The peloton Journey
An in-depth interview with
Dr. Josh Switkes, Founder and Chief Executive Officer
The following is a transcript of the audio available via the player above. The audio file is the definitive source.
Voiceover: Tech, cars, machines, big trucks. Keep your engine running. Click subscribe. Subscribe with the little button in your podcast app, or click the three dots in the little circle, or visit us at techcarsmachines.com and GTKPartners.com, where our subscribe buttons are much bigger.
Ali Tabibian: Welcome, welcome, welcome to this episode [00:00:30] of Tech, Cars Machines, and maybe this time we should call it Tech, Trucks, Machines, because that's what were going to be talking about; mainly 18 wheelers in fact. So, big trucks, and we're going to be talking about these big trucks with Josh Switkes, founder and chief executive of Peloton Technology, an extraordinary informative gentleman. When you hear the interview, you'll see why I describe him that way and Peloton is a truck platooning company. What does that mean? It basically means that Peloton enables tailgating. They [00:01:00] help two trucks find each other. They help them get close to each other and follow each other really closely and as it turns out, the way the aerodynamics work out, both trucks experience reduced air friction and therefore save a lot of fuel, which is a big deal as you can imagine, in the trucking industry.
Now in second we're going to play you the audio portion of a little clip that the company has that'll give you a sense for what's going on inside the truck as this platooning is being set up, but I like to think of the company as having two major pieces of technology. [00:01:30] One is inside the truck and that basically allows for coordination between the two vehicles. They're the radars and the cameras and the other cool self driving stuff that help the trucks stay in tight formation safely. Most uniquely it's the truck up front, which is the scout looking ahead about 850 feet, that controls the cruise and the gas and the brake for both trucks via a direct communications link and that's kind of cute and unique.
By the way, fuel saving is where this all started. [00:02:00] It turns out safety is impacted positively too. The second piece of the technology is the network operating center. We'll go through that a little bit more but it's basically a control center that can nationally control where these trucks are, helping them find each other and give them other useful information.
I'd like to suggest that you look for three things in this episode. One is listen carefully to haw the company is very carefully defining the problem it's trying to solve. [00:02:30] These are not automated trucks, they're just adding to the drivers and the trucks capabilities. They're not trying to replace the driver, in fact they're trying hard not to change the experience. One video display on the windshield, one big button to engage and disengage.
Josh Switkes: The key there is that allows us to constrain down the type of situations the system encounters tremendously. If we had to build and validate the system for every situation in the world we'd be working for a very long time on it. By constraining it down to the ones that [00:03:00] really matter, which for trucks are, the big highways, the US highways, we really simplify that problem.
Ali Tabibian: As we discussed in episode two, transportation is an enormous market, if you're solving a meaningful problem, you're going to wind up with a big outcome, therefore there is no need to exaggerate. What you really need to do is define the problem correctly so that you're not over reaching in terms on the capabilities of technology as they are today so you can deliver something meaningful to [00:03:30] people in the short term. As the industry likes to say, "If you bought it, a truck brought it," and that truck typically consumes $100,000 of fuel a year so any difference you can make is big news.
Long term the value of Peloton might substantially be to what the company calls a network operating center. Again, a quick explanation from Josh. More later.
Josh Switkes: We also have truck to cloud communication so we've build what we call the network operation center, which is a cloud service, which is supervising [00:04:00] and coordinating the trucks.
Ali Tabibian: Third, I'd like you to take measure of Josh's personality. Notice how controlled and measured he is. Extremely smart person obviously but still controlled and measured. This matters if you're in a business that's trying to address an industrial segment. When I come across companies that are trying to sell to big industries, transportation, heavy machinery, etc., what I suggest is that they, to some extent, set aside what is most typical [00:04:30] about Silicon Valley culture. And the reason for that is, over the last 10- 15 years, most of the really big money in the Valley has been made by consumer facing companies. Those companies by their nature, they're selling things one customer at a time. They have to do a lot of work to get attention and so those companies tend to be more extroverted, more ebullient, just have bigger personalities to them and to the extent that marketing personality gets absorbed by their marketing executives. That's not the kind of personality [00:05:00] that a large industrial company is looking for when it's looking to adopt technology that substantially affects safety and reliability. A lot of this may just be a matter of perceptions but it's real. So listen carefully to how Josh's delivery and his personality almost naturally would make someone comfortable that what he's going to deliver is going to be safe and reliable and he's not necessarily looking to blow you away and get your attention.
