MANDY HAMERSTON: Differential pressure is one of the biggest things you have to worry about as a diver. So, anywhere there’s suction because you don’t want to get stuck underwater.

REID HOFFMAN: That’s Mandy Hamerston recounting just one of the many uncomfortable situations she’s plunged herself into during her career as a commercial diver.

HAMERSTON: We did one job, there was a water-intake system that helped with the heating and cooling of three of the skyscrapers downtown, and it’s this big structure. There are four pumps. 

They had to keep one pump running, and they’re going to have it on low. And in my mind, I had pictured it being about five feet, maybe 10 feet tall. We get up to it, it’s like 15, closer to 20 feet.

As you’re getting close with this dark ominous object, we know to kind of stay away from this area.

Being comfortable in uncomfortable situations is what you do as a diver.

HOFFMAN: One way to be comfortable with that discomfort is planning each dive — whether it’s in 30-degree-water beneath 18 feet of arctic ice, a lake filled with poisonous industrial chemicals, or navigating around huge suction pipes that can trap you underwater. 

It’s a skill I admire, especially as often, even seeing just a few feet ahead is a challenge when you’re working underwater.

HAMERSTON: Usually visibility isn’t that great for commercial divers. I’ve had a lot where you put your hand in front of your face plate, and you see nothing, it’s just black, and you can’t even see changes of shadows. You’re like, “Okay, this is going to be a good one.”

HOFFMAN: That’s why it’s essential that Mandy gets as familiar as she can with the dive site before going under.

HAMERSTON: We talk about the job site, what it looked like, what our hazards are, making sure that we’re all aware of potentially where they are underwater.

“Okay, here are my reference points that I need to hit, so I know where I’m going.” Having that internal mind map of knowing, okay, I know I should have three or four steps, and then I should hit this marking. If I haven’t, backtrack, find my point, and then go from there.

HOFFMAN: For Mandy, setting clear waypoints can mean the difference between a successful dive and an aborted mission. But exactly when and where she hits these waypoints isn’t set in stone; she needs to keep them flexible to deal with sudden, unexpected changes in the water around her.

The same is true when you’re navigating the ever-turbulent seas of scaling a business. 

That’s why I believe setting short, flexible waypoints is a powerful tool for navigating your scale journey.

[THEME MUSIC]

Chapter 1: Why you need to set short & flexible waypoints

HOFFMAN: I’m Reid Hoffman, co-founder of LinkedIn, partner at Greylock, and your host. And I believe setting short, flexible waypoints is a powerful tool for navigating your scale journey.

Imagine you’re setting off on a daunting expedition. It could be hiking the Appalachian trail, or rafting the Zambezi River in Zambia. Or it could be a cross-country drive to see the family at the height of the holidays.

The important thing is your expedition is big. Maybe even audacious. And doing it all in one go would be foolhardy.

That’s why an experienced guide across difficult terrain will tell you to set achievable, incremental goals that get you to your endpoint. Otherwise your expedition is doomed.

These incremental goals are called waypoints.

The further your destination, the more waypoints you’ll need — and the more consideration you’ll need to give each one, and the role it plays in moving you toward your end goal.

This is why waypoints are also vital when you’re embarking on the daunting expedition of scaling a business.

I wanted to speak to Chris Urmson about this because, as a pioneer of developing and deploying autonomous vehicles, Chris knows all about setting achievable waypoints to reach an incredibly ambitious goal. In fact, I like to call Chris “the Henry Ford of autonomous vehicles” because he’s so far ahead of anyone else in his understanding of self-driving tech and how to scale it.

I know the story intimately because I’ve been tracking Chris’s journey closely since before he founded Aurora. I was so struck by his vision, his mission, and his plan to bring it to scale that I helped lead Aurora’s first round of funding, and then its listing on the NASDAQ in 2021. I also sit on its board.

But by far the best person to tell you about Aurora’s incredibly ambitious goal — and the waypoints it’s aiming to hit along the way — is Chris himself.

