16: Keller Rinaudo — Building the Sky Ambulance

Keller Rinaudo, co-founder and CEO of Zipline

Listen to the episode on iTunes.

My guest this week is Keller Rinaudo, the CEO of Zipline, a drone-delivery company focused on healthcare. Zipline provides to Rwanda and other governments the ability to do instant deliveries at national scale and is now delivering up to 20% of the country’s national blood supply. The ultimate goal is to put the 12 million citizens within a 15–30 minute delivery window of any essential medial product they need. Rwanda is an interesting example of a country that could end up leapfrogging real road infrastructure and going straight to autonomous aviation.

This was a fascinating and wide-ranging chat about policy and innovation and the future of infrastructure. It’s an incredible story of how a young band of Rwandans are doing something that has never been done before and having a huge impact on patient health, potentially paving the way for other governments to follow. We discuss how young governments are more able and willing to take risks, and how the US government by comparison has become risk-averse and ossified, and what it would take for us to become a leader in this space. We also get into why an immigration policy that keeps technical talent in this country, as well as access to healthcare, are requisites for entrepreneurship, and why having chill parents with low expectations took the pressure off and allowed Keller to pursue a huge mission. Keller is a professionally ranked rock climber and all-round fascinating guy, so I hope you enjoy this conversation as much as I did.

Keller, great to have you — let’s start with a quick overview of what Zipline is doing. Essentially instant delivery for life-saving healthcare?

KR: That’s a great explanation. Zipline is looking to build instant delivery for the planet and our mission is to deliver urgent medical products to people in difficult to reach and remote places. Today we’re operating at national scale in Rwanda. We’re delivering a significant percentage of the national blood supply on a day to day basis and we allow hospitals across the country to get instant access to any blood product that a patient needs on either a routine or an emergency basis.

AMLG: I read that by this past summer of 2017 you’d climbed to 20 deliveries per facility per week, is that right?

KR: That sounds right on average. It depends on the size of the hospital. There are certain hospitals that are smaller and hospitals that are larger but around 20 deliveries per hospital per week. The most important thing to realize is the hospitals we serve, they only receive blood deliveries via the Zipline system and most of them are receiving multiple deliveries a day.

AMLG: So you’ve built seven of 21 planned facilities in Rwanda which is your first market right?

KR: We’re actually at eight. That number is changing as we’re adding hospitals to the system every day. In the long run our goal is to be serving all healthcare facilities across Rwanda. Something like 40 hospitals. Then there are an additional 400 health centers that don’t do blood transfusions but do need access to a whole host of medical products that are hard to get access to. In the long run the vision of the Rwandan government is to put each of their 13 million citizens within a 15 minute delivery of any essential medical product they could need.

A distribution center worker loads a sachet of blood into a Zip

AMLG: In terms of how it works — the staff in these clinics, they text a distribution center where the workers pack the medications into a box that they then load it into the Zip. Then it takes off and goes and airdrops the payload and returns to the distribution center without having landed at all?

KR: Yeah. Although sometimes what we do sounds a bit weird or like science fiction, people who come and see it are always shocked by the simplicity of it. The surprising thing about this system particularly when you’re talking to doctors who are using it on a day to day basis, it’s so simple. It’s like sending a text message and then instantly receiving the product that you needed to treat a patient or even save a patient’s life. We’re doing that using autonomous electric vehicles. They weigh about 13 kilograms. They fly at about 100 kilometers an hour. There’s really no human involved during the delivery of it. From the moment the vehicle leaves our distribution center to the moment it returns, it’s making all of its own decisions. It’s flying itself to the hospital delivering the product and then returning home. When we deliver it’s important to know we don’t land the plane. The plane is essentially coming within 30 feet of the ground and then dropping the payload in — we call it an air brake — you can kind of think it as a simple cost-effective parachute that ensures that the package falls right on the doorstep of the hospital in a gentle magical way.

AMLG: I’ve watched the video — it looks like a little red shoe box. So you fill it with sachets of blood and there’s a QR code so the vehicle automatically knows exactly where to go. Then it parachutes down in a gentle, magical way?

KR: Yes and you can actually catch the package. If they’re standing out there a lot of times they’ll catch the package because it’s so precise. We can deliver into about two to three parking spaces. The experience for the user is bit like using a ride sharing service — you’re indicating that you need something. You’re getting a text message back saying “Thanks for the order. Zip has been dispatched. It’s 12 minutes away.” Then you get a second text message saying “Zips one minute away please walk outside to receive the package.” That’s it. You do not need anything other than a cell phone to place an order and get an instant delivery of a product needed to treat a patient. There is no infrastructure required, very little training, anybody can use it.

AMLG: It’s probably more reliable than my Uber turning up where I think it’s going to turn up at the time I think it’s going to turn up. Because there’s no humans in the loop unless something goes wrong, right?

KR: There is a human in the loop in the sense that we have an air traffic controller that is in communication with the vehicles at all times and can issue high level commands to different vehicles in the fleet if necessary. But those cases of intervention are exceedingly rare. The vast majority of the time these vehicles make their own decisions, monitor their own health, successfully complete missions and return to the distribution center.

