clock menu more-arrow no yes mobile

Filed under:

Full transcript: Color Genomics President Othman Laraki on Recode Decode

Genetics is now a software and data problem instead of a raw laboratory and science problem.


On a recent episode of Recode Decode, hosted by Kara Swisher, Othman Laraki, the CEO of genetic testing company Color Genomics, talked about how the data generated by our bodies has already become an important medical tool — and, as the cost of genetic testing continues to decrease, how it will become available to everyone.

You can read some of the highlights from the interview at that link, or listen to it in the audio player above. Below, we’ve posted a lightly edited complete transcript of their conversation.

If you like this, be sure to subscribe to Recode Decode on iTunes, Google Play Music, TuneIn and Stitcher.

Kara Swisher: Today in the red chair is Othman Laraki, the president of Color Genomics. Othman co-founded Color in March 2013 coming off of more than three years as vice president of product at Twitter. Color analyzes the genes of its users to see if they’re at risk for certain hereditary cancers. We’re going to talk about the future of health, tech and much more. Othman, welcome to Recode Decode.

Othman Laraki: Thank you for having me.

There’s so much to talk about. I just recently did a podcast with Vic Gundotra who’s doing EEGs on little small devices and just recently got money from the Mayo Clinic, and I recently interviewed Anne Wojcicki from 23andMe, one of the original genomics companies.

So why don’t you give us a little background on how you got to this first, so people can understand who you are. Because coming a VP of product at Twitter is not really where ...

It’s not quite the traditional background for health.


I guess in terms of background, yeah, my background prior to this had entirely been in the software world. I started several software companies, I was at Google for a number of years. And most recently I started a company that Twitter acquired, I was VP of product.

Which was?

It was called Mixer Labs. We were one of the early developer APIs. Kind of how now there’s a big wave of companies like Twilio and Stripe that are taking a piece of infrastructure and turning it into API for developers to build the applications, we were doing that with geo information.

And so we made it possible for ... at the time, there was a big wave of check-in apps, etc., when the iPhone and Android had just started using GPS through their API. And we built an API that made it easy for people to develop geo-enabled applications.

I see.

And then at Twitter I was working on product.

What did you work on?

I ran three groups: Growth, revenue and international. Essentially they’re all focused on different aspects of scaling the business. We were very privileged to be part of a pretty incredible time in Twitter’s history where this product was growing really fast and we were dealing with some of the really interesting scaling challenges and seeing this amazing product really reaching a very large population.

Yeah, now growing not-so-fast is their big issue.


Every company hits that, don’t they? So you decided to do this. Why? How did you get to this?

Something that I hadn’t talked about before Color was ... I have a fair amount of cancer history in my family, and my grandmother had passed away from breast cancer. My mother survived two of them, and then after her second diagnosis, her doctor’s like, “Well, you know, this seems pretty unlikely that you just have this,” so they recommended that she get genetically tested. And she discovered that she has a BRCA2 mutation, a gene called BRCA2.

The BRCA gene.

Yeah, it’s one of the two well-known BRCA genes, and they’re associated with dramatic increases in breast and ovarian cancer risk. I also got tested, and it turned out that I’m also a carrier as well. So it had been part of my personal and family life, I guess, forever.

But what happened was, we were just coming of this incredible phase in genetics where the technology behind sequencing had gone through a 15-year period when the costs were dropping at an exponential rate, and all of a sudden it started becoming possible to do genetics at a cost point that was completely impossible before.

One of the side effects is that it turned genetics into a software and data problem instead of a raw laboratory and science problem. What that means is that, all of a sudden, you can start building genetic products, or products that use genetics, but in a way that makes it accessible to the entire population, not just for a few extreme cases.

So you wanted to go into this versus, I don’t know, making photo apps at Facebook.

Photo apps are fun, but ... [laughs]

All right, if you say so. Because God knows we need another one.