CHRIS URMSON: The mission is to deliver the benefits of self-driving technology safely, quickly, and broadly.

For all of us, getting around is important. For those of us who don’t have the privilege of driving, having that access and freedom that the rest of us take for granted, that is super meaningful. And if you think about the 40,000-plus people who die on our roads in the U.S. every year, the vast majority of those are due to personal vehicles. 

HOFFMAN: Those are just the headline benefits of a technology I firmly believe will be as transformative to society, the economy, and our daily lives as the internal combustion engine, the transistor, and the internet. 

And like any revolutionary technology, many of the benefits are yet to be imagined.

But I’m getting ahead of myself here — time to pull back and set some shorter waypoints as we chart our way through Chris’s story.

Chapter 2: Chris Urmson’s introduction to self-driving vehicles

The Atacama Desert in Chile was an unlikely early waypoint for Chris in his journey, and there wasn’t a single paved road in sight.

URMSON: At the time I was working at Carnegie Mellon as a graduate student, and we were working with this robot down in the Atacama Desert. It was a really neat four-wheel thing. But this robot drove at between 15 and 30 centimeters a second. So if you imagine a little old lady with a walker moving across the Atacama Desert, that’s what this was like.

HOFFMAN: The ultimate destination for this robot — or at least its descendents — was Mars. The Atacama Desert was just a stand in for the extra-terrestrial terrain it would eventually face.

Although the distance between each of that robot’s waypoints was measured in centimeters, it marked a huge leap in developing the technology. It meant that Chris and his team soon had their eyes on their next waypoint, the DARPA Grand Challenge — a race for autonomous vehicles set by the U.S. government’s Defense Advanced Research Projects Agency.

URMSON: The idea was to get a vehicle to drive from Los Angeles to Las Vegas, and to do it at 50 miles an hour and going across the desert. And that just sounded cool and exciting. 

HOFFMAN: But there was far more to it than just showcasing technology.

URMSON: In the Iraq war and in Afghanistan, we were losing more of our young men and women on the supply lines than we were on the front lines. And that this was a way that we could automate moving material around on the battlefield so that those soldiers could better defend themselves. It seemed like a worthwhile, important mission. And so I jumped into it.

HOFFMAN: This time the robot was substantially faster. However, the terrain — and the conditions of the competition — set things disastrously off course.

URMSON: We watched this thing head out in the desert, and it was doing 40 miles an hour, which was amazing, right? You saw this giant tailfin off the top of the robot, over the shrubs and bushes out in the desert, but it went about seven and a half miles out of what was supposed to be 150 miles and kind of ended up high-centered on a berm, wheel spinning, smoke coming out of it. It was heartbreaking to see it end that way.

HOFFMAN: Undeterred, Chris and his team re-entered the competition two years later and won. This attracted the attention of Google founders Larry Page and Sergey Brin.

Chapter 3: How Chris Urmson was recruited by Google

URMSON: I was at Carnegie Mellon, and I just joined the faculty, and Google reached out. So Sebastian Thrun, who is a friend from the DARPA Challenges days, had kind of gone to Google to help start what is now Street View and been talking with Larry and Sergey, and they wanted to do self-driving cars. And so Sebastian and I had been talking about doing something together before he went to Google, and Larry said basically, “Why don’t we do that here?” And so I just become faculty, and this crazy little search engine company that had nothing to do with anything that I was doing was interested in having me come out and help start this thing.

HOFFMAN: Okay, this was 2009, so while Google may have seemed crazy to Chris, it certainly wasn’t little. It had firmly established itself as a leader in search and other services. It was also branching out into mobile with Android. 

Google’s autonomous vehicle research had just started as an offshoot of its Street View mapping. And Google had the same long-term goal as Chris — to make autonomous vehicles an everyday reality. However, the waypoints that Google set to reach that goal were more ambitious than the ones Chris had seen in academia.