Rwanda is located in East Africa, and mountains dominate the central and western parts of the country.

AMLG: In terms of Rwanda — its a pretty wild, mountainous country. What did you have to do technically to support the navigation system, did you just use 3D satellite maps and pair these with manual ground surveys, or what?

KR: One of the characteristics of Rwanda that made it a good place to start with this technology is that it’s a mountainous country. It’s known as the Land of a Thousand Hills. It can often take, by nature of that topography, a long time. Roads tend to be windy. We actually took open source topographical maps of the country and then loaded those — along with more precise 3D surveys of the delivery sites that we serve — we loaded those both into the navigation system of the airplanes so that when you have a package the vehicle is scanning the barcode of that package and then instantly has its mission. It knows where it needs to go and there’s no programming in coordinates. That’s all done ahead of time because each path is predesignated from the distribution centre to a hospital we serve. Every time you’re doing a delivery to that hospital the plane is flying in the exact same path. This is how we ensure that the system operates in a predictable, reliable, ultimately boring way. Logistics should be boring. There shouldn’t be any surprises.

AMLG: What about the design process, what didn’t you expect — were there breakthroughs in how you designed the system to be this efficient and this simple? Was the aircraft catching method where you hook it on landing what you decided to do from the get go?

KR: When we were getting started this had never been done before. It still has not been done by anyone else in the world. We had no idea if it would work. Some of the more intricate parts of the system — the way we recover the airplanes for example — the airplane doesn’t have any landing gear and we don’t have runways.

A Zip being snatched out of the sky, with onlookers observing the “sky ambulance”

Recovering the plane from 100 kilometers an hour, snatching it out of the air and gently bringing it to a halt is an exceedingly tricky problem. The solutions that we initially tried, we were shocked that some of them worked as well as they did. On the technology side there was always doubt in the back of our minds about whether this was even possible. When we were building it everybody was telling us it wasn’t possible and that can mess with your head.

AMLG: On the technological side or business side?

KR: On every side. The overwhelming advice we got was, this is not technologically possible. Even if it were technologically possible it wouldn’t work reliably. Even if it did work reliably there’s no willingness to pay for it and no need for it in different parts of the world. Even if there were a need for it the technology won’t be able to operate at scale. Over the course of the last three years every year we’ve had to disprove one of those notions. Now I think they are all disproved. But even now people who look at what we’re doing will say, well OK the technology works and it works reliably and it works at scale and it turns out there is a need for it, but only in that country that you’re in. It won’t apply to other countries. That’s the next mistaken notion that we’re working on correcting now.

But in terms of being surprised, whenever you’re trying to do something for the first time in the world there’s a lot of uncertainty from a technology perspective. That first time that you see something that you build work is always miraculous and surprising. The other surprising thing to us is that we knew this was going to look weird. Having an autonomous electric vehicle delivering urgent medical products from the sky in a remote part of Rwanda, that looks kind of crazy.

AMLG: Looks crazy to the locals?

KR: Yeah. We do a survey flight to the hospital before we begin delivering medical products, just to make sure that the route is working perfectly and the numbers look good. And the doctor there was telling me they were having problems because all the patients were climbing out of bed to see the vehicle as it was coming by. The doctors were trying to keep the patients in the beds because it wasn’t good for them to be getting out of bed.

AMLG: Unintended consequences!

KR: Exactly. There is an element of this that is radically different. There’s a magic or science fiction to it. But the surprising thing to me was that after seven days all of the magic and science fiction is gone and the doctors treat this as the most obvious thing in the world. They expect the service, they rely on it and they find it boring. It’s amazing how fast you go from science fiction to this is just the way we do it.

A Zip drops medical supplies at a local health facility

AMLG: You make it look easy but I know how much resistance there is on many fronts to get to this, it’s remarkable. One of my favorite things you’ve said is that the locals call it the “sky ambulance” — it’s obvious, the sky ambulance is coming.

KR: It’s like, of course.

AMLG: In terms of the technological breakthrough — you’re building the drones yourself, you’re designing manufacturing and operating a completely new fleet of vehicles every four to five months. You’re borrowing a lot from existing flight control systems and best practice in the aerospace industry, but how is that rapid iteration working? What kind of changes do you make every four to five months?

KR: Zipline designs the flight computer, which means that we are actually designing the boards and the microprocessors that are making decisions on board the vehicle. We’re designing the overall avionics system. We design the flight controls, which is the math that allows the vehicle to fly. We design the guidance and navigation system, which is how the plane finds its way out to where it needs to go. We also design the air traffic control algorithms and the communication architecture of the plane. We design the airframe and we design the entire distribution centre that needs to be able to launch and recover these planes at high volume on a daily basis in a reliable way. By owning the full stack — one thing most people don’t realize about Boeing is that Boeing is only a final integrator, and when you go and try to build a plane like the 787 it’s this complicated rats nest of subcontractors subcontracting to subcontractors, and that leads to projects being expensive and slow.