[laughs] Yeah, exactly. But it was one of those cases where, yeah, it really felt like an opportunity to use a lot of the background that we had on the software side to have a very meaningful impact on the lives of a lot of people. And it was really one of those cases of combining people from very different fields and taking what used to be science and turning it into engineering, essentially.

My three other co-founders, all of them have scientific or health backgrounds. That combination, I think, is what really allowed us to make a difference.

So it does remain science, though. It is a science.

Oh, for sure.

And one of the issues in the past with companies like 23andMe and several others were issues around — and obviously Theranos has sort of colored everything — this idea of disrupting it in such a way that seems fantastic, and then maybe isn’t quite so much. I think there’s been a lot of headwinds for companies like this.

Absolutely, and that was actually one of the big initial thoughts behind how we approached the space. We’re very depth-first focused, where we said, “We’re going to choose one area where the science is well established and where there’s a lot of impact to be had.”

The connection of BRCA1 and cancer has literally existed for 20 years now and today, 20 years later, there are millions of people in the U.S. who are carriers of these mutations that still don’t know. And the science, by now, is very not debated anymore. So what we decided to do is to go for something like that, where there’s relatively small scientific risk, in some sense, and it was very much more of an application challenge: How do you do this with very high quality and very high skill ...

In a small area.


But obviously you’re going larger, correct?

Yeah, definitely.

So where do you fit in? On one hand you have something like Theranos, which again, could be issues of fraud there, there’s obviously some things gone haywire at that particular company, which the Wall Street Journal has chronicled beautifully.

And then you have 23andMe, which had issues — regulatory issues — but nobody disputed that these spit tests were interesting and you could find all kinds of things. And I was showing someone my results before the FDA got involved — which were fascinating, actually, and full of great information. And the results today are not very helpful, particularly, and that’s one of the difficulties: They give you very little information and treat it like a game versus something that’s really helpful to your health.

So how do you look at how Silicon Valley looks at these things and how you should be regulated by the government?

There are a few aspects to that. First of all, I think in general, in the early days of an industry there’s a temptation to go very broad and to try to do a lot of things to cover the entire space. That was again one of our big choices, to say, “We’re going to take something [where] there’s a very well established existing model, but we want to just do it at scale. That’s the first piece.

Secondly, when you’re dealing with health, obviously — and this is very personal, too, when dealing with people and their family’s health — the single most important thing is quality, in that the information that you’re delivering to people is a) accurate and b) made available in a way that is digested and used in a way that’s consistent.


Yeah, consistent with the impact of that information, right? If I’m reporting on the likelihood of having a unibrow, [KS laughs] the risk of that information is relatively minimal, right?

Well, not really.

[laughs] But ...

It’s a devastating illness.

But it’s actually been relatively easy to deal with. There’s 50-cent tweezers, and that’s a pretty low-risk procedure. As long as you don’t poke yourself in the eye.

But when you’re dealing with something like cancer, the importance and the impact of that information is much, much higher. So what we’ve done there is ... our standard was clinical practice, so for example, every test is physician-ordered, and every test includes genetic counseling.

So they know what to do. The idea is that the patient is getting information unencumbered by information, I guess.

Historically, people would think that there was two separate axes: One is quality and the other one is access. And if you improve access, it means low quality, or a low focus on quality, in some sense. And for us, our goal was to think about how can we increase access and maintain the expectation of quality.

So we’ll get into that idea of the FDA wanting some more genetic counseling and other information which was lacking in some of the earlier efforts. Explain how people go through your process, because it’s different than other genomics. There’s been a lot of genomics companies from Silicon Valley.

Yeah, totally. I think one of the big differences with us is that historically, when genetics used to be just the lab side of genetics and it was incredibly expensive and incredibly difficult, any company that did genetics defined itself as a genetics company.

One of the areas where we’re a big departure from pretty much everyone else is that we think of genetics as a part of a bigger service, where when people come to us, they don’t necessarily care as much about whether they have a BRCA mutation. What they care about is how do they and their family members manage risk of cancer.

And that’s how we think about our service. What are the things that we can do to help people manage the most important risks in their lives? Genetics is an important component, important building block, but it’s not the product. And that has translated in a few ways.