Okay, that's enough [00:05:30] of my soliloquy. We're going to play, for about a minute, a little audio snippet from the Peloton platooning video that you can find on their website. And we'll put links to it in the episode web pages. What you're going to hear is a truck staring off, getting on the road. They system advises that there's a platooning opportunity available and guides the trucks toward each other with, basically, some speed instructions, and then essentially what the truck drivers do when the get close to each other is push that big Peloton [00:06:00] button and the system then links them electronically and from that point on, not the steering but the pedal work is essentially being done by the truck up front. You'll hear how the system automatically manages for a car that cuts in between the two vehicles and you'll also listen to the two trucks disengage, and then we're going to go right to the interview, which starts with an attempt by me to get a quote from Alex, Josh's dog, who is typically the first person that runs out to greet me when I'm at Peloton.
Peloton Music: [00:06:30] Partner available. Proceed at 54 miles per hour. Proceed at [00:07:00] speed limit. Rendezvous in one minute. Ready to platoon. Now platooning
Follower truck:: Hey lead truck, how's it going up front?
Lead truck: How's it going back there?
Follower truck: We got a challenger coming up on us real fast.
Peloton Music: Cut in [00:07:30] detected.
Follower truck: Alright the challenger is exiting now.
Lead truck: 10-4 where you headed?
Follower truck: I'm going home. It's my daughter's first birthday.
Lead truck: Before you know it she'll be running. I'm going to stop for fuel up ahead.
Follower truck: Sounds good. I'm going to keep going.
Lead truck: Stay safe out there.
Follower truck: Thanks I will.
Peloton Music: Platoon ending.
Ali Tabibian: Alex are you going to [00:08:00] give me some audio? Woof.
Josh Switkes: I don't think she will. If someone walks up to the window she might bark at them but that's the closest you can get.
Ali Tabibian: So Josh, thank you very much for taking the time. It was a pleasure having you at our conference, and it's a pleasure to do this again. Let's maybe start in the beginning. How did you come up with the idea and what was right about the time?
Josh Switkes: Yeah so I had been working at a big automotive company [00:08:30] starting with my Ph.d research at Stanford, which ... toward the end we worked with Volkswagen. Then I worked full-time for Volkswagen. Then I got the start up bug. I went to a power train startup, so making engines more efficient and then I really got the bug to start up my own company.
Together with a few friends we were brain storming company ideas and we saw that a few things had really come together. One was a reduction in the cost of hardware, computing hardware. [00:09:00] The second was a rise in fuel prices over the years and together those made the value of technology that could save fuel and could increase safety, made that extremely valuable. So we started looking at where that could be applied. We were all automotive guys, not trucking, and we just happened upon the statistics around trucking. I remember I wrote in my notebook one evening that each year $100 billion was spent on diesel fuel for heavy trucks in the US. [00:09:30] I was still working at another company. A few days later I looked at my notebook and I read that $100 million, oh that's pretty good. And I said "Wait a minute. I wrote billion. Is that right?" I looked it up. Sure enough it was billion.
Ali Tabibian: Right.
Josh Switkes: So then I said "I can't comprehend a number that large. I need some scale of reference. So at the time I looked up the total global internet advertising market. At the time- it's grown, but at the time it was about $50 billion. I said "Wait a minute, [00:10:00] there's literally thousands of startups trying to, directly or indirectly, get a little piece of that advertising market worldwide, while, just in the US, the fuel use is twice that. I said "Okay we've got something here. We need to look at how we can apply this technology to the trucking industry."
