Founder Stories: Andrew Conrad, Founder and CEO of Verily

Andrew, thank you for being here and escaping Google's labs, secret, top secret labs. I assume it's not easy to get out of there. So I'm assuming why you did come here was to tell me a secret.

Yeah, I actually have a parole officer in the back room there. They don't let me out un—

We only have you for 28 minutes.

OK, yes. So maybe I could tell you just briefly what the division that I lead is. It's called Google Life Sciences. And Google Life Sciences has a sort of an interesting mission. It's at Google X, and Google X's job is to take on big problems and try to find clever solutions to big problems. So one of the big problems that we decided to tackle at Google X was health care. And the way in which we envision doing this is trying to sort of convert the paradigm of medicine, which is currently sort of reactive and episodic, to a new paradigm that's sort of proactive and continuous. And so the way that we've envisioned doing that is by creating a whole bunch of gizmos that some of them you may have heard of, like the contact lens that measures glucose is a pretty prime example of that.

Are you wearing one of those right now?

I'm not wearing the contact lens. I'm not a diabetic. But the contact lens does two things. It actually helps your vision and measures glucose. So it does two things. And today, I think we're going to talk about the sort of a next thing.

So what is it? What is the secret? That's why you come here, right?

Yes, we come here to— they come here to let me test out secrets. So the next project we're going to talk about is something we're calling the nanoparticle platform. And that is a complicated thing. It's a little more complicated than the contact lens. And what we're doing there is nanoparticles and nanotechnology is a sort of term reserved for very small particles, 10 to a billion times smaller than a meter. They're so small that thousands of them could fit within a red blood cell or millions of them fit within a grain of sand. And our idea was to functionalize these nanoparticles to make them do what we want. Nanoparticles are the are the nexus between biology and engineering. So we can make these nanoparticles behave in ways that we want them to do. But essentially the idea is simple. You swallow a pill with the nanoparticles and they're decorated with antibodies or molecules that detect other molecules. They course through your body. And because the core of these particles are magnetic, you can call them somewhere. You can, if you look at your wrist right here, you'll see that there's these superficial veins. Just by putting a magnet there, you can trap them and you can ask them what they saw. The analogy in medicine is like, imagine you want to explore Parisian culture and you do it by flying a helicopter over Paris once a year. That's what doctors do now. And what we're hoping to do is these little particles go out and mingle with the people. We call them back to one place and we ask them, hey, what'd you see? Did you find cancer? Did you see something that looks like a fragile plaque? Did you see too much sodium? What did you see?

So I want to make sure I understand this right, and also I got a C in biology. So for anyone else who got a C in biology, we have this nice diagram up here. So I take a pill. Is it safe to take this pill? What happens to this stuff in my body after you don't need it?

Yeah. Again, the core of these nanoparticles is something called iron oxide. It's pretty ubiquitous and it's really safe. And that's just the little part that gives it its magnetic characteristic. It's coated or painted or decorated, whatever word they like to use in nanotechnology now, with things like antibodies or proteins or molecules that you're quite familiar with. And they're just the vehicle that drives those molecules in a directed way around the body. It's sort of like a self-driving mini car. And it drives it around the body, and then we can make it park where we want. And so there's a lot of evidence that these things are safe. There are already some nanoparticle medicines on the market and many contrast agents when you get MRI, for example, use similar nanoparticles.

Okay, so I take this pill, these course through my body, they look for anything bad. Yeah. And then I wear this device. Yeah. And it, and it can, it tells the information to the device. There's some sort of screen there. It relates to a doctor.

It confesses it up to the doctor, to the cloud, to whoever needs to see it, the doctor, the physician. But essentially, imagine this. So if 5 mL of blood passed through those veins in an hour, instead of going to the doctor and having blood drawn where they take a few mL of blood, this way you could sample hundreds of mL of blood because blood passes, the nanoparticles collect, you ask them the question. And what happens is you're able to ask it once an hour, 5 times a day, whenever you want, so you get a more continuous monitoring instead of this episodic monitoring when you go to the doctor and he, he or she draws blood and says, oh, this looks a little bad, or this looks a little good. In this case, you'll go to the doctor and the doctor will say, well, over the last year, 312 days you were looking good, but this last 2 months you've been showing worse signs for this or that, and we need to correct that.

