Podcast Episode #2: The Approach of Autonomous Trucking
You can find a summary of the podcast and the links here.
Here is the full transcript for that podcast.
Mush Khan: All right. Well, welcome everybody to The NOW Network podcast today. We’ve got some incredible guests. We’ve got Allan Rutter, who is with the Texas A&M Transportation Institute and Dan Goff, who’s with Kodiak. Today we’re going to talk a lot about what’s happening in freight as it relates to automation and autonomous vehicles. What’s happening from a big policy standpoint and from an overall trend standpoint. I’m really looking forward to today’s conversation.
Before we dig in. I’m going to turn it over to each of you. So, you can do a quick intro about yourself and your organization and then we’ll jump in. So, Dan, why don’t you to kick us off with your intro.
Dan Goff: Sure. Hi, I’m Dan Goff. I’m the head of policy at Kodiak Robotics. Kodiak is a startup dedicated to putting self-driving long haul trucks on the road. We were founded a little over two years ago by a guy named Don Burnette, who was really one of the original pioneers in the self-driving industry. He was a grad student at Carnegie Mellon in the late 2000s when Google kind of swept through and said anybody who’s interested in sort of trying to put the research you’re doing in the practice. Come with us.
He spent about five or six years at Google and started to really think that Google was sort of working on the wrong use case. And that’s the sort of traditional view all of us have of these futuristic robo taxis, passenger vehicles that take people wherever they want to go was actually the wrong use case for the technology he was developing.
And that it was much better suited for trucks and the reason for that is the intuitive reason that when you’re driving around and in a city you see a lot of weird stuff happening on the roads, very unpredictable driving environment, you know, as we all see in our daily lives. You have cars, and you have bikes and pedestrians and pets and all of these things that are great and are why living in cities is fun, but it also really, really hard for computers to understand.
Whereas for a truck, if you’re looking at basically long haul truck that’s mostly driving on highways. You’re talking about one of the most predictable highway driving environments there is. Everybody really should be going in the same direction. People sort of act much more predictably on highways, with the occasional crazy thing happens. But it’s much more predictable and importantly, what you should do when something weird happens is a lot more obvious. If you see a pedestrian 100 feet ahead in the middle of you know the street you live on, the right thing to do is almost always to keep driving because that person’s crossing the street jaywalking. But every once in a while, that sort of that instinct is wrong. Whereas if you’re on a highway. If you’re on I 45 and you see a pedestrian 100 feet ahead of you in the middle of the lane, you should stop where you’re going.
Given sort of this increasing interest, he decided to pivot his focus from passenger cars to trucks. He was one of the founders of one of the first self-driving trucking companies, company called Otto, that was quickly acquired by Uber. Became head of software lead at Uber and then left in April of 2018 and found Kodiak. His co-founder was a VC investor guy named Paz Eshel, who was looking at the space and kind of came to the same conclusion that he did.
He didn’t see any companies that he thought were in the right trucking space. So together they founded Kodiak and in just over two years, we’ve gone from sort of the two of them founding the company to a team of 75 people and we’re based in Silicon Valley, but we actually do most of our testing and deployment in Texas. We deliver freight daily between Dallas and Houston. Have a fleet of trucks out of our South Dallas operations hub. So, we think Texas is going to be the first place we see these technologies on the road and if you are in Texas, you can actually see our trucks on I-45 on a daily basis.
Mush Khan: Great, a couple of points. Number one, you mentioned sort of the unpredictability of in city driving. I’ve got a 17-year-old son, so he’s part of that sort of weird ecosystem of unpredictable drivers. And Dan, I think you’d mentioned to me on the phone the other day that Kodiak has already made a couple hundred deliveries between Dallas and Houston. Am I remembering, that right?
Dan Goff: I think we hit 400 last week. Yes. So, we started last July, at the end of last July. So, it’s been just about a year and our model is to focus just on the highway portion. So today we have a safety driver and what we call right seat operator or System Operator. So, two people in every truck. There’s a joke in the autonomous vehicle industry that you can tell that the vehicle drives itself when it has two drivers in it at all times.
So we have our safety driver drive to the distribution center, picks up the load, drives to the highway, presses play, and the truck drives itself until it gets to Houston and then they pull the truck off the highway and do the final delivery.