One is translated in terms of how we build the product and how we make it available to people, but it also impacts how we think about distribution. So, for example, when you think about it as a service, we work with a number of large self-insured employers. Companies like Visa, Blackrock, Tencent, etc., that have large populations where they’re bearing the cost of health care for their employees, and they offer Color either heavily subsidised or entirely subsidized to their employee base.

So it’s not a consumer product.

Yeah, exactly. It’s a product where we really try to help ...

Move through the enterprise.

Yeah, exactly. So in that case, obviously for employers it’s a big benefit, because there’s an ROI component to it, but also there’s a very direct human impact to it.

And now, every week pretty much, we get these incredible stories like — a couple months ago, we had this woman who didn’t have a huge amount of family history, got tested through her employer and discovered she was a carrier of a BRCA1 mutation. And because of that she was like, "Okay, it came from one of my two parents." She went and got her parents tested, and it turned out her mother is the carrier. And her mother had never had cancer before, but because of that, she went to see her doctor, who got suspicious, and they found that she had Stage 3 ovarian cancer. But it was not yet metastasized, so [it was] much easier to treat. This same woman, by default, would have shown up a year, two years from now with most likely a late-stage cancer where it’s a very different outcome.

So when you get these kinds of warnings for people, it’s like a small bit of information ...

So when people get it, they can get it just from a regular doctor assigning it to you? Or do you have to get it through work?

People can work directly with their doctor.

But they can’t call you up and say, “I’d like the genetic test.”

They can come on the Color website, purchase the test. If they have a doctor that they’re already working with, their doctor can easily go on our website and place the order. If not, if they don’t, we can help them with a referral to third-party physicians who they can work with as well.

And then once the results come in, they can talk to the doctor about them.


But you don’t have doctors on staff.

No, we don’t have doctors on staff. We have board-certified counselors who help with giving people the results, the interpretation of the information, as well as educating doctors oftentimes, because most doctors ...

They don’t know genetic tests.

When they went to medical school they had, like, one lesson on genetics, probably, that was more focused on the pregnancy side. Very few have cancer genetics training.

Sure. So how much does that cost?

So the test is $250.

Right, but it can be deeply discounted.

Normally, even today, these tests cost anywhere between $2,000 and $4,000, and the side effect of that is that historically, even today, if someone just wants to go and get tested, if they don’t want to pay thousands of dollars out of pocket, it would go through their insurance, and very few people qualify through insurance. You need to have had an extreme level of family history in order for insurance to cover [genetic testing].

You have to be super-sick to find out if you’re super-sick.

[laughs] Yeah, exactly. Once it’s very obvious they’ll ... but it turns out that, for example, half of women who are carriers of BRCA1 or BRCA2 mutations don’t have the family history that would allow them to get insurance coverage.

Oh wow, I don’t think most people usually know this. All right, and how much money have you raised?

We’ve raised about $60 million.

Whoa. That’s a lot of money. From who?

Some of our main investors have been Formation 8, Joe Lonsdale, Khosla Ventures, General Catalyst, Hemant Taneja — who actually, most recently, was in a lot of news with the Stripe investment — and a number of other high-profile philanthropists or people who care a lot about health investments.

… Who want to get into this health care [field]. We’re going to talk about that and more when we get back. We’re here with Othman Laraki, who is CEO of Color Genomics.


I’m here with Othman Laraki, president of Color Genomics. And we’re talking about genetics, genomics and all kinds of health care startup issues. Talk a little about the ecosystem of health care startups. We’ve had a few in here, as I’ve said, but it’s been a difficult area for investment for a long time. And now it’s sort of heating up.

With health care, a lot of investors like to say everything takes 10 times as long and costs 10 times as much because it’s an area where there’s so many other variables to deal with than just building software.

But on the flip side, it’s an area where we can have huge impact on, obviously, people’s lives as well as on the economy. I think it’s one of those [situations] where now there are a few environmental things that have changed that really make it worthwhile to very heavily invest in it. One is [that] we are coming to a breaking point, right? It’s such a heavy part of our economy.