But when one small team of hardworking engineers can own the entire system from scratch you can move fast. If something’s going wrong you just turn to your right and say, hey so-and-so you built this system, we need to change it by this afternoon. The other thing that we’ve done that’s enabled the speed of the company is we do all of our engineering, manufacturing and flight operations in the same place. When you step outside there are planes flying. So if a plane looks like it’s not flying right or a component you designed looks like it’s not working the right way, you can design something different, run into the manufacturing shop, get something built in a different way and then test it that day. The rate of iteration is 100 times faster than in a traditional aerospace company.

AMLG: It reminds me of SpaceX and how they rethought the operations and the layout of the engineering floor so they could have that quick communication loop.

KR: A significant portion of our engineering team is from SpaceX and we have learned a vast amount from how SpaceX designed those rockets. It’s quite difficult to design aerospace systems quickly but also in a safe way. Those two goals are in tension with one another. SpaceX has led the way in terms of showing how that’s possible and now we’re trying to show that it’s possible with airplanes not just rockets.

AMLG: How big is your team now?

KR: We have something like 60 people full-time.

AMLG: How do you keep that number capped as you expand the business to other countries and other use cases?

KR: We won’t keep it capped as we expand into new countries. One of the most important things to understand about what’s happening in Rwanda right now is that distribution center is being led by an extraordinary team of full-time Zipline employees, but they are also native Rwandans. The technical lead at that distribution center is an extraordinary engineer named Abdul. The rest of the team are hardworking, super smart, and driven to make sure that the system has a positive impact on patient health, which is our mission.

Each country we go into we hire a full team of people to run the distribution centers in that country. That team is not U.S. expats, they’re predominantly citizens of the country we’re launching in. That said from a headquarters perspective we focus on hiring people who love taking on huge hairy technical problems who can do the work of 30 or 40 engineers at a place like Boeing. That’s the only thing a startup can do to survive in this space. You can’t spend $32 billion dollars developing an airplane like Boeing did with the 787. You have to figure out a way of doing that for five or six orders of magnitude less.

AMLG: The lean aerospace startup.

KR: Exactly. That means people have to own more, move faster, be willing to take risk. And — although this sounds crazy in an aerospace startup — especially under test conditions you have to be willing to crash. If you’re not willing to crash then you’re in this regime of ultra risk aversion that will prevent you from ever doing anything new.

AMLG: In terms of crashes, presumably you’ve had drones that have been downed as you’ve been in operation since last fall. What do you do in those situations and how do you recover those planes?

KR: When I talk about crashes I’m mainly talking about what’s happening at our test site. When we started, especially with early prototypes, we took a lot of risk. We operated on an empty piece of land so it made a lot more sense to go fast. We thought, if the plane doesn’t work right we’ll learn from it and it’s not the end of the world. And we did. Many of those early prototypes crashed. But the learnings that we have for each and every one of those instances allowed us to build safety and reliability into the system in a way that would have taken orders of magnitude more resources if we were trying to guess what was going to go wrong.

AMLG: Have you had any planes crash since you’ve started operating? I presume it has to happen.

KR: One of the big learnings for us from the testing we were doing in Half Moon Bay at our offices was it was important to design an ultimate failsafe into the system. If every other component and safety mechanism on the plane fails you must be able to ensure that the plane is going to come to the ground in a safe and reliable way. So we actually designed a parachute into the plane. It works much like the Cirrus, a high end general aviation aircraft that you can buy today. The Cirrus has a similar system where you have a ballistically deployed parachute. We’ve designed a simpler system into our airplanes such that if all the other safety components of the vehicle fail the plane can pull its parachute and come to the ground so gently that you can catch it. We use that commonly under test circumstances at our headquarters. It’s a good thing we built it because we have experienced significant anomalies, things that we weren’t expecting having to do with the fact that Rwanda is a different environment than the environment we were testing in.

AMLG: There’s a strong monsoon season right, its a rainy country?

Bacha Valley. Rwanda is known as “the land of a thousand hills”

KR: Strong winds, strong rain. It can be very humid. It can be very hot. You’ve got to prepare for the unexpected. As a result these planes that otherwise might have glided to the ground and gotten damaged are able to use the safety mechanism, pull a parachute and come down gently, typically in what we call a secondary landing site. We set up secondary landing sites along the path so that if the plane isn’t comfortable flying itself home, if it’s not 100 percent confident it can make it back to the distribution center, the plane is designed to proceed to the secondary landing site and pull its parachute. That has happened in Rwanda. We work closely with the Rwandan civil aviation authority to document that, learn from it and improve the system over time. It’s important that as you’re iterating and improving this is 100 percent safe for people on the ground. That’s why the parachute is critical.

AMLG: So there’s a double parachute situation, one for the payload and one for the drone as a failsafe. And you’ve said each Zip costs about the same as a motorcycle, so it makes sense to retrieve them if in theory that were to happen.