Mush Khan: Thanks, Dan. Appreciate that background on Kodiak and certainly an exciting journey after two years it’s amazing the progress that your team has made. Allan. How about you, tell us about you and the work that you’re doing right now with the Texas A&M Transportation Institute.
Allan Rutter: Howdy. Thanks for having us on. Texas A&M Transportation Institute just celebrated its 70th birthday. We got created by the Texas legislature as an applied research organization. We’re affiliated with Texas A&M university. Out headquarters is just outside of Bryan.
Have about 700 people who are full time researchers and grad students doing everything from roadsides to pavements to driver behavior to transportation planning and performance.
Do a lot of work under contract to public and private sector sponsors. I joined TTI about seven years ago. I head up a division of freight and investment analysis and help coordinate our freight practice across the organization. Previously, a long history within the public sector working around the Capitol in Austin for a long time, including working for different governors. I was the head of the Federal Railroad Administration for President Bush and the head of the North Texas Tollway authority for a couple of years.
Seven years is about as long as I’ve lasted anywhere, so I’m really looking forward to pressing that a little farther with my TTI boys. We are doing work in freight. Working on helping public sector understand how freight works and what they need to do to plan their transportation networks accordingly, and the whole business of what’s going to happen with automated trucking is really important to people who own and maintain our highways.
Mush Khan: Great. Allan, thank you very much. So, you let off of the Texas A&M greeting of howdy. I think that’s a requirement for anyone that works at A&M.
Allan Rutter: I’m a diversity hire at TTI. I went to the school in Austin. And because I’m a native Texan. I’ve been saying howdy for a long time. I didn’t realize it was a registered trademark of the Aggies. But I still use it to fit in.
Mush Khan: Tell us about some of the bigger trends that are happening in the area of freight in particular what we might say sort of more of a commercial or industrial freight world and what are the things that are sort of top of mind at the institute today and also with you.
Allan Rutter: Well, I think what we’re trying to figure out is how to keep abreast of the pace of change in technology, particularly in the freight space. I think you’re seeing an awful lot of increase of application of technology and use of it everywhere on how freight moves. Warehouses and distribution centers are being automated. There’s an awful lot of encouragement to think about the driving function, how to make that safer and frankly, how to leverage a limited driver pool to get more out of those folks.
The promise of vehicle automation is to not only have some benefits to those that labor, but also to make it safer for everybody, safer for those drivers, safer for the people on the highway. How people who own the asset are maintaining it? What do they need to do about that physical footprint of the road to make it available and then to fit some of that automation that’s happening?
So, a lot of what we do with our public sector sponsors is helping them understand how fast things are moving. There’s an awful lot of interest in freight automation. The fleet is smaller than automobiles. Like he said the primary activity of highway driving is a little easier to think about how to automate that in advance in some driver’s systems. So, part of our job is to help people get ahead of or at least stay consistent with what’s happening in the industry.
Mush Khan: Great, thanks Allan. Dan, what about you, how do you and the team at Kodiak view sort of the macro changes that are happening in the macro trends that are happening?
Dan Goff: Sure. I think, I think the first macro trend that I think Alan kind of alluded to is that all of a sudden, maybe not quite all of a sudden, that people are really paying attention to freight automation.
You know, when we were founded, again sort of little over two years ago, self-driving trucks were kind of considered less interesting, maybe more of a backwater than robo taxis. But in the last two years. The industry has really come around to our point of view and our view of the world and understanding both, you know the importance of the technical advantages to tackling this problem. First, the power of the business case and sort of the broad range of benefits that come from freight automation.
So, we’ve seen Waymo, which is Google’s self-driving successor and part of Alphabet announced in the last maybe four or five months that they’re going to be focusing on freight alongside robo taxis. Aurora, which is another of probably the best known of the independent self-driving developers announced as sort of a new focus on freight as well. So, we’re really seeing people paying attention to this technology for the first time, which we think is great. I mean, I think it’s certainly validation of the importance.