URMSON: The objectives at a university are very different from the objectives at a company. And that has a lot of ripple effects. And that kind of difference in incentive structure, that difference in mental alignment means conversations went very differently. Instead of, “No, that sounds too risky or too scary,” it was, “Okay. There’s some risk and some scary stuff here. This is how we’re going to work around it.” Right? And that mindset of, “Okay, let’s go make this happen” from this part of the world that you normally think of as a blockade was like, “Oh, this is interesting.”

HOFFMAN: This difference in attitude to risk highlights an important point about waypoints: waypoints aren’t simply about eliminating risk; they are about maximizing your potential for success at speed. Setting the correct waypoints will help you avoid false starts and wasteful backtracking. But if you weigh them too heavily to avoid risk, those waypoints can become an infinite set of steps that get you nowhere.

As a leader, you also need to take account of the environment you’re operating in.

This is why Chris took the decision to leave the prospect of faculty tenure and embrace Google’s offer.

URMSON: And then we got to work with amazing people, and we got to push the ball forward. And we started to see like, “Oh, this can be a thing,” right? “That this technology is meaningful, and it’s ready to come out of the university. It’s ready to be a business. We’re going to have a chance to change the world with this.” And so that combination of an environment that was excited to make this happen, people that were amazing, and an opportunity was just too compelling.

HOFFMAN: In 2009, Chris made the jump to Google. When he got there, Google’s founders Larry and Sergey already had some waypoints mapped out. 

Chapter 4: How Google achieved their self-driving goals

URMSON: Larry and Sergey set a goal for us to drive 50,000 miles on public roads, and then to drive 100,000 miles on public roads, and then to drive these thousand miles of really interesting roads. And it was actually really cleverly constructed because the first two were really about getting a volume of data. And then the third one was about: how do we understand the breadth of the problem, make sure we’re not kind of overly narrowly solving the problem?

HOFFMAN: This gets to the heart of how carefully planning your waypoints from the start can be a valuable gift to “future you.” Waypoints shouldn’t just be linear points that you mark off on your journey. Each should build towards the next by giving you more experience, more data, or more options down the road. 

That’s exactly what these early waypoints at Google did. They helped test, iterate, and prove out the product. They also gave vital direction in what future waypoints to set.

URMSON: We wanted to drive between Santa Cruz and Monterey Bay, right along the coastline. And it’s a narrow road. Two-directional traffic. And we at the time had problems where the way that the radar would track objects, it would generate these virtual erroneous cars that you thought were going to crash into us.

And so then the self-driving car would swerve. And you could imagine driving down the Pacific Coast Highway, swerving away from oncoming cars would be terrifying. 

HOFFMAN: So Chris set another waypoint to avoid any more terror on that highway.

URMSON: Before I let the team go and try and run that, we set this goal of driving up and down old Highway 101. Because that’s a long stretch of head-on traffic. And we had to do that 10 times before we would go and try the Pacific Coast. 

HOFFMAN: Once they hit this new waypoint, they were confident they were still on the right course.

URMSON: I think we knew it was going to be solvable when we finished that thousand miles of driving. We drove down the Pacific Coast Highway to Los Angeles. We drove through downtown San Francisco and down Lombard Street. And so the breadth of the things we had to tackle meant that we were exposed to pedestrians in the roadway and people cutting you off and making left turns across traffic and merging at high speed and zipper merges and all of those interesting traffic problems that are like, “Oh, geez, you can’t solve that. You can’t solve that.”

HOFFMAN: Notice how when Chris and his team encountered an unexpected problem, they set a shorter — but not easier — waypoint. This is something to keep in mind. Resetting your waypoints isn’t always about avoiding difficult terrain. Sometimes, you want to take that terrain head on, but in a way that lets you learn and adapt as quickly as possible.

In the case of Chris’s team at Google, they were still a long way from fully solving a lot of problems they faced. But by setting realistic — yet ambitious — waypoints, they were discovering a clear route toward eventually overcoming them.