KR: Yeah we retrieve them. Again this is an exceedingly rare event, it doesn’t happen often. But when a plane pulls its parachute we will go pick up the plane and usually the plane’s flying the next day.

AMLG: Lets get into the economics — you’ve said the business has been profitable from day one and that the costs will come down as you continue to improve the supply chain and as volume goes up. Can you speak to the unit economics and how they will change as you expand?

KR: One of the exciting things about working with the Rwandan government is they think about projects like this differently. A lot of places in the developing world have an attitude of, we want this for free or we want it to be done by a non-profit or we want it to be philanthropy. The Rwandan government’s approach is different. You always hear the president of Rwanda talking about trade not aid.

AMLG: I think there’s a misconception that you’re doing philanthropy, but you’re not — you’re making money on these deliveries right?

With the rapid adoption of mobile phones, many Africans have bypassed state-owned phone companies and banks in favor of mobile payments. 10 year old company M-Pesa, a service that lets you send money via cellphone, now has over 30 million users. M-Pesa processed ~6 billion transactions in 2016. [Source]

KR: Yes that’s important. We don’t make a lot of money on the deliveries, it’s not a high margin business. But it is important to understand that philanthropy doesn’t scale. Sustainable profitable businesses do scale. The best example of this are cellphone networks in Africa and how absolutely transformative that business model has been for everybody’s lives across the continent.

We want to show that it’s possible to use this kind of technology to similarly leapfrog. In the same way that cellphones allowed many countries to leapfrog the absence of landlines we think this kind of technology can show that it’s possible to leapfrog the absence of roads or low quality roads to make fast deliveries in a highly cost-effective way. It’s important to show that this is sustainable and that is the key thing. This is not some philanthropic thing that’s going to go on for a year and when the funding dries up it’s done. This can support itself. It’s economically viable and it can fund future growth both in Rwanda and in other countries across the world.

Population and GDP of Rwanda [Source]

AMLG: It seems that Rwanda has been a great initial test bed. As you look at other countries in the developing world, a lot of local governments are tied closely to the business community, there’s corruption, bureaucracy, long sales cycles. How do you evaluate which markets you’re going to go into?

KR: There was a reason that we chose Rwanda first. The government there is extraordinarily disciplined. They make decisions fast and they’re investing heavily in healthcare and technology. As a U.S. citizen I wish our country could make decisions that quickly and that responsibly. I assure you we can’t.

AMLG: You’ve said its bittersweet that you’re having to start in Africa and that you couldn’t do this right away in the U.S.?

KR: Well at this point I consider myself at least an honorary Rwandan. The people of that country are incredibly deserving and hardworking. The level of ambition of a country that has limited resources relative to a country like the U.S., to say we don’t just want to have the best healthcare system in East Africa, we don’t just want to have the best healthcare system in the developing world, why don’t we just go build one of the best healthcare systems in the world and let’s use technology and robotics and artificial intelligence to get there. That’s extraordinary. But yes it is bittersweet. We’re a U.S. company and the reality is that the same positive impact that this system is having in Rwanda, it could have that same positive impact on patient health in the U.S. There are people who live in remote parts of this country who don’t have the same access to healthcare that you have if you’re living in a city or on the coasts. That’s a shame. We’re essentially waiting for the U.S. government to catch up to a country like Rwanda, in terms of using technology to make people’s lives better.

AMLG: Right. As you said it’s probably a lot about attitudes as well — it takes a special culture and government to say we are going to be first, we’re going to lead. And the U.S. in this case wants to be first to be second and have it proved out. I know the FAA’s line of sight restrictions are still in place but they do relax regulations for certain companies in certain areas to test out. You made a recent announcement with the White House about doing tests in the U.S. — the Navajo reservation delivery or Smith Island in D.C.? Can you talk about those?

KR: As I mentioned there are these places where you can actually look at the data — people who live in rural places in the U.S. have worse health outcomes than people who live in urban cities. A big part of that has to do with access. If you live 150 miles from the nearest hospital and you have a medical emergency the chances of you surviving are worse. It’s not that surprising. Particularly in places in the U.S. where the roads might not be good or where you are remote or where you have Critical Access Hospitals or clinics that may be stocking out of products or may only hold a small number of medical products on hand and not the ones you might need, having instant delivery is an obvious idea. One thing that blows my mind is that we have instant delivery for hamburgers but we don’t have instant delivery for medicine. That makes no sense.

AMLG: The U.S. has certain priorities, burgers are a matter of national pride.

KR: Haha. But the fact that the fast food industry is out-innovating the medical industry is a scary sign of how slow the medical industry is moving. That said, because we’re working closely with the government and with the FAA, we know that the FAA is pushing hard. Everybody does want to encourage this industry. We don’t want the U.S. to fall behind in terms of core infrastructure projects of the future. That said the FAA needs data. So we are now generating as much data as we can to allow regulators across the world to become more and more comfortable and essentially to understand that this is inevitable.