And when is the period where you would see someone wearing this in their life? Is it for someone like my mom who's had cancer 3 times? Is it someone like me who has a mother who has had cancer? Is it for people who are currently suffering?

Yeah, so the most— the first test cases, because again, I think the central thing that happens at Google Life Sciences, what we do is we partner. So what we're going to end up doing is partnering with someone else to bring this along. We're the inventors of it. Like, we made a few hundred contact lenses, then we licensed it out to Alcon and Novartis who make billions of contact lenses. So we're going to invent the technology, then find a partner. But I imagine what the partner would do would be looking at the test case like your mom. The easiest test case in this is for patients who have had some surgical intervention or chemotherapy and they're worried about recurrence.— it's a super prevalent problem, like women with breast cancer often reoccur within the next 5 years. So they could wear this device, and instead of waiting for a lump or a bump or something bad or pain in their back or whatever the symptoms were, it could say, hey, the amount of circulating tumor cells went from 1 the first week to 5 the next week to 10 the next week. We should think about intervening, changing therapy, doing something different.

And all that's instantaneous?

It's all the time. It's continuous, and it's and it's available all the time instead of having to go to the doctor with an invasive process of having to draw blood and wait for your appointment and park your car. It's sort of automatic. And I think that's— when we were thinking about trying to make devices that people would use, we had to make it simple and automatic and not invasive. It couldn't make you carry a backpack around with a radar dish and ask you to jab a needle in your arm every 4 minutes. People just weren't going to do it. So we have to try to find things that add extra value, like the contact lens actually corrects your vision perfectly and can measure glucose, or this can measure many of the parameters that you would normally get. We hope it can measure many of the parameters that you can probably see by drawing blood, but without the blood draw part.

You'd think that when you think about it, you think it's kind of crazy. Google, you work at Google, Why did you go work at Google, and what is Google's interest in this type of information? I mean, Google has tons of information. Why this information?

Well, to be clear, Google X is a little different than Google. Google X is sort of an unbounded thing. It just has a very simple job, and its job is to take on big problems. And you know, like with cars, we're taking on transportation. With balloons, we take on communication. And healthcare, we just sat and thought, gosh, what's the way to address healthcare? It's just not sustainable. Letting people get sick isn't the right answer. And think about this, in no other complex system do we do that. Like, can you imagine changing the oil on your car when the engine's broken? It's like no one would do that, but that's what we do in healthcare. We wait till we go to the mechanic when we're broken. And that we don't think is fundamentally the right idea. I think what made it why Google is Google X is an amazing place. And on my team, I think the thing that's given us great hope and encouragement is that I have an astrophysicist on the team. And I have a cancer biologist and an oncologist and a cardiologist and an electrical engineer and a mechanical engineer. And it's a disparate bunch of really top-notch people. That I don't think would go anywhere else. And the most interesting thing, these little nanoparticles fluoresce. You can see it through the veins. And they twinkle very much like a star twinkles in the sky. And if you know the spectrum of a star, they can tell you if it's old or new or burning this or burning that. And so the astrophysicist provides really interesting insights into the nanoparticle and the oncologist teaches us about the progression of the disease. And the electrical engineer builds these devices that allow us to measure these things and communicate with Bluetooth and Wi-Fi and all those fun things and send it off to the cloud. So it's really a remarkable collection of people that Google X brings. And I think that's why Google X and not why anywhere else.

And to press you on the point of why Google What's in it for Google? I mean, obviously, all this part of Google X is solving the world's problems, but not until you actually put out a product. So how long is it going to take for this type of project to make it out? And even then, what's the benefit for Google?

So again, right now we're actively looking for partners, just like the contact lens. The contact lens will do good in the world, we hope. Do well. I'm a mom, English teacher, I don't know, whatever.

I'll teach you about English.

Yes, you help me about that. You're the journalism major. Because the contact lens, like we built a couple hundred prototypes and they'll build billions of them and they'll, Novartis will understand how to deliver it to patients and how to deliver it to doctors and pharmacies and all that stuff. We won't do that. We'll—

So you're licensing this technology?

Yeah, that's what we do. And if the products work the way we hope, then obviously there's economic reward for it, but mostly there's we fulfill the central mission of Google X, which is to solve the world's problems, find innovative and crazy ways to solve the world's problems.