URMSON: While we hadn’t solved them in a truly durable, broadly applicable way, we’re like, “Okay, we’ve kind of understand this problem. We know that there’s ways we can improve on what we are, but we will get there from here.” Within the first couple of years of being there, it’s like, “Okay, there’s hard work in front of us. Getting to reliability here is going to be a real effort, robustness here is going to be a real effort, but this will be solvable.” And that was exciting.

HOFFMAN: By 2016, Chris’s team at Google had achieved 1.8 million miles of autonomous driving. It was a huge achievement, but that year, Chris decided it was time for him to part ways with Google.

URMSON: I’d been there at that point for seven and a half years. It was an amazing experience, right? I can’t thank Google, Larry, and Sergey enough for the opportunity to be part of that. And the team that I worked with as well, right? Amazing people. At some point, I kind of lost confidence that we were going to get there on the business side, that we could build the partnerships we needed to, to be able to scale and deliver the product in a way that was going to have a huge impact in the world. 

Chapter 5: Why Chris Urmson left Google

HOFFMAN: Google still shared Chris’s grand vision. However, Chris didn’t see how he could set and reach the waypoints he thought were needed to complete the journey.

URMSON: As a person who was kind of responsible for that team, if I’m not bought in, I felt like it was going to be very hard to lead that team, because these are amazing people, and they deserve better. And so, I say there’s three things you do in life when you have a business problem, right? You try and fix it. You get in line, or you get out of the way. And I worked hard to try and fix it and then ultimately decided it was time to get out of the way.

HOFFMAN: Once word got out that Chris was parting ways with Google, he wasn’t short on offers from other companies looking to chart their own route in the autonomous vehicle industry.

URMSON: I had the opportunity to go to a number of different tech companies. I had the opportunity to go to a number of different automotive companies. These are all amazing companies. But we would’ve been the side bet. We would’ve been the thing that wasn’t core to the business at all of these places. If you look at the automotive companies, these companies have been around for a century. They’re good at what they do, but what they do is not figure out how to make the car drive itself. It’s not motivate software engineers and hardware engineers to work together to build a brand new technology.

You think about the tech companies. Again, incredible companies, trillions of dollars of value, but this is going to be something on the side that when times get tough, may not be a focus. 

HOFFMAN: Instead, Chris wanted to be somewhere that he could set his own waypoints and reset them often and rapidly.

So Chris set out to build his own team that would share not just his vision, but his views about how to get there.

URMSON: It felt like there’s this moment where if I got the right core together, we could build something special because it was clear that we were still very early in the development cycle of this technology. And so, if we could hit the ground running with the right people, the right philosophy, we could have these huge social benefits, huge economic impact, and build a hell of a company.

HOFFMAN: This would also let him set waypoints that would motivate his team to get to their ambitious goal.

URMSON: One of the biggest challenges we have is that it is one of these kinds of grand problems, which means you don’t get the easy dopamine hit of I shipped the thing today, and I shipped it again today and again. It’s: how do we break it into parts where people can have the victories on this long journey? If it’s a marathon you want to celebrate each mile, not have to wait until the end of the 26 miles. Where can we create internal milestones? How do you create these discrete moments on something that is really a continuous progress of robustness and reliability? Because you can get in a self-driving car today, and it mostly works, and it’s the rub is getting from mostly works to works all the time. I think that’s kind of the challenge and that’s the way we’ve been tackling it is creating what we can of milestones along the journey.

HOFFMAN: This is another important component of waypoint setting: when you do it right, it will become a natural motivating force for your team. Hitting waypoints together will give a clear sense of progress and shared achievement and keep everyone engaged with the mission.

In 2017, Chris founded Aurora Innovation with Sterling Anderson, the former director of Tesla Autopilot, and Drew Bagnell, Uber’s former autonomy and perception lead. Together, they had the expertise and the shared vision to get them to their destination. 

URMSON: Building a car is really hard. These are multi-billion-dollar investments, and they’ve been figuring out how to do this well for a century. So to think that we’re going to solve the self-driving problem and the build-a-car problem, and, in some cases, we’re going to build Uber as well, that feels like a three miracle problem to solve.