Additionally, we’re seeing a cross of technology types and a bit of increasing interest in freight all together and in perhaps a divergence in the kinds of technologies that we’re seeing or multiplication. So, we’ve all seen sort of the spectacular growth and interest in electrification, but mostly for first and last mile delivery. But increasingly, I know there’s talk of how do you electrify over the road trucking as well. And we’re sort of seeing all of these new freight technologies come in at the same time as well as what’s happening in warehouses, automating what happens in yards and all those kinds of things. We’re seeing just this huge proliferation of these technologies and I think that’s in the medium to long run it’s going to be really great for consumers. I mean, first of all, efficiency is good for consumers, there’s no question, but also for the average drivers.
There are going to be a lot of benefits to our technology. Technology as well, you know, our trucks won’t speed. They’ll never drive drunk, they’ll never going to look at their phone. They don’t care if it’s three in the afternoon or three in the morning. So, we’re hoping that it affects the average people on the roads. When these kinds of automation technologies start to proliferate is you will actually see fewer trucks on the road. The reality is that very few people love driving next to an 18-wheeler on the highway. Reducing the need to have those trucks driving down major interstates and during rush hour will be a great advantage to a lot of people.
Safety is incredibly important. And I think we’re not alone in this. We are building what is called the safety case, which is basically a front to back argument that our vehicles are safe, and our goal is to at least make a very comprehensive argument that our trucks are safer than a human driver and that’s going to be really great for drivers as well.
Mush Khan: Dan, let’s dig in deeper to the point of fewer trucks on the road, I guess, at any given time. So, tell us more about how autonomous trucking and just sort of technology applied to the freight world in general is going to produce that result.
Sure. One of the core safety regulations, a really important safety regulation that the federal government enforces today is to maintain the safety of trust of what’s called hours of service regulations. Which basically say that a truck driver can only drive 11 hours a day out of a 14 hour workday. And that’s to make sure that truck drivers are well rested and aren’t fatigued when they’re on the road.
You know, our computers in addition to not driving drunk will also not generally speaking, not getting fatigued. So, the utilization rates of our trucks are at least significantly higher than with human driven trucks. So, the actual sort of overall need for trucks will probably drop to some extent with automation and additionally we’re not going to optimize around human schedules. We’re not going to optimize around hours of service, which is basically how the industry optimizes asset utilization today. We’re going to work pretty hard to avoid the truck getting stuck in traffic and to contribute to traffic. So, we’re going to be sort of focusing on how you actually route around urban areas in the middle of the day and/or at rush hour. And how do you not contribute to congestion. The overall impact will be you’ll see, a lot more sort of stuff moving overnight, and to some extent that happens already, but to a much greater degree than we see right now.
Mush Khan: Yeah, that makes a lot of sense. Allen, both you and Dan talk a bit about safety. Tell us about your view on how well the various regulating agencies and entities out there understand the impact of freight related technology on things like safety?
I’ll just sort of couch it with my own observation I made and a couple years ago watching a congressional hearing with Mark Zuckerberg with Facebook and the regulators attempting to sort of figure out how to even ask questions about technology and its impact on our world and it strikes me sometimes the regulators may not have a great understanding of how quickly technology is improving.
What do you think about that? What do you think about how well regulators really understand what this technology is? Some of the things that Dan’s talked about in terms of the value proposition to society. Tell us more about that.
Allan Rutter: Let me start on that. I mean, I think there are two distinctions, one is policymakers, people in elected office and regulators. If you think about who’s at one of those congressional hearings. Tend to be older, tend to be guys, and they usually have been doing that job for a long time. The good thing about the highway environment is most of your elected officials still have a car and still drive.
So, they’re somewhat familiar with that that experience, but they’re not as on top of how fast things are changing. So, making policy tends to be a lagging indicator rather than a leading one. I think one of the interesting things that Texas legislators had done. Kelly Hancock, is a state senator from the Fort Worth area, sponsored some legislation a couple years ago to create a regulatory environment in Texas that says, hey, if you want to experiment on automated vehicles, trucks, autos, here’s how that happens. We want you to have some insurance. We want you to be reputable but we’re not going to tell you what to do or how you do it. We will create an environment which you can experiment because we want Texas to be a place where people can experiment. And to do so safely. So that’s really been the reason why you see Kodiak and TuSimple and Embark, and a whole lot of people in Texas. Is because Texas legislators created an environment or a regulatory standpoint that allows that experimentation to happen. I think we haven’t seen it at the National federal level, Congress, be able to create a regulatory environment for automated vehicles, freight or automobiles. Most of the attention has been on automobiles because of Uber and everybody else’s experiments. And when you have a pedestrian fatality that gets everybody’s attention. Most of those testing regimes haven’t included freight, that allows for this experimentation to continue to happen. I think regulators themselves are trying to figure out how to encourage the adoption of this technology which they know to be a safer operation, how to make that happen. How to do so in a way that makes other drivers comfortable.