Does that mean that eventually, let's take this product and the nanoparticle platform as an example, the information won't flow through Google? It won't flow through Google servers? You won't have any information to sell to marketers or insurance companies?

No, no, no. We have no interest in that whatsoever. Our partner will be handling that. But it'd be like saying that somehow GE is in control of your X-ray. It's just not.

Are they?

No. No, and this is something that we're not—

Are they making a microwave with my X-ray?

Yes. No, this is our partners take care of all that stuff. We're the inventors, the creators of the technology. And they're the disseminators and the marketers and the commercializers. Of the technology. We don't have much to do with it at that point.

So not that preventing cancer and figuring out if we have cancer earlier isn't a big project, but what else are you working on in your secret labs with this team of eclectic people?

Well, I mean, obviously what we're focused on is this fundamental switch. And I really— I know this sounds like a broken record, but we're passionate about this. This idea of switching things from reactive to proactive. I don't think the solution lies in any other central sort of algorithmic transformation. What we're doing is we're trying to provide the tools that make that feasible. If you just think about it, we have sort of a— one of the guys who works for us has a sort of interesting story that really brings this home to us. He's a great engineer named Tom Stanis, and he was riding his bike at Stanford, as 96% of Googlers do, and he was hit by a car. And he gets knocked out, and he wakes up in Stanford Hospital. And they have some good news and bad news for him. Good— bad news, I guess, is that he broke his collarbone, he broke his foot, but he's gonna be okay, except for the fact that they found a lump on his kidney when they were looking to see if the car had caused internal bleeding. They found a lump on his kidney, and that lump turned out to be kidney cancer, renal cell carcinoma, which was in its earliest stages because it had been discovered at a much earlier stage because you have two kidneys and usually don't detect cancer there early. So they cut it out. He was out of the hospital in 3.5 days, limping with a cast and other things, but he was cancer-free. And we thought about this, if that Had he not been hit by a car, he would have come into the hospital with late-stage renal cell carcinoma, had an 8% chance of living for 5 more years. And we figured that we were either going to hit everyone with the car or we'd have to develop—

we have self-driving cars.

Yeah, but we were now going to make them self-targeting cars. And we realized that since we were getting rid of all auto accidents as it were, that was not a practical solution. So we had to look for something different to do. And I think that this— that was the sort of the clarifying moment that there has to be a better way to do this. There really does have to be a better way to do this. And I think the nexus of technology and biology now, if you think about genome sequences used to cost $100 million and now they cost $1,000, that we're in this exponential change. And I think that the timing is right for us to have really, hopefully have really new and important impact in this space. And the community, the life science community beyond Google, our partners like Novartis or Duke or Stanford or the hospitals are all working to try to maybe turn the battleship the other way.

And how long do you think that may take? I mean, obviously this is a problem process and what you've talked about here with the nanoparticle platform, how long till technology like that is in our hands, is in the hands of doctors and patients that the shift in this industry can take place?

Well, it's going to take a long time to get it to its endpoint. But with each increment, we steer the ship in a different direction. And it will be years, a few years before people wear these devices, people like your mom.

A few years? Uh-huh. For what you're announcing today?

Yeah, for what your mom would wear because she'd be the earliest case because they know she's at extremely high risk for recurrence of cancer. And then years after that, people who are just generally at high risk, maybe people with family members, and then people who are smokers and people who are at risk for heart attack. You could imagine as the technology improves that it spreads to a wider and wider group. And what we hope, you know, like if you buy a new car It can tell you if your tires are running flat, if the engine's not tuned, if the— it's amazing how many sensors sit on a car. And why not have those many sensors sit on something far more important than your transportation?

So you're saying in less than a decade?

Yeah, we would definitely hope it's years, not decades until this is deployed.

Wow. Wow. I think we can go to some questions. Jeff is going to call on some hands. We've got one back there already.