AMLG: It’s a bit like autonomous cars, where they’re collecting tons of data to show the safety of their vehicles on the road. So you are porting all this data that you are currently documenting to be able to come to the U.S.?

KR: We’re not overly focused on the U.S. The reality is that we’re an international company. There are many countries all of which have different healthcare systems and in each case the need is a bit different. As with any new technology or innovation, there are unknowns around what are the effects, how do you regulate it, how do you manage it. That’s true for everything. The most innovative disciplined countries out there are going to be the ones that seize the opportunity first. Of course the countries that seize the opportunity first, that used to be the U.S. The U.S. led the way in aviation for the first half of the 20th century. But it may be different countries in the future. Those countries that lead the way are the ones that reap the outsized benefits of that technology because they’re usually the ones that own the manufacturing base and the innovation base for that technology into the future.

AMLG: I do want to talk hypothetically at least about the U.S. market and specifically the healthcare market. From what I’ve read it seems like there’s two things. One is what you’ve sort of outlined, which is the rural use cases, there’s a high rural population in the U.S. so that makes sense. The other would be things around personalized medicine, so rare immunotherapies, cell gene therapies, biologics that require temperature control, humidity control and monitoring of the shipments while they’re moving through the supply chain. As you’ve talked to health networks and pharmacies in the U.S. What have you found about their response? The U.S. healthcare system is notoriously bureaucratic and it’s one of the glaring areas where there’s been no innovation and people have tried time and time again to crack it. How in theory could you crack innovation and the U.S. healthcare market?

Keller Rinaudo speaks at a conference

KR: I think it has to come from the doctors. A bureaucrat in a regulatory agency having to do with healthcare in the U.S. is going to give you one answer. But a month ago I gave a presentation in front of about 1000 doctors in the U.S. and was absolutely mobbed after the conversation. All I’m doing is saying, here’s what we’re doing in Rwanda. And doctors are saying “wow, here is how I would use this. I just had this case the other day where I could have used instant delivery.”

AMLG: So I have to ask, doctors have a lot of pain points in their business and they’ve wanted change in the system, but pharmacies, hospitals insurance companies, PBMs — there’s a lot of legacy interests there that people haven’t been able to crack through. You can have the enthusiasm from doctors but it hasn’t really worked for other businesses who are trying to change the system. What could you do differently? Could you take advantage of existing supply chains, how could you integrate there, how could you serve hospitals? What’s the wedge?

KR: There are lots of ways. There are lots of healthcare logistics companies in the U.S. For instance UPS has a large healthcare logistics branch. There are companies that specialize in this — McKesson, AmerisourceBergen, Cardinal Health. These are some of the biggest companies that nobody has ever heard of. They are huge, and hugely complex systems that are designed to deliver medicine to health centers and hospitals across the country. The way that you start is you focus on urgent and emergency medical products and you make it clear that if a patient is dying the doctor can get the product they need to say that patients life.

It’s a very simple idea. Even the bureaucracy of the healthcare system is not going to be sufficient to prevent someone someone who is dying or someone’s family from using that system or insisting that it be used. In our conversations with those healthcare logistics companies in the U.S., they’re super innovation-forward because their desire is to provide the highest level of service to the hospitals and health centers that they deliver medical products to. It’s actually not that complicated. It’s not like you’re trying to get a new medical compound. You’re just saying, hey this is 10 to 20 times as fast. If you need a medical product quickly you can get it.

AMLG: And it basically needs to be cheaper than a helicopter which is how emergency supplies get delivered?

KR: Yes. A pretty low bar. We can be conceivably a hundred to a thousand times less expensive than a helicopter.

As of Sept 2015 there were ~1,300 “Critical Access Hospitals” located in rural areas in the U.S. [Source]

AMLG: In terms of how you would get to scale in the U.S. — there’s about 6,000 hospitals. How many of these distribution centers would you need to cover the U.S.?

KR: I think that with about 20 distribution centers you could cover 70 to 80 percent of the U.S. population. When you think about that that’s incredible scale. Because each distribution center is a relatively low fixed cost. We can set it up in two weeks. You could build an instant delivery network for the entire country in six to eight months if the regulatory regime were amenable to it.

AMLG: We’ve mentioned a couple of these logistics companies. Amazon recently made an announcement that they’re interested in going into the prescription side of the healthcare system. They’re definitely thinking about healthcare — and who knows with them with the skunkworks and the silos — but it seems like you’ve got a several year lead over what we publicly know about Google and Amazon’s program. How do you think about potentially competing with those giants not on the burrito delivery but on more critical item delivery?

KR: We don’t focus on these big companies that have corporate incubator projects where they’re building prototypes in a research lab. That’s not that interesting to us. We’re more focused on our customers.

AMLG: And with folks like UPS or FedEx you’re in more of a partnership model?

UPS announces their public-private partnership with Zipline in 2016.