Sure, which people have said for years and they never seem to fix it.

Yeah, maybe it’s like global warming, [laughs] like at some point, hopefully, we fix it before everything breaks.

Probably not.

[laughs] Yeah, I guess — historically.


But I think the other piece is that there’ve been a few changes from a scientific as well as environmental standpoint where, for example, the cost of genetic sequencing has, all of a sudden, made it possible to completely rethink the cost structure of this information. And there’s also the impact of using machine learning in health care.

Which has been the big thing.

Yeah, I think it’s one of the big areas where it allows you to extract signal from this very diffuse data set.

So when you get that data, it’s something we talked to Vic Gundotra about because he’s taking these EEGs, and there’s all kinds of signals there for other diseases through a simple EEG. Talk about that idea of signals, because he felt that artificial intelligence is going to change all of health care and a lot of people both negatively and positively think that we won’t need radiologists, we won’t need all kinds of medical personnel, and that unless you understand the use of artificial intelligence you can’t be a medical practitioner in the future.

I disagree with that, actually, because I think we already use artificial intelligence in a lot of interesting ways today, and you don’t have to be an AI expert. Actually, it’s the opposite.

Right, I don’t think it’s an AI expert. I think it’s that medicine will be relying more heavily than ever ...

Yes, so I think we’ll rely on it, but in fact, I think it will allow people to be able to not need to dig into the data as much. It’s almost like self-driving cars — they allow transportation at a different scale.

With health care, if you look at how we practice medicine today, it’s more of what we call an expert system, where you have a decision tree that you try to make fit into people’s heads, and we call those doctors. If you go see your doctor, ideally you’d like them to be able to use your genome, your family history, your glucose readings from the last year, etc. But all that data, we can’t process it as humans. So the question is, how do we use software to enable us to work with that data on a very different scale?

See, I would like a robot doctor, because I feel like they’d be better. [laughs]

Or a robot-assisted doctor.

No. I want a robot doctor.

Oh, you just want a robot doctor. [laughs]

Only because, you know, we were talking about self-driving cars last night at this big Silicon Valley event, and someone pointed out something very ... I think it might have been Joe Lonsdale, I can’t remember, pointed out that when a self-driving car gets in an accident, all the cars learn about that accident. When a person gets in an accident, just that person learns, and it doesn’t go any further.

So it seems as if ... you know, I’ve had some issues around blood, and I had a stroke a long time ago, and I just have to keep repeating myself to lots of people. And they still don’t have the right information. And I always feel like it should be existing somewhere where it would follow me digitally, and then I would also be subject to what happened to other people. There are patterns, there’s gotta be.

I think it’s almost like there are two separate problems. One is how information is accessed and propagated. I think sometimes, in health care ...

Yeah, they’re scared.

Yeah. Privacy is taken as an excuse to lock data up, but ultimately, we own our health data and we need for it to be used for ourselves.

Well, even anonymously, like if you had this many EKGs or something else, and there’s lots of signals, it could benefit you.

Yeah, and we found actually with Color, everyone who takes a test ...

What happens to it?

They have the option to enable their anonymous data to be used for research.

Do most people agree to that?

Almost everyone opts in.

And then where does that go? What happens?

What we do there is, for example, we contribute aggregate data to a number of databases that are used to classify variants in terms of their impact on cancer and things like that. These things really matter. And, for example, a Color customer, by allowing us to do that, can save the life of someone in France who happens to have the same mutation that can now be classified. That really has a huge impact.

The other piece, though, is just being able to use ... Right now, the way data is used in medicine, I think, is we only use the super-salient signals. You come in with a bleeding wound, that’s the data point, right?


But there’s all this low-volume data that is carrying as much signal that’s completely not being used. Like one example — I say [it] jokingly, but it’s real — is that when you take your car in for an oil change, your mechanic will plug your car into a computer ...

And know what happened.