HOFFMAN: Of course, Aurora was starting from scratch while its competitors like Google and Tesla were already dots on the horizon. It posed the question: how could they set smart, flexible waypoints that would get them into a position to overtake? 

Faced by this daunting challenge, it was time for Chris and his team to decide upon their first waypoint. 

[AD BREAK]

HOFFMAN: We’re back with Aurora’s Chris Urmson.

If you’re enjoying this episode on the importance of setting waypoints, be sure to bring your friends and colleagues along for the ride. You can do that right now — just hit the Share button in your podcast app.

And to listen to my full conversation with Chris, become a Masters of Scale member at

mastersofscale.com/membership. You’ll be able to hear some things we couldn’t fit into this episode, like Chris’s early adventures in the Olympics of the Mind, a deep dive into sensor technology, and keeping safety at the core of their mission. You won’t want to miss it.

Chapter 6: The origin story of Aurora

Where we left off, Chris had just co-founded Aurora. His goal was still the same — to make safe autonomous vehicles a part of our daily lives. 

Chris could now start setting waypoints that drew on his previous experience creating self-driving tech. And he could do it from a fresh foundation. Previously, a lot of the development had been tied up in the trial-and-error of unforseen hurdles that every new tech faces. Now Chris could take those learnings, plus advances in technology, to set more effective waypoints.

But the challenge ahead was still daunting. Let’s recap how Chris puts it:

URMSON: To think that we’re going to solve the self-driving problem and the build-a-car problem, and, in some cases, we’re going to build Uber as well, that feels like a three miracle problem to solve.

HOFFMAN: Chris recognized that the challenge of making autonomous vehicles a part of our everyday lives was huge. 

Maybe even miraculous. So it was time to start charting waypoints that led to making one of those miracles a reality.

URMSON: As a start-up, I think if you’re going to create outside value, you probably have to shoot for a miracle somewhere, but one miracle seems like the right amount. And so that comes back to focus again for us. So let’s focus on what we do best in the world. Let’s make the self-driving system.

HOFFMAN: It was around this time that I was drawn deeper into Chris’s story and led Greylock in series A funding. One of the things that drew me to Chris and Aurora was the clarity of Chris’s vision and the waypoints he had set to achieve it. They gave me a clear line of sight on how this ambitious mission could be achieved and what each step along the way would entail.

And this is another advantage of clear, achievable waypoints — they reveal your vision and just how you’ll get there to others. This can be the deciding factor in securing financing, forging essential partnerships, and winning customers.

Focusing on the AI system that would drive the vehicles informed the next waypoints for Chris.

URMSON: Well, driving miles on the road, you have to do that, but it turns out that if somebody dies in a car accident every 85 million miles, trying to exclusively use on-road driving to convince yourself you’re sufficiently safe is not actually going to get there. It’s going to be part of the toolkit, but we’re going to have to invest in simulation. 

HOFFMAN: Like any person, more miles driven by the AI made it a better driver. But driving millions of miles isn’t very practical — even for artificial drivers. That’s where software comes in. Aurora developed advanced simulation software that could repeatedly test the AI in extreme situations — the kind of situations that only crop up rarely in the real world.

By partnering with others in the auto, transport, and tech industries, Aurora could get access to increasingly large pools of real-world data. But they needed more than on-the-road experience.

URMSON: Deep learning hadn’t been a thing five, six years before Aurora started, and now it’s such a powerful tool that we need to be able to fold that into our architecture. Let’s think about how with high definition maps, we can leverage that technology and move on from the challenges of having giant monolithic maps to using charted maps that allow us to update them more quickly — all these really interesting technological insights that create a foundation for us to build something that’ll scale.

HOFFMAN: This was a critical insight that may feel like adding miles to their journey, but this waypoint would actually help them form a solid basis to prove out the tech, and to test it to rigorous standards of safety and reliability. Then when the time came, they could roll it out at mass scale.