KR: Yes UPS is one of our biggest partners, they have been a key partner in the work we’re doing in Rwanda. They’re hopefully going to be a partner in the work that we’re doing in the U.S. and many other countries. We’re now talking to different logistics companies about working in other countries and helping them make logistics networks automated and instant, which is the future of logistics networks. The writing is on the wall. That is where logistics is going. It’s an exciting transformation that’s going to play out over the next 10 years. Everybody’s interested in that transformation and all of these businesses are going to have to adapt in response to it.

The core goal of our company is to build the best possible product that can save lives and have a big impact on patient health. We do that by spending lots of time with our customers, understanding how the product can better serve them, how can we do a better job of serving doctors and nurses. Then we iterate really fast. Everybody and their grandmother is talking about drone delivery right now. What that usually looks like is someone’s buying an off-the-shelf quadcopter and duct taping the Snickers bar to the bottom of it and then manually flying one kilometer in perfect weather and saying, we did just a drone delivery.

AMLG: I basically do that with my family…

KR: Haha that’s awesome and should be encouraged. But it’s only through intense focus on customers and customer need that you build the right product. That’s what we’re focused on in this company.

AMLG: Yeah it seems like you’re really in a concrete use case and getting real learnings and real data and making real progress. But moving from the specifics I have a 10,000 foot question. You mentioned the future of logistics and you’ve said in the past that if cars were invented today they would be illegal. In the U.S. we’ve got legacy infrastructure and it’s hard to do that leapfrog to whatever step might be ahead of us. If you were designing the logistics system in the U.S. from scratch what would you do?

KR: It’s something we give a lot of thought to. It’s hard to predict two years out versus even 10 years out. But a couple of things are becoming obvious. When you look at the trend of how logistics has worked over the last 100 years, things are getting faster. That’s that’s obvious to anybody. We use to deliver mail using horses. So things are getting faster. And things are getting automated. For example if you send mail now we use the Internet rather than a horse. If you’re sending a package that package is going through a highly automated sorting facility at a UPS or FedEx warehouse.

We think that as technology is increasing those two trends are becoming more inevitable and are playing out to their final state. The obvious final state would be that when you order something you get it instantly and there aren’t humans involved every step of the way, physically handling whatever it is that you ordered. That’s the 35,000ft view. But from a more specific view of what’s happening today. Instant delivery is being used by millions of people across the country. There are many large businesses that are scaling quickly. Obviously Amazon is doing this, Instacart, eBay Now, Google Shopping Express — all of these companies are trying to figure out how you can get someone who’s buying something online what they order as quick as humanly possible. But the way that we’re doing instant delivery, the overarching technology platform that we’re building on top of, is teenagers driving 2000 pound gas combustion vehicles to deliver you something that weighs two or three pounds. That’s a crazy solution. If an alien landed on the planet today and looked at that solution it would make no sense. The only reason that makes sense is in the context of how we’ve done logistics for the last 100 years. When you think about it from a physics first principle perspective, the idea that you have a human sitting in the front seat moving the steering wheel back and forth and a small package sitting on the back seat. It’s just weird.

AMLG: From first principles humans don’t make much sense with anything at all

KR: Well hopefully there are some things that humans make sense for, otherwise we’re going to have a hard time staying busy. With logistics you’re doing the same thing over and over again. Its one of the most natural things that should be automated in the world. Not only that but if you’re delivering a package instantly that weighs three to five pounds that means you can’t batch process. So the vehicle is going to be carrying just one package. If the vehicle’s carrying one package and the average weight of the package is three or four pounds, it doesn’t make sense to use a vehicle that weighs 4000 pounds.

AMLG: Lets get into your background. You’ve always had this interest in robotics. You went to Harvard, studied biotech — I looked up a publication you had in Nature in 2007 — you must have seen like 20 years old when that was published?

KR: I think I was 19.

AMLG: I barely understand the title, “A universal RNAi-based logic evaluator that operates in mammalian cells” that’s the title of your paper from when you’re 19??

As a 19-year-old student at Harvard in 2007, Keller became one of the youngest ever published authors in Nature Biotechnology.

KR: That’s science speak for a cool computer made of DNA that can recognize cancer. The idea was that we could use DNA to actually do computation in cells, so it can behave like a doctor on a cell by cell basis and it can make diagnoses and then take actions should a cell in your body be unhealthy.

AMLG: Another massive frontier there. You’ve said you realized that getting products into people’s hands in biotech and pharma takes a long time and you wanted to do something that had impact faster?

KR: Yeah so I chose aviation — in retrospect I’m not sure that was the smartest decision. Also a heavily regulated market.

AMLG: A lot less pipettes though, pipettes can really drain morale.

KR: I’ve done my fair share of pipetting. The other thing that I did at school was a bunch of friends and I built a climbing wall. That was an interesting exercise in trying to convince the Harvard administration, which is not known for taking risks with their 400 year old buildings to build this weird project that this freshman felt strongly about building. That was the other side of my experience at school where I found it was cool to build something physical that clearly had an impact.

Zipline co-founder and COO Will Hetzler met Keller at Harvard.

AMLG: You did that with your current COO who you went to Harvard with?