Yeah. They’ll gather more data from your car than your doctor will get if you go in for cancer treatment. And that’s incredible, right? We have the ability to extract that data, to read that data from our bodies and to be able to use it at scale. And I think we are in that transition time where [we are] building the software for that, but the medical system is also adapting to the availability of this incredible new tool set.

So talk about the idea of privacy. Because that’s always used. There’s always all kinds of regulations, and obviously the FDA has been very — many people think — slow; other people think their job is to slow things down.

But a lot of people actually go to Europe or other places to do some of their regulatory issues, because it’s more open to trying things. Talk a little bit about that issue of privacy. Because Silicon Valley always bumps up against this, like constantly, “Why can’t we?!” kind of thing.

Yeah, it’s a good question. I was discussing this with someone recently. For example, if you take the internet, there’s been a relatively high level of self-regulation by internet companies. And also by organizations that are the privacy watchdogs that apply pressure to ensure companies are also self-policing.

Yeah, how’s that fake news thing working out on Facebook?

I think there’s a lot of soul searching now.

No, there isn’t. They don’t have souls. [laughs].They have to have souls to start with. Anyway, sorry. I was just revealing a personal bias I have, but go ahead. I can’t believe I said it out loud.

But you know, I think with regard to private information, the first big thing that’s happening is that people now are starting to have more and more of the expectation that the data is theirs and that they get to decide. One of the big changes is that we went from the majority of your health data being generated and kept in hospital records ...

Mm-hm, on paper.

Yeah, on paper, with terrible handwriting that even the people who wrote it can’t read it, most times. To a point now where maybe some of the most important health data being generated about you is on your phone and while you’re walking.

Walking, sleeping …

You’ve got your heart rate monitor; with glucose monitors, now it’s becoming easier and easier to track that. Your genome, for example. These are data sets now that people are in control of. And I think that will change the power dynamics.

And people will push for this?

I think so.

Because people get very babyish when it comes to money and medicine. The paternalistic system really does jump into place.

I think the system, or in general, when we talk about the quote-unquote “system,” it tends to be more punished for risk-taking and less rewarded for missing opportunities.

Well, because risk has some downside. You kill people.

Yeah, exactly. But also it doesn’t get punished for the missed opportunities. Whereas individuals are very aligned with not missing those opportunities. For me, knowing that I’m a BRCA2 carrier is very important for my kids, etc.

Important for your kids, for everything else. But there is this paternalistic system that you have to get through. Having been in it — I’m a plague to doctors, because I’m always like ... You know, they’ll give me some answer that’s not correct, or I listen to what they’re saying and I go, “You don’t know, do you?” And they give me some mumbo-jumbo, and I say, "I speak English. You don’t know." [laughter] And they’re like, "Ehhh." And I go, "Just say you don’t know and I’m good with that."

And then you have something like Theranos [where] it comes true. That maybe this doesn’t have the efficacy that was promised. So it’s an interesting dichotomy between Silicon Valley and the medical community where they obviously would welcome more information, at the same time they like the priest kind of attitude.

Yeah, I think maybe it’s worth separating two things. I think one is how do you establish that a product works? And I think for that the medical system has some pretty reasonable standards, you can criticize it but ... So there’s the genes, there’s also, how do you establish a product works? One thing that we’ve done is we’ve been very transparent with our data. And so when we validated our test, we validated it against a very well accepted test and we published that very openly.

Right, which is important.

And that kind of peer review nature of medicine I think is actually quite effective. Another aspect of that is, medicine is a very peer-driven group where there are what we call key opinion leaders where they’re just well-known scientists who tend to have strong opinions that are respected and hopefully in general [laughs] there for a good reason.

So when you work with people that have that kind of credibility, those enable you to establish that you had someone who really can go very deep with you. You passed their sniff test, essentially. And so there are things there that allow companies, whether it’s Silicon Valley or not, to be able to say, “Look, what we do is real and it works and it’s well established.”