Aurora set about working toward these waypoints with trusted partners in the automobile, transport, and tech industries. 

Chapter 7: Why Aurora focused on autonomous trucks

Progress was good. But then in 2020, they decided to focus on autonomous trucks.

This may sound like a narrowing of vision and a sudden change in waypoints until you hear Chris explain the reasons behind the shift.

URMSON: The trucking industry today in the U.S. is about a $700 billion industry, whereas ride-hailing is about a $35 billion industry.

So it’s 20 times bigger. The unit economics are stronger, and you have to go less places to make it useful. You can focus on driving freeways and the highways. And that ability to have a bigger market, to have easier scaling we expected, to have better economics, that meant that this was the right place to start.

HOFFMAN: Trucks drive set routes repeatedly on large roads. So setting up self-driving trucks would be faster and generate revenue sooner. 

Sometimes setting realistic waypoints that build toward your goal means you need to take stock of your grand journey. Like Chris, you may well realize that your ultimate goal can itself be broken down into separate waypoints.

So while Aurora’s immediate focus turned to trucks, passenger vehicles and the ride-hailing market remained a waypoint — just a more distant one. 

URMSON: Because we’d architected the system the right way, we knew we’d be able to go and enter the ride-hailing market afterwards. And so it was really strategy tactics on how to build the business and how to scale the business in the right way.

HOFFMAN: This wasn’t a pivot, it was a change of waypoint that would drive scale, and allow Aurora to get to the overall destination far faster.

Waypoints can take you on all kinds of side quests, and that’s fine. You may discover unexpected new ideas or shortcuts to your main goal. The important thing is not to get lost. For Chris and his team, developing self-driving tech for trucks will be a major achievement, as well as an important waypoint on their overall goal of deploying safe, robust, self-driving tech to all vehicles.

URMSON: It’s the same sensors, it’s the same computer, it’s the same software. There are differences, a big truck bends in the middle in a way that a car doesn’t, the cab moves relative to the frame of the truck in a way that a car doesn’t.

But if you architect the system in the right way, you can be encompassing of both of these. 

HOFFMAN: We’ve already seen how a waypoint is not necessarily a shortcut. A good illustration of this in the Aurora story is their decision to use a mix of different sensors — cameras, radar, and LiDAR — to feed information into the autonomous driving system. I asked Chris to expand on this decision.

HOFFMAN: I think it’s worth to linger on having many different sensors and sensor fusion. Because, obviously, Elon, who we enormously respect and part of what he is doing with Tesla says, “I only need, essentially, the equivalent of cell phone cameras and a few of them because the human eyes’ good enough.”

And the Aurora thesis is no, actually in fact, it’s much better to have every possible sensor with sensor fusion. 

URMSON: Even as I drive through the world, you see something, and you’re confused by it. You can’t tell exactly how far away it is. And that’s after millions of years of very focused evolution on the brain that became the human brain.

And so, by using different types of sensors, we’re able to compliment them. We’re able to say, “Okay. The camera is really high resolution and gives me very good color. The LiDAR allows me to instantly measure how far away things are and get the true 3D structure. Radar allows me to see through weather conditions that both cameras and LiDARs might have challenges with and allows instantaneous measure of how fast things are moving.

HOFFMAN: What I want to focus on here is how perfecting these different types of sensors added extra cost, time, and complexity. It’s another clear example of how adding waypoints can slow you down in the short term but is absolutely the right call for your long-term journey.

With these new waypoints set, Chris and his team could map out exactly where to go. But building the tech was one thing; they also needed to map out the waypoints for the business model that would bring it to scale.

Aurora built partnerships with trucking companies, gaining unmatched insight into the on-the-ground issues faced by their future customers.

They also forged powerful, strategic, and mutually beneficial relationships with automakers and regulators to build the right kind of business to get self-driving technology deployed at scale. 

Chapter 8: Inside the Aurora and Uber partnership

Now it was time to build a partnership that would move them closer to passenger vehicles. So Chris reached out to Uber CEO Dara Khosrowshahi about uniting Aurora and Uber’s self-driving car team.