People who are on the I-45, as I pull out of Bucees and get on the highway out of Madisonville, how are they going to be comfortable about that truck they’re merging next to, how automated, it might be. That’s going to take a little while and a little public education. The good news is that folks like Dan and the Kodiak guys, by demonstrating this on an ongoing basis. Part of that education is to say, you probably didn’t know this, but for the last year, we’ve been on the same road you’ve been. That’s going to be an important part of that public education.
Mush Khan: Yeah. Makes a lot of sense. Well, certainly it’s exciting as a Texan to hear that the Texas is one of the states that’s leading the way and providing a progressive way to look at adoption of technology within the framework of regulations and policies, etc.
Let’s dig into one core topic in the time that we have remaining. Want to make sort of a really practical conversation. So, Dan, let’s start with you. Let’s say that you’re a freight company or you’re somewhere within the supply chain ecosystem, can be a freight company it could be a buyer of freight, warehousing, etc. What kinds of things should you be thinking about today to take the first steps on adopting a new way of thinking around your business? Because it can be pretty daunting as we hear about these sorts of technologies and how things are moving. It can be pretty daunting to think about where I even begin if I’m an owner of these businesses so what are your thoughts on that Dan?
Dan Goff: Sure. I think it’s a first step to think about this. I mean, as you just said, our trucks have been on the road for a year, we’ve delivered 400 loads for customers. Basically, nobody’s noticed. Sometimes, I wish people would notice more to the extent that I’m involved in our public outreach.
And so, I think one of the interesting things about automated freight is that, in fact, we’re not really disrupting anybody in the way that people think about Silicon Valley’s sort of cutting in and breaking stuff. In fact, you know the American freight system is incredibly efficient and incredibly high, well-oiled machine and we’re really looking to make just one part of that machine, a little bit more efficient and a little bit safer.
I hate to say this because, you know, one thing that we realized sort of fairly quickly is that in the end we’re a trucking company more than a technology company in a lot of ways. So, if you’re interested in in the technology, give us or one of our competitors a call and say, hey, I’d like to ship something with you. And in fact, you’ll probably see that it won’t make that much of a difference in what you’re doing, which I think is great. I don’t think at least in the short term, people really need to reorient themselves around automation. In the longer one I think some of the different changes, however, are going to be perhaps a little more profound. So, as I said, again, we’re focused on just middle mile what we call middle mile, but just the sort of interstate highway portion of the logistics chain.
So that means we’re going to be looking for places where we can switch. First of all, we’re preserving a pretty significant role for traditional human truck drivers for first and last mile and actually think we’re going to be creating a lot of jobs in that space. And it means also we’re going to need places where we can switch between a switch a load from a traditional truck to one of our trucks somewhere highway adjacent. If you own a warehouse or in the warehousing business. I think it’s time to at least start thinking about how do we prepare our warehouses and prepare our infrastructure for self-driving trucks. Are we optimally placed close to highways? Do we have sort of the right space to switch loads? Again, you know, once you’re already switching loads, maybe it makes sense to switch to an electric tractor, which will, once the infrastructure build out, probably be a little bit cheaper and more efficient. So, are we prepared for those kinds of switches?
But those questions are a little bit of a ways off and right now if you’re interested. It’s something that actually is relatively easy to try, so long as you’re on one of the routes that we or competitors are focused on right now. This technology is not quite on rails, but we are just doing Dallas to Houston right now. And that’s something that I don’t know if we’re going to change that next year or the next.
Mush Khan: Great. Thanks, Dan. Allan, how about you? What would be your conversation with someone who is a business owner or leader in the space? How should they start taking steps to really embrace technology like this?