I think separately from that there’s the question of control and how it changes the jobs of doctors. I think it’s a big distribution. There are doctors that, when we started Color three-four years ago, we’re still on the side of increasing access is dangerous, etc., and by now it’s like there’s so much data that shows how big of a cost it is for people to not have access that they’ve really swung to the other side.

What is your business? What is your actual business? Selling tests to companies? Or what?

We’re in the business of enabling people to access data from their bodies that enables them to ...

And paying for it.

And paying for it. The model is direct pay. We don’t go through insurance, people just make that choice with them and their doctors or with their employers. And it enables them to get access to information on cancer—where some people have a very high risk of cancer—and then there are well-established guidelines about what they can about it.

Once they find out. So they can only do one test now? Or is there more?

Yeah, it’s a single test. We cover 30 genes that cover the eight most common hereditary cancers.

Such as? Explain.

Yeah, breast, ovarian, colon, prostate, skin, etc.

And so they’ll say you have a proclivity for this versus the population, or what?

Yeah, so we can use a concrete example. There’s a gene called APC where if you have a mutation in it you have an extremely high risk of colon cancer. The average population is about 3 percent colon cancer risk. If you have the APC mutation you literally go up to between 70 percent and 100 percent risk. So you know, the likelihood of [you getting colon cancer] is astronomical. But if someone knows about it — while colonoscopies are not pleasant, they are very effective.

I just had one. I enjoyed it very much.

[laughs] Okay, so you’re maybe ...

I have the colon of a 20-year-old, from what I understand.

Well, that’s good.

I was very proud.

And the good thing about colonoscopies in addition to I guess the enjoyment in some cases is that they’re actually very effective. When polyps are detected early, it could allow someone for example with an APC mutation to literally potentially add decades to their lives.

And so they can do different things, right?

Yeah, exactly.

Actually, Katie Couric made me take it. [laughs]

Oh really? Oh yeah, because she’s been a huge proponent ...

Yeah, she’s a proponent. She screamed at me until I did it. But what was interesting is it was helpful. It was good because I was thinking what could you do then, what action could you take if you had that or something else. Same thing with breast cancer.

Yeah, exactly.

But you don’t give them advice then, right? You have this now, you need to stop doing blank or get a test.

So for each one of these there’s — whether it’s cancer risk or gene — there are national guidelines for you to do if you have a mutation in one of these.

For skin cancer it’s increased screening, obviously increasing vigilance around sun exposure, etc., but what you end up having is you end up ... When people have this elevated risk and when their doctors know about it, they can adjust their screening in a way that increases the chance of catching it.

Awareness. You can get on top of a lot of these things if you can get them early, that’s quite helpful. Do you ever worry about people finding these things out and freaking out? Like having the freak-out? Because I remember getting some of the genetic information from 23andMe and it was disturbing.

Yeah, so a big part of it is for people to be informed ahead of making that choice, so ensuring that people are making a very informed decision.

I don’t know if you’ve met a human being lately, [laughter] oh sure, yeah yeah yeah, ah!

So we’ve tried to reduce the friction there, for example by making videos.

Education people.

... making it easier for them on the education front. That’s also why we require a physician order so there’s always a doctor involved. And that’s also why we include genetic counseling where anyone who takes a test ...

So these are your genetic counselors.

Yeah, they’re our counselors, they’re all board certified, but we’ve trained them specifically around ...

To go through the guidelines or whatever it was, whatever they happen to be. Like what if someone gets one and they’re like, they have all of them or something like that.

Yeah, each person has different permutations and risks so there are certain cases where people have mutations of multiple genes and the guidelines adjust based on that.

Right, and you don’t do things like, “How scary.” Or like, "Twenty percent, you should be a little bit nervous, versus 70 ...” If I saw 70 I’d freak out.

Yeah, in general with all these things ... In my personal case, it’s obviously not good news, but when you know about it you’re like, okay, well, it is what it is, but now that I know, here are the three things that I can do.

So you have this gene, you’re a carrier of this gene. And you have children. Did you know that before?