One back here, if we get the mic. Hi, Andy. Miguel Helf with Fortune Magazine. I was wondering if you could explain to us at a level that a non-biologist might understand, how do these particles work? What do they detect? What do they see? What do they read? And then to follow up on your last comment that it's going to take a decade, so what needs to be solved for this to be viable? So that's a very complex, multi-part question. Let's see how I do. So imagine the nanoparticles have a core of iron oxide. That's just the magnetic bead that lets us drag it around. And on the surface, we could put antibodies. So an antibody binds to a specific antigen or a protein. So in the case of a cancer cell, there's a protein on the surface of many cancer cells called EGFR, epidermal growth factor. And the nanoparticles are labeled with the detector of it. And as hundreds of those particles stick on the cell, a cell that wasn't magnetic is now magnetic, right? Those bonds will pull it and trap it in the magnet. And then if we ask the little nanoparticles to shake— see, I told you I dance. If we ask the little nanoparticles to shake, they shake very differently if they're bound to a big cell than if they're bound— if they're free, right? They can move very freely in a pulsing magnetic field if they're not encumbered by a big thing 2,000 times bigger than them. And so by knowing that, we'll be able to count the total number of particles there and the ones that are bound and unbound. And then if you count the bound ones, you'll know that, oops, they're bound to something, and how many things are they bound to? Does that make sense? And the second part of the question was, what do you have to do to get that to work? There was comp— you've got to get good partners who are used to both medicines and medical devices. You have regulatory agencies you have to clear this through. You have stuff that has to make it happen. The experiments that need to be done. And it's a course, but it's a pretty understandable course of how new devices and new drugs are developed. I think we had another question here up front. And while that mic is coming over, if you are a little shy, want to ask your question over the app, you're welcome to. I'll pull it right up here. I think the biggest problem of mankind is death. Are you working on death? I couldn't agree with you more. Some of you I know are quite affluent and would think taxes are also an equally magnitude problem, but I think death is the biggest problem. And I think a really interesting thing about what we're doing is we're trying to stay off death by preventing disease, right? And that's the thing for everyone. Most people die of diseases. The cars, I guess, are preventing accidents. But so yeah, I think fundamentally our foe is death. Our foe is unnecessary death because most people get sick when they shouldn't. We have the technology to intervene in some of this and we just should spend more energy and effort on it. But ultimately, yeah, death is our foe. That could be a t-shirt.

That's why my My question to you has been, is the healthcare industry ready for this type of innovation? If we know that we may get sick, or if we catch things earlier, is the industry and are the medical professionals ready to sort of flip that on its head?

Yeah, fundamentally it's much better to not have sick people. I don't think many people like it. If you think about it, if you could keep people out of the hospitals, the hospital system healthcare systems would be more efficient if you could keep them in a clinic instead of a hospital, if you could keep them in an office instead of a clinic, if you could keep them at home instead of in an office. It would benefit everyone. And if you detect disease earlier, the remedies for those diseases would just be more efficacious. Those expensive drugs would just work better, or the surgeries would be more effective. And so I hope that I think the healthcare industry really is ready for this notion of trying to be preventative instead of reactionary and transactional. I think the times are coming. If they're not ready, they should be because it's coming. More questions from the audience? One back here in the corner.

Hi.

So this is fantastic. How do you envision this actually being productized? Is it like a targeted pill or is it like a cocktail of multivitamins with like 36, 45 things that are constantly being monitored? How does this actually— Yeah, so in the early iterations, obviously there's some ones specific for detecting certain types of cancers. If it's your mom, we're going to be worried about what type of cancer we occurs. But the formulations, it's pretty easy to decorate them with different things. It's plausible to have many different formulations that would be tuned, personalized to the person, what they're most likely. If you're a smoker who loves the sun and has 3 relatives with cancer, you probably should be thinking about that. If you're overweight and love, you know, red meat, you should— and smoke— you should probably be thinking about things that about heart attack. Probably think about exercise too. But all of the— so it should be prescribed by a physician and the prescriptions should— we should provide a wide spectrum of possibilities and physicians should decide the best thing for that patient along with that patient. I got a question here via the WSJD Live app from Ken O'Brien. It's a good one. How are these nanoparticles removed? What are the ramifications of having these build up in your body.

So my first question there to you.

