited, or at least hopeful, optimistic that the more. data that we have, the better decisions we can make to improve, not just ourselves, but the society around us. The interesting thing around the social dynamics around biohacking and eutrophics that there's always been two veins of groups. One, you have the community base, and these are people doing n-1 self-experiment. And then you have academia that's being funded by Department of Defense or DARPA to create super soldiers or just more enhanced humans. And I think we're at this interesting point where these two different groups are really converging to one. So we can start seeing

some of these military research, you know, blue sky ideas being productized and something that can be in consumer hands. That's actually the inevitable progression of technology too, right? Because like you have like things like the home brew club, which sort of took like this DARPA-based, like, you know, centralized notion of computing. And then you also have like the communities of personal competing, you know, PC enthusiasts and the mainstream. And we are an interesting point, I think with technology in general where we're seeing a lot of this where there are certain things

that are being developed at a government level and people at a consumer level are plugging into that in different ways, which I think is really relevant. Yeah. One of the things that we have to figure out as a society is how do we think about some of the questions

about what happens when our entire lives are monitored, not just by our wearables, but by our homes, by our bathrooms, by our cars, wearables are changing and evolving but they're not going to go away what kinds of systems do we want to have set up such that we were protected further down the line so we have Gina the genomic information non-discrimination act which prevents your genomic information from being used against

you in certain ways but all of this data that we're collecting now I like to think about it The same way that in the 1950s and 60s, x-rays were used in shoe stores, you could go in and have an x-ray taken of your foot to see if the shoe fit well and it was this novelty. But over time, we realized that there were actually very dangerous side effects to x-rays. And it didn't mean that the x-rays shouldn't be used or should be outlawed completely, but that there were particular circumstances in which we should use x-rays, but we also needed to exercise caution. And I think that we're going to realize that data is a lot like x-rays

in that there are very many positive benefits that we will be able to get out of having this data. But it's also important to exercise some amount of caution because data has become permanent and once it exists about us, it's there. And there are going to be a lot of potentially dystopian kinds of futures that might come up that. And so I'd say that for even if you think that you're not collecting data about yourself,

it's being collected anyway. And so it's important to think about some of the longer term implications of that. You can avoid it. So how do we think about moving forward positively? I'm so glad you brought that up because, you know, one of the things that I don't like is when we preemptively decide things before there's sort of this permissionless innovation and creativity that can happen.

But to your point, we absolutely have to be more thoughtful about those consequences. I mean, one great example at a very evolutionary scale of this conversation, we had Yuval Harari on this podcast. And he's the author of Sapiens and Homo Deus. And one of the big things that he pointed out is like we're as we enter a world where we may be able to augment ourselves with technology, it could actually entrench inequalities because a certain class of people might be able to afford a certain class of devices or augmentation that other people might not. And in fact, what was previously, you know, an inequality of, you know,

societal, cultural level becomes a technological one then. And then one that actually really affects outcomes. You have escape velocity. Yeah. Yeah. Yeah. Exactly. So to your point, there's responsibility to data. There's just broader question of things connected to surveillance and privacy and then of course a larger consideration around inclusion so that making sure that a certain elite class isn't the only one benefiting from such technologies. Okay. Well, thank you for joining the A6 and Z podcast. Thanks, Mal. Thank you. Thank you so much.