Yeah, I have two young boys, but I knew about it beforehand. For me, I have increased risk for a few different cancers.

Did you have them tested?

No. At least for these genes currently they’re adult onset which means that there’s no actionability change until you’re in your 20s anyways. So I think it will be their choice.

There’s an interesting thing I saw last night at the breakthrough prizes where someone ... I’m blanking on her name ... but the reason she got it was she was doing genetic testing on a certain gene that created a disease and then you pass onto your kids. And the person knew about it and then they were able to bypass it for their kids, it was really fantastic. So this guy had it, manifested it, but his kids will never get it, which I thought was really interesting.

Yeah, even now, today, you can do in vitro selection of ... and certain people are carriers of ...

This wasn’t that but it was ...

Oh it wasn’t that, okay.

Yeah, it was really interesting. But the concept was ... that’s what people worry about of all kinds of diseases. We’re going to talk a little bit about the future. Do you think it’s a good thing? Because there’s a play called “Twilight of the Golds” where they found out — this woman who had a gay brother found out — their baby still to be born was going to be gay and then considered getting rid of it. This was back in the ‘80s with AIDS issues and stuff like that. And I remember it being questions about what you do when you have all this information, all this genetic information. Where does it go? What are we going to get to if you’re out. In the business it’s going to be all about this. You’re going to be able to digitize yourself.

The way I think about it is, each one of us, we have a torrent of data in our bodies. We’re generating a tremendous amount of data and today we make use of almost none of that data.

None of it, yeah. And often make fun of people. Like there’s all these jokes about people in Japan, they have these [appliances] that do everything.

But I think what happens, at least first of all from an individual self-care standpoint, all the choices that we can make around everyday life choices that can have a very big impact. When people start having a lens into that, it becomes much easier to improve our choices.

Another thing that I think is going to be quite interesting is right now the things that are easiest to act on are things where the actionability is very specific. Like for cancer, people tend to really follow the guidelines because it’s ... you just need to go get screened once a year or you need to go choose to get a colonoscopy or whatnot.

Things that have historically been very difficult are things that are more habit modification. So it’s like, okay, if you consume less sugar ...

So they have information, they won’t stop eating sugar unless oh no, you’re actually going to get cancer if you keep doing this.

And I think part of that is like that feedback loop. And so I think being able to plug into, for example, a continuous glucose monitor, if every time you pop a Snickers bar you get a graph that you see that ...

Yeah, no, I’m a proponent of [this]. I just had some Fitbit or one of them, or Nike was in here and I was like, these are useless because they’re not saying, “You ate that donut, here’s what happened, you need to get up.” And it sort of orders you around as a body, it starts telling you things. But where does that go to? How much information — should you have all that information about yourself? Constantly? Is that a good thing? And where does it go? How much information will we be getting? We’ll have like a monitor on our neck or something like that?

I wouldn’t be surprised.

Or in our eye or whatever.

The one thing that’s likely to be relatively limited is our ability to input that data and read it ourselves, like the bandwidth of human eyes or ears or whatever. And so I think what likely happens is that we just build more tools that enable us to use a lot of data in a human actionable way.

I think part of your question too, though, is okay, what happens with perfect information, do we start making weird societal choices that ...

That’s one of them, and the other is what do we become as human beings? There’s a whole push in Silicon Valley around the perfect — the singularity, the this, the that, the idea that we become computers ...

We just interviewed Elon Musk who was talking about neural networks so we can keep up with computers. There’s all kinds of things like that. And I think health care is where it really intersects. And the fact that we can do this at such a high level and yet most people in this country have terrible health care, across the world have terrible health care. So how does that integrate down to people who don’t have health care or are not well served versus this sort of elite group of people who are monitoring everything they do?

That’s a really interesting question. I think in some ways, if you look at the two extremes, on the hand there’s one extreme that’s for the people who can put infinite resources to do either good or crazy things, right? And do you give superpowers to some people, who would be living eternally, etc. I think that’s always been there, like Mary Shelley’s “Frankenstein,” that myth or anxiety has always existed relative to humans and technology.