Yeah, they're inert, these nanoparticles. They're very inert inherently. And I know you had a pee demonstration, but I am free to say pee. And your kidneys do a good job, and your liver and your bowels do a good job of getting rid of them. So they just go away the old-fashioned way. They don't need to— there's no complex procedure to get rid of them. We've got another question up here in the front. We got a mic. Andrew, can you share with us, I mean, perhaps some example of how you guys are actually testing this? Has someone done it in the world, perhaps in India or China or whatever? Are you guys already testing it on rabbits? Is this— what part of it is still kind of a science fiction and what part of it is really something that you have seen with your eyes? Okay. So that's a good question. We've done 4 families of very important experiments. And yes, I've actually seen all 4 of the experiments work. Because in science, our job every day is to kill our idea, right? Unlike everything else where we're trying— marketing people are trying to make things go no matter what, we're trying to kill things no no matter what. So there's 4 experiments, families of experiments that we've seen that give us a little bit of hope. Sorry, Mike. That give us a little bit of hope. And that first set of experiences, we've been able to functionalize nanoparticles, meaning that we can very, very thoughtfully decorate their surfaces and know the affinity and the avidity at which they will bind to certain types of cells, molecules, or markers. So we can make them. We can do an experiment where you take a few cancer cells and mix them in with a million normal cells in a beaker, pump them through a fake arm, put a gizmo on, and say, can you see? Does this one have cancer cells in it or not? And with that experiment in hand, you can say, hmm, those— and then we can look under a microscope and say, look, the particles are only bound to these cells. They're bound exactly the way we thought. The ratio in which they're bound is what we thought they would be bound, the strength in which they're bound is the right strength. So we call that nanoparticle gymnastics. And we haven't done one experiment, we've probably done hundreds of thousands of experiments exploring the parameters of nanoparticle binding and the characteristics of nanoparticles. I think the next big thing that we have to be able to know we can do is that we can concentrate them. So if If there is only one nanoparticle, it wouldn't be that informative because you'd need to know, well, how does that work. So this concentration notion, by very carefully designed magnetic fields that have fields that are pulsed and separated and can shake, we can really begin to collect those nanoparticles and separate the ones that are free from the ones that are bound. Because that's an important thing to know, just bringing back the nanoparticles home and not having them them tell you anything, it's not valuable. So we've done a lot of experiments demonstrating that. And we still— tons of work to do, but we've been encouraged by these results. The next thing that we really need to do is then can you query those nanoparticles? Can you ask them questions? Are you bound? How many are you? How's it looking down there? What's going on? And we found ways to do that. If anybody's ever shown a laser pointer through their finger, you know how near IR or red light will pass through your hand quite well. Well, the red light will pass really easily into the veins here. And so we can fluoresce these particles and have them twinkle back. And we can count them. We can shake them using radio waves. And we can do something called NMR, nuclear magnetic resonance spectroscopy. She knows that. And we can see if they're bound or unbound. And so if we can query them, then the next thing we have to be able to do is build devices that are portable and small enough that can do that work. We've had some fun trying to do that.

By the way, his watch is not one of those devices.

No, this is just a watch. It tells me what time it is. And the last thing is, I don't know if it was the Wall Street Journal helped us talk about this before, which is this baseline study. Finally, if that all worked, We better not cause mad hyperchondriasis out throughout the world by making everyone paranoid that they're all getting something. So the baseline study, which is this huge study that we're doing with 10,000 people to measure what a healthy person should be instead of measuring what sick people are, we'll be able to use devices like this on folks to say, oh, okay, if you have a few blips here, it doesn't mean anything because you have to calibrate it. You have to have a ground truth. And absent that chain of experiments and that chain of knowledge, it's never gonna be a product. I think we had another question.

I think we've got time for one more.

In the back, anyone? Right there. Great job on this, by the way. So I had a question. If you can basically mark a cancer cell, most of the— half of the medication we take is is actually just to activate our immune system. So if you can mark it, why do we need to measure it? Why can't you just tell the immune system, go after all these types of cells? It's not— has not escaped our thinking. It's a little harder to do because what you want to make sure is that you have to get very careful specificity and other things. But obviously what our partners envision doing is once you can detect them, getting rid of them or dealing with them is going to be a much more straightforward thing to do. The idea of immunity and immune modulation is a key component to this. But you've got to be careful about it, because you don't want your immune system addressing the particles too early, because then they'll go away too fast and not be able to find the rare cells. So it's something we think about. It's a good and thoughtful question.

All right, we've got to get Andy back to the secret lab to starve up death. Thank you. Thank you so much.