URMSON: From my lens, it made a lot of sense that, together, we could sort through the best technology, build a best of breed stack, and have critical mass talent, while also putting in place a partnership with the world’s number one ride hailing platform, while putting in place a partnership with the world’s number one car manufacturer, Toyota, which had just invested and was working with ATG, that this would be a win all around.

HOFFMAN: From Uber’s standpoint, the decision also made a lot of sense. What mattered to Uber was not owning the unit that developed the tech, but rather that it had access to the best technology for deploying self-driving tech at scale in its fleet. In short, Uber was flexible in the waypoints it was following to get self-driving technology into its vehicles.

Aurora absorbed Uber’s autonomous driving unit, ATG, in 2020. The merger highlights how synching waypoints isn’t just important between partners, it’s important within companies. This is particularly challenging when you are trying to merge two different teams. 

HOFFMAN: It is, generally speaking, super difficult to make a merger work. The really important thing is the ATG people quickly feeling like they’re Aurorans, feeling like they also have their hands on the steering wheel, they’re a participant in the construction of the journey, which they totally are. What were some of the things that you did in navigating that very difficult challenge that other technologists, executives, et cetera might keep in mind if they find themselves in a similar challenge?

URMSON: Trying to be open, trying to have humility, trying to make sure that we were approaching this as equals. We took the teams that had been at Aurora, we took the teams that had been at ATG, and we said, “Okay, you both work on the same problem, write a six pager each explaining what you’ve been doing, teach each other what was good, what was bad, what you’d do differently, what the limitations are, so that we understand the scope,” because there was an awful lot of technology overlap.

Then we said, “Okay, we’re not going to call anybody’s baby ugly.” As we integrate the technology, we’re going to say, “Here’s what matters. The first priority is shipping a product in trucking. Right behind that is shipping a product in ride hailing.” 

HOFFMAN: Getting both teams together to re-evaluate their waypoints as early as possible meant that they could move forward quickly with a shared sense of purpose. However, like all scale mergers, the integration of Aurora and ATG didn’t go totally smoothly. In fact, by trying to keep everyone’s waypoints in line, they actually knocked some off course.

URMSON: There were places where, out of respect for leaders and people on both sides, we tried to create maybe in some cases unnatural breaks in the organization, so that leaders from both sides could be leaders. It turns out that was a mistake because people notice it’s unnatural. Then you have two leaders that are frustrated, instead of perhaps one leader who’s disappointed. We had situations where those two leaders would kind of knock each other out, and we might have lost the person from both sides. If I were to do it again, that’s something where I might work a little harder for simplicity and clarity in the organization.

HOFFMAN: It’s not enough for everyone to have a clear sight of the waypoints and why they’re important to pursuing your mission. People also need to feel like they have input in setting those waypoints, and in flagging when a waypoint needs to be changed.

It can be the difference between feeling like a valued member of a ship’s crew and a stowaway.

It’s also important to set clear waypoints for your customers. Failing to do so can leave them confused or even angry.

This is something Stacy Brown-Philpot learned the hard way when she took over as CEO of TaskRabbit, the platform for hiring people to help with tasks like running errands, moving house or fixing a gate. Stacy brought in a sweeping reorganization of the platform that meant service providers — or Taskers — got more bookings, and users got a far more reliable service.

STACY BROWN-PHILPOT: We tested it in London because they don’t know what TaskRabbit is and how it works. They’ve heard of it. Lots of people were in London and wanted TaskRabbit to be there, but they didn’t know what the product was.

The assignment rate went up, our close rate went up to over 80%, not 50% anymore. We knew that we were on to something, and then we brought it back to the U.S.

We told them the same day that we told TechCrunch and USA Today, and all of our clients, everybody, and that was a mistake because they revolted, and they were upset. Mostly because we didn’t tell them that we were going to do this, some because some people weren’t going to have work anymore, and they were going to have work in a different way.