I think the thing that has almost always happened has been around technology being a way to democratize progress in a way that’s scalable to everyone. For example, cancer testing. With what we’re doing literally today, 20 years after these discoveries have been made, only about 15 percent of people who have high risk of breast or ovarian cancer in the U.S. know about it.

Right, so why isn’t everybody getting tested?

But that’s where I think we’re going, where it’s literally now economically feasible and profitable to enable every single person to get tested.

Although isn’t your business predicated on people that are already probably going to take good care of themselves to start with?

Um, yes and no. If you compare it to pap smears, right. So there are about 70 million pap smears that are done in the U.S. every year. The positive rate on pap smears is literally almost an order of magnitude lower than what you’d see with cancer risk, but now all of a sudden it’s the same cost. So people are willing to do these things that have from an impact standpoint ...

So you’re saying you don’t have to do those anymore.

Well no, I think the economics around those are still positive, but my point more is that being able to get insight into your cancer risk has all of a sudden become ...


... way way worthwhile for ...

Lots of things. So do you see that iterating downward? Because a lot of things SIlicon Valley does is all about, you know ...

Actually, when I look at our customer base, I mean it is all over the U.S. We literally get people who write into our support mailing list who are like, “Wow, thank you for doing this, I’ve had some family history but was rejected by Medicare. I’ve been trying to save $3,000 for testing for a year and now I can do it.”

It’s very broad. In fact, one of the things we are able to do because of the dramatic cost difference is that whenever someone gets tested or buys a test, we have this program from Day One that’s called the Every Woman Program where a person can contribute to get other people tested who can’t afford it. And so we have a very large number of people in places like, for example, Mohr House or Penn School of Medicine, or SF General, who got access to testing who literally can’t even afford the $250. Whereas when you’re dealing with something that costs thousands of dollars it’s much harder to subsidize.

Do you imagine it ever gets to be that low? Like everybody walks into a Walgreens and gets genetic tested all the time?

Yeah, I think within the next few years it will.

Just walk in.

In some ways, your genome, it’s no different than your lipids, your glucose. And it just used to be much more expensive and all this scaffolding of rationalization has been built around why you should only look at it under these conditions. But in reality it’s just data, it’s just information that we use [just like] a lot of other information and here it’s all of a sudden now this information’s accessible and you can use it to really impact people.

So last question and then we have to go, how does it feel doing this as opposed to working at Twitter — not to denigrate Twitter, but I joke that a lot of companies here are essentially doing assisted living for millennials, [laughter] like a lot of the companies seem ridiculous on every level. Convenient and fun and interesting, but what’s the difference doing something that’s obviously more meaningful.

I think one of the big differences, when you’re dealing with something like Google or Twitter, etc., the number of people you’re touching is in the hundreds or a billion people, and so the magnitude of humanity is incredibly ...

Sure, but assisting Donald Trump to tweet better is not really a life goal that I’d imagine you have.

Yeah, that would be a new one. But I think what’s very different here is that you get sometimes a [glimpse] of one that makes your last year worthwhile. Sometimes you’ll get literally one person and you’re like wow, even discounting everything else, this one person today is alive, maybe wouldn’t have been if we hadn’t been doing our work. And so I think that is a very, very different feeling.

It really changes the feeling of responsibility of what we’re doing as well as our mental math around why it’s worthwhile. And that’s been a huge factor in our ability to build a team like we did. We have one of the things that I feel most fortunate about is we’ve been able to attract to the company just an incredible group of people, some of the best people we’ve ever worked with through all these amazing companies. And these are people that could be getting paid more, but we’re all connected to this in a very personal way.

Yeah, they could all be doing and Uber for milkshakes or something.

Milkshakes are good, [KS laughs] it’s like delivering happiness.

[laughs] Not for your genetics they’re not! [laughter] They’re delicious. Anyway, Othman, thank you so much, it was great talking to you and thanks for coming by.

This article originally appeared on