Allan Rutter: Well, I think they have the opportunity to think about like how Dan described their operation as a freight provider, not a technology company. And so, if you’re a shipper or you if you have a fleet, or if you’re in the business as a broker or 3PL, your job is to provide your clients the ability to take stuff from one place to another, keeping an eye on how that happens and how to add efficiencies.
The discussion or consideration of whether that highway trip is going to be automated is part of how reliable is that? If that’s happening on off hours maybe you get some reliability benefits. Maybe there’s some cost benefits that accrue to you. And those same 3PLs and shippers and carriers are also trying to figure out how to gain more business intelligence out of the ton of data that’s being created by their trucks in normal trucks right now.
A whole lot of OEMs are starting to realize that they are not only in the equipment manufacturing business. They are in the service provision business. And so, a lot of folks who provide tires or lubricants or engines are offering people the opportunity to buy service in the same way that HP actually sells ink not printers. These guys are providing ongoing service to help create diagnostic information that helps people gain more business intelligence. How the drivers are performing and how they can be coached, how their engines are performing and how to get better mileage.
There’s a whole lot of benefits that these guys are adopting technology incrementally all over the place, automation in the vehicle movement is one part of an automation space that’s increasing their ability to keep squeezing more efficiencies out of the freight world.
Mush Khan: So it seems like both of you talked about this concept of different parts of the supply chain beginning to light up with automation or technology or data, and it certainly seems like we’re moving towards a connection of all those things together. Maybe we’re still some time away from that. So that beautifully highly efficient connectivity, but I can see how that world is evolving. So, one final question for both of you.
If you had a magic wand and you could predict or create what the world would look like in five years’ time as it relates to the freight movements that we’re talking about, what would you want to see what do you think would be an incredible outcome for society and customers and participants in the supply chain. So, Dan. I’m going to start with you.
Dan Goff: That’s a good question. I think from an infrastructure perspective, one thing that the industry is really focused on is making this as lightweight as possible.
We don’t want this to be a 50-year infrastructure project. This is not building the interstate system again. And it’s really uncool, but if there’s one thing that policymakers could do to sort of accelerate adoption of this technology. I’m going to go with two things. One is that highway adjacent parking that I mentioned earlier, and I think it’s important to understand machine learning. People talk a lot about machine learning and self-driving is an application machine learning technology. Machine learning is not magic, it’s a very high functioning pattern recognition.
So, making those patterns as consistent as possible, turns out to be really important. Making sure that when people are doing construction they actually bothered to repaint the lane lines, even if it’s only for a few weeks of construction, turns out to be really helpful for these kinds of technologies.
People often asked us what we’re looking for and expecting us to say we need 5G covering every inch of highway and all these vehicles, to infrastructure, to kinds of technologies and like lane lines, actually, it’s pretty important.
I do think the public education part is really important, as Allan talked about, that people are really going to like these technologies but are going to be weird at first. And we were planning on bringing a truck to SXSW, we were pretty excited about that. We were going to do a demo. And you know, it’s hard for people to get sort of hands on experience with these kinds of technologies these days. But I think that’s really important as well.
Mush Khan: Yeah. Great. Thank you, Dan. Allan. How about you?
Allan Rutter: I’d say five years from now, we’re going to see an explosion of different kinds of freight technology beyond what we can even expect right now. I think we’re going to see different kinds of propulsion, power. A lot more electric, a lot more fuel cells, a lot less diesel.
We’re going to see some changes in business practices. A lot more hub and spoke, a lot more regional, which hand off to some of those inner-city movements. I think the ability to leverage the efficiencies that come from this increasing automation of long-haul interstate movements offers some abilities to have trucking almost function a little like air airlines did a couple of decades ago.
So, I think there’s going to be real variety of kinds of opportunities out there, all of which means that as we as consumers buy more stuff and buy more stuff online that all that can get to our houses as soon as we want, efficiently and cheaply and safely. So, I really expect some really dramatic, unusual, and very interesting changes in the future.
Mush Khan: Well, Great, well thank you both of you for your time this morning. Such a great conversation.
Thank you for your time this morning and to all of our listeners. Thank you for listening into this conversation hope everyone has an amazing day today. Thank you.
Interested in learning more about our software or want join our mailing list? Click the button below.