Ali Tabibian: That's really interesting, Josh, in a prior episode our listeners heard me go through a little bit of a calculation on how much scale the transportation industry have. You know we work in the field of bits and bytes and are surrounded by people who do, you just [00:10:30] forget that. And on the fuel side, burns around about a billion gallons of fuel a day. It's like filling up the Trans america Tower twice a day and setting a match to it every time. And that's a quarter of the global fuel consumption. I remember when we talked once before, and this is one thing that's really kind of stayed with me since we talked at that time, from the conception to the animation of what Peloton does to where you are today, it's actually been a fairly [00:11:00] linear path. Is that accurate?
Josh Switkes: Yeah.
Ali Tabibian: By start up standards linear.
Josh Switkes: Yeah, I mean, I think in every startup there are ups and downs and what I like to say is, part of the job of the CEO is to smooth out and filter out those ups and downs so that it is linear up overall. So, yeah, we saw this big opportunity to apply this technology to trucking and we said "how can we test that before we really build it?" That's [00:11:30] sort of the key to all startups is you don't want to go too far before you know if it really make sense.
With a technology like this where it involves real vehicles and it involves real safety it takes time to make it real so we started testing it how we could. You know, fleet executives, maintenance guys, safety guys, just to get their reaction. Then we went out and prototyped it in what way we could without funding. So we would take pictures of trucks close together and see what the view looked [00:12:00] like out the window of a truck to see whether it's going to be comfortable for the driver. We bought a very small truck, basically a U-Haul truck, and started doing some prototyping in there.
Ali Tabibian: Amazing.
Josh Switkes: So we kind of built it up as ... Trying to make it as linear progress as possible.
Ali Tabibian: But you also have a network operating center, essentially. Maybe give us a minute on that to set the stage of what's going on in the background as well.
Josh Switkes: Sure, yeah, so our approach is to use connectivity in multiple ways. So we have truck to truck communication, so [00:12:30] the front truck is communicating to the rear truck exactly what it's doing so that rear truck can react nearly instantly and very accurately. We also have truck to cloud communication. So we've built what we call the network operation center, which is a cloud service, which is supervising and coordinating the trucks. It's making sure that the platooning is only happening when you're on the right type of road, in the right weather conditions, with the right partner, at the right following distance and, the key here is [00:13:00] the cloud can have, does have in our system, information that the truck cannot have or that would be very difficult for the truck to have on its own. For example, what are the conditions a half a mile ahead? Well there's no sensor that can detect that but it's easy for the cloud to know that there's traffic ahead, there's a severe weather alert, that type of thing. So we combine that with on vehicle, on truck intelligence to make sure the system is safe and operates properly.
[00:13:30] A real key there is that allows us to constrain down the type of situations the system encounters, constrain that down tremendously. If we had to build and validate the system for every situation in the world, we'd be working for very long time on it. By constraining it down to the ones that really matter, which for trucks are the big highways, the big interstate, the US highways, we really simplify that problem.
Ali Tabibian: And that's a really revealing statement actually and we'll come back to it if you don't mind when we talk about some of the data analysis [00:14:00] and artificial intelligence. Speaking of development and technical development, anything interesting or surprising you think somebody should know looking at the truck in terms of what was inside of it? What was kind of intriguing? What surprised you that you had to develop?
Josh Switkes: One of the nice things about trucks compared to passenger cars is they're built in a much more open and modular fashion than cars. The reason is that trucking fleets demand extreme customization. [00:14:30] So if you're a mid size fleet you might buy 1000 trucks from one truck maker per year and when you go to that truck maker you say "I'm buying 1000 trucks this year. I want this engine. I want these brakes. I want this transmission. I want the wheel base to be exactly this length" and so on. So it's much more customization than you get on a car where you might chose leather or cloth, which stereo, and maybe the engine size. As a result of that customization, [00:15:00] truck makers have had to build the systems ... The truck makers and the suppliers have had to build the systems to be modular and inter operable because they can't go through the same design cycle a car maker would where three or five years before production they've locked in all the details of the design. Instead this engine has to work with all the different brake systems. This transmission has to work with different wheel base settings and things like that. So that means there's standardization [00:15:30] of much of the communicate between components on the vehicle, which makes it easier to integrate into them whether as an after market system or a factory system.