KR: Yeah exactly.

AMLG: Did you just stay in touch from those days or did he come straight to Romotive with you?

KR: We stayed in touch and eventually I convinced him to quit a cushy consulting job and come join us.

AMLG: What were you like as a teenager. In high school were you also putting rock walls into your high school campus?

KR: I went to high school in Phoenix, Arizona. I went to North High School, a public school. I think we were classified as a failing school under No Child Left Behind for my entire tenure there, which was an interesting experience. I wasn’t that academically focused in high school weirdly. I was working a 40 hour a week job at a restaurant. I learned more working in that restaurant and figuring out how to get a lot done and how to keep customers happy when they were cranky that is relevant to my job today than I did in high school.

AMLG: So you weren’t tinkering in a basement with robotics, you were learning how to get stuff done?

KR: I would say so. I was also fascinated by science and always knew I wanted to be a scientist. But high school in Phoenix Arizona is probably a different experience than high school in other parts of the country.

AMLG: As you think about kids today and how we can inspire more people to build at the frontier, are there any takeaways from your childhood or what got you to Harvard and what got you into entrepreneurship?

KR: I think it comes down to the fact that I had parents who cared about me a lot and didn’t give a shit what I did. That’s so empowering instead of having “get a job get a job” or “make sure that what you’re doing is respectable and don’t fail, don’t embarrass us.” Many of the things I did were potentially very embarrassing to my parents.

AMLG: Like what?

Keller is one of the top rock climbers in the country

KR: For example when I graduated from college — it may have been the case I was slightly brainwashed, I went and worked for a consulting firm for two and a half months. I hated it and left. Then I lived out of my car for six months rock climbing. I got sponsors and was professionally rock climbing for eight months and my parents at that point were probably wondering, OK what was the point of going to college. Is that formal education being put to full use. But they were super supportive and it was only in messing around and hanging out that I was able to get interested in robotics and it became obvious to me that wow this would be awesome. I don’t have anything better to do, I don’t have a full time job, so why not go and start building these robots and learning about it. The more I learned about it the more fascinated I got. Doing something new, particularly doing something as weird as what we got started doing, there’s a permissiveness of unemployment maybe and low expectations. People who are really driven, it’s like you gotta get to that next step in your career, and you end up being taken advantage of by institutions that prey on those kinds of personalities.

AMLG: It’s like Peter Thiel’s thing about why people need to drop out of the Ivy Leagues, because you get on this track and there are these hoops and you have to jump through the hoops and it never seems like you can do something unconventional.

KR: Totally.

AMLG: What would you say is the most surprising thing people don’t know about you?

KR: I’m much less interesting than everybody else at Zipline. I’m also a small part of the team that has built everything that you see when you go to Rwanda. I don’t know if anything about me is that interesting.

AMLG: OK but what do you think about in your spare time?

KR: I’m also obsessed with the educational system in the U.S. One of the things I think about when I spend time in Rwanda and see how disciplined and talented the workforce is there, is potentially how screwed the U.S. workforce is. There might be a sense of entitlement in this country — that we are entitled to jobs, we’re entitled to be able to do the same thing for our whole lives and not have to learn something new that people in other countries do not have.

AMLG: A loss of scrappiness almost?

KR: Yes I would say so. One of the most interesting things I’ve been thinking about over the last year is— I’m reading a biography of Benjamin Franklin and when you read about the founding fathers you see this incredibly entrepreneurial spirit. In the early days the U.S. was a startup and was a very entrepreneurial country. Other countries are taking on that mantle now and are a lot more entrepreneurial than the U.S. It’s cool and really good for those countries. But as a U.S. citizen I would love to see us head more in that direction, more in this spirit that we were in in 1776, when there was so much uncertainty, risk tolerance and knowledge that we have to stick together and do new things. We do not have that right now.

AMLG: The Mark Zuckerbergs, the people that break out with their startups, do you think they’re exceptions in their willingness to take risk? Or is it more on a systemic level that the systems don’t allow for a lot of risk taking?

KR: The majority of venture capital in the world is still invested in the state of California, which is a crazy example of the power law at play. So yes it’s still the case that the majority of the valuable technology companies are being built in the U.S. but my point is I’m looking at the rate of change. Not at where we are. I think that things are changing and the U.S. is losing its place as the leading entrepreneurial country in the world.

AMLG: Is that fixable?

KR: I don’t know actually. It seems so easy to fix yet so impossible to fix. What’s crazy is it’s probably obvious to at least 80 percent of the country to implement reforms that would make it easier for people to take risks. Two big things are, firstly making it easy for highly talented technical people to come to the U.S. Another thing would be not kicking engineering or science PhDs out of the country once they complete their PhDs in the hard sciences, which is a crazy thing that we’re currently doing. Another thing is making sure people have access to healthcare so that you can afford to do what I did. The reason I was able to start Zipline was I was on my parents’ healthcare plan as a 24 year old. Had I not been able to do that maybe I would have kept my job so I didn’t lose my health insurance. There are so many obvious things that 90 percent of the country would agree with. Yet the political situation makes it impossible.