HOFFMAN: Eventually, that new system was embraced by Taskrabbit users. But if Stacy had clearly signposted this new waypoint, she would have avoided some serious pain.

Clearly signposting your waypoints is especially important if you’re asking people to get on board with a world-changing product like autonomous vehicles.

Aurora has covered a lot of ground in the five years since it was founded. That’s because they’ve stayed committed to setting short, flexible waypoints — and continue to do so. Chris sees such waypoints as essential to framing and addressing the challenges and opportunities facing Aurora and the entire field of autonomous driving.

Chapter 9: How Chris Urmson sees the future of autonomous vehicles

Remember how Chris laid out his vision for a world transformed by autonomous vehicles at the start of this episode?

Now listen to how he clearly lays out Aurora’s current waypoints, and how they will lead to the ultimate vision of making self-driving vehicles a part of our everyday lives.

URMSON: Initially, the focus will be on long haul driving, but over time, we’ll expand that into more and more kinds of urban driving. First, it’ll go between two terminals, then it’ll go between terminals and depots. Ultimately, it’ll go between a warehouse and a store, delivering the goods where they need to go.

HOFFMAN: As well as setting waypoints for the development of the tech, Chris sets out waypoints for how this will improve the trucking industry.

URMSON: It’ll be safer. Because they have this variety of sensors, because they have software that doesn’t get distracted, doesn’t get frustrated, doesn’t get angry, it’s constantly paying attention. Our trucks are going to drive at 65 miles an hour instead of 75 miles an hour. That means a 25% improvement of fuel efficiency. That’s like a huge win for the environment right there. It’ll make it easier for people to get goods. Instead of it taking two days for something to ship from Houston to LA, it’ll take a day.

HOFFMAN: Chris is also clear about the waypoints for a group of people who might, understandably, be concerned about how this will impact them: truck drivers.

URMSON: It’s important to understand that we are dramatically short truck drivers today, something like 80,000 drivers short today, and expect to be 160,000 drivers short by the end of the decade. My expectation is that if you are a truck driver today, and you would like to finish your career as a truck driver, you’ll likely be able to because this technology is going to fill a need that is unmet by the labor supply pool. I think the other part of it, though, is that we’re going to create a whole bunch of new, interesting jobs. There will be displacement over time, that is one of the things that I worry about, but the more I understand about the labor situation and the market here, the more I have confidence this will be a softer transition than many.

HOFFMAN: And Chris also has the waypoints set for rolling the tech out beyond trucking.

URMSON: What you’ll see is transit get better eventually because instead of having these 50 person modules, buses that are driving down the road, you’ll be able to have tailored transit that’s four, six people. It’ll be much easier to get where you’re going, and it’ll raise the level of accessibility for folks at the lowest parts of our economic spectrum and strata. I think that will be magical.

HOFFMAN: I want to highlight here not just how Aurora has hit the waypoints it has set so far but also how it hasn’t hesitated to alter them in service of reaching its goal. Chris started with a focus on cars, then switched that focus to trucks when he saw how it would move them more quickly to the general rollout of self-driving tech. Then, when the Uber partnership came on the table, Chris once again adjusted his waypoints to take account of that huge opportunity.

This highlights another key advantage of having clear yet flexible waypoints. You can rapidly steer toward new opportunities without hurtling hopelessly off track. You can also adjust to changing circumstances in the market, your competitors, and the economy. In short, flexible waypoints let you adapt quickly without losing momentum.

That’s why Chris is leaving room to set new waypoints once the unseen possibilities of the tech become clear. And the dream remains clear and achievable.

URMSON: I think the most exciting thing, though, is that we don’t really know what all the impacts are going to be. It was the horse, and then the steam engine, then the combustion engine, and now automation. It’s going to be incredible to see what happens.

HOFFMAN: This is perhaps the most powerful thing about waypoints: they let us map out our dreams, for ourselves and for others. And this is one huge step toward making those dreams a reality.

I’m Reid Hoffman. Thanks for listening.