AMLG: Do they have that in places like Rwanda? In China there’s not a social safety net and that’s why there’s such high savings. But they’re still a very entrepreneurial culture.

KR: It’s obviously a more complicated question when you look from country to country, but certain countries, their governments are younger and haven’t been around as long and haven’t had as much time to ossify and become dependent on the processes and rules rather than the spirit that gave rise to those processes and rules. Maybe the U.S. is now more focused on the rules than we are on the spirit that the country was created in.

AMLG: So what do you think about draining the swamp or throwing a bomb into Washington to shake it up?

KR: Oh boy I don’t know how to answer that question. The most important thing to me is that we let people vote and that we use voting systems that are representative of people’s desires. The scary thing is that there are large groups of people in the U.S. who are potentially being disenfranchised and that the voting systems aren’t doing a good job of capturing what the actual U.S. citizenry want. Anyway it’s a much more complicated discussion.

AMLG: Tangential question — if you had several billion dollars what would you do with it?

In August 2017 Zipline announced expansion to Tanzania. [See the full announcement]

KR: The cool thing about innovation today is you do not need a billion bucks. I don’t think we could successfully use that much money right now as a company. Capital markets are so crazy that if we did need that money we would have access to it. But we are more focused on perfecting the product and launching in a disciplined way with new countries that are excited about leading the world in autonomous vehicles and saving lives. We are focused on a couple of new countries over the next six months, both getting fully to national scale in Rwanda and starting to show this can work in other countries in East Africa as well as other middle income countries in the world, with the goal that eventually we are able to build this kind of infrastructure in the U.S. as well. The most important thing to understand is that when we launch in a country, people are relying on us with their lives. So instead of hyper growth and being OK if it breaks, we’re more tempered. Yes we want to grow aggressively but it’s important that this infrastructure always works. When people are depending on it with their lives it’s not acceptable to say oh our servers are down.

AMLG: Reliability is pretty high up there.

KR: Exactly. So we are balancing. We’ve got to make sure people can rely on this with their lives.

AMLG: I spent a month in the jungle in Papua New Guinea a few years ago. It strikes me that there’s a lot of similarities — getting up to the highlands took a long time, the roads are definitely not safe physically or from bandits and it’s a mountainous country. I wonder if that would be a country you consider?

KR: We’re really interested in Southeast Asia and countries that are archipelagos are perfect examples — it’s so hard and so expensive to move stuff around in archipelagos. It’s a good example of where this kind of technology can have a huge impact.

AMLG: If you could see any new technology in your lifetime what would you most like to see?

KR: I would love to see us get rid of cars. Cars are really dangerous and it would be cool if we didn’t see hunks of metal traveling past your house at 50 miles an hour. The kinds of vehicles that move humans around, you want them to either go up or down, to go into tunnels or into the sky. I’m excited for both those changes.

AMLG: So Hyperloops underground and VTOL aircraft?

KR: And flying motorcycles above.

AMLG: Clear up the streets for the humans.

KR: Yeah — it’s awesome when you don’t have noisy, gas combustion, dangerous vehicles. Larry Page pointed out that 50 percent of the area of the city of Los Angeles is parking lots and roads. It’s amazing to think about what can happen in the places we live if you can reclaim that space. Also, my girlfriend is doing her PhD in computational genomics at Stanford, so I still get to live vicariously through her on thebiotech side and it’s exciting to see the changes that are coming out of that industry. Its going to make human life dramatically better. For sure life extension is possible, but I also think its possible to radically reduce morbidity so that ageing is a better process. That can have a huge impact on the rate of progress of humanity.

AMLG: Have you followed the CRISPR advances?

KR: I’m very familiar. You don’t appreciate how powerful CRISPR is unless like I was 10 years ago you were doing experiments genetically modifying organisms or constructs. It took so long and was so complicated. Now it’s as easy as editing a word document. It’s incredible.

AMLG: It’s terrifying but incredible and powerful at the same time. Any final thoughts to share?

Some of the Zipline Rwandan team

KR: The important thing at the heart of this story and the thing I constantly return to and that motivates me is the amazing team of people at the distribution center in Rwanda who are working their butts off and having a tremendous impact on patient health and doing something that’s never been done before. The fact that they are the first ones in the world to do it and the fact that this small government with limited resources is the one that’s taking the risk and leading the world in terms of showing how to use drones to save lives instead of kill people, is extraordinary. I hope lots of other countries will use Rwanda as a role model in that respect.

AMLG: All these people who said this is not possible — how have you dealt with that? A lot of founders I talk to, especially those in frontier areas, face the same thing. What’s your mental process, how do you have the strength to power through?

KR: I just go visit Rwanda and meet with doctors and patients.

AMLG: That’s some serious motivation right there. Keller — thank you so much for coming on today and telling us the incredible Zipline story. It’s exciting to hear what you’re doing and I can’t wait to see what happens next.



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store