The other thing I had mentioned on the development side is, we have tried to, and been pretty successful, in prioritizing what part of development we do at which phase of Peloton. What I mean by that is, we started by building a prototype that was completely safe to operate with an engineer in the truck.
Ali Tabibian: With a chaperone.
Josh Switkes: With a chaperone, [00:16:00] and that was sufficient for fuel economy testing. So we did very rigorous fuel economy testing with a system that was, certainly, was not intended for commercial use. Then, once that was proven, and we got more investors involved and more partners involved, then we moved on to production development where we needed, and have focused, on the safety side of things.
Ali Tabibian: Let's talk a little bit about your customers, what were your assumptions about their care abouts when you started [00:16:30] and what do you know today, assuming that that's different?
Josh Switkes: So I think there's... When we first started we said as I relayed we were excited about the fuel savings.
Ali Tabibian: Right.
Josh Switkes: And we said "Wow this is ground breaking for fleets." They have- And this was all correct, I'll explain in a minute why this was not the full picture. They have typically one to two percent net margins and fuel is 30-35% of their operating expense. [00:17:00] So saving them 5-10% of fuel, you know, it can double or triple their profit in some cases. It's a huge driver. So that was correct and we knew that would get them excited. What we- as we started talking with fleet even with the animation, one of the first topics every time people would bring up every time, is like "Well, that's great, if it's safe."
Ali Tabibian: Right.
Josh Switkes: Safety is critical for many reasons but for the customer it's critical [00:17:30] financially in terms of their own liability. In terms of the down time when there's a collision and in terms of their brand. A truck accident- usually when there's a truck accident, it's on the news. So when we first started Peloton we kind of viewed it as, this is how we described it to people, we said "You know, basically here's platooning. It saves a ton of fuel. Here's how we're going to make it safe." Unintentionally that description was interpreted [00:18:00] as "Okay there's a trade off here. It's a little less safe but it's worth it for the fuel savings." That was not our intent, but that's definitely how people interpreted it. So we started looking ... We got that reaction from a lot of people and we started looking at the statistics around safety. And we realized that actually the platooning system would have a profound improvement on safety. You know, profound, a decrease in collisions, and so it's actually a huge benefit of the system. So we then, both externally and [00:18:30] for our own purposes, realized that the real message here was, you put the system on your truck. You're much safer and you save fuel. There's something in it for everyone, and that made it an even more compelling financial argument for the fleets. It really kind of rounded out the story.
Ali Tabibian: It's really interesting to hear you say that. I remember, just in the time that we've known each other, I can tell from the companies collateral that the focus has gone from fuel, [00:19:00] to, I think safety is actually mentioned first on your website collateral as an example.
Josh Switkes: Yeah we usually mention it first. Certainly when we talk to a fleet, the financial benefit is probably roughly equal probably from the safety and fuel savings. The good news is they don't have to chose. They get both.
Ali Tabibian: Always a good thing.
Josh Switkes: Yeah
Ali Tabibian: That's one lack of choice that's a positive. What's the ... How much custom hardware, custom infrastructure, [00:19:30] custom software did you have to develop? Are you really integrating things more or are there a lot of components that you essentially have to build of from scratch?
Josh Switkes: We used some off the shelf hardware components but we built them into our custom ECU that runs our software. Our software is ... certainly we're using existing operating systems and things like that, but basically all the software is custom. On the hardware side, the architecture of the hardware is a critical part of how we make the system safe. So [00:20:00] we've done a lot of work there. That being said as we more forward our goal is to be mostly a software company. The hardware enables the software to run in the trucks, but over time trucks will have more of the hardware components needed just from the factory. So we're looking and working with the truck makers on ways to more tightly integrate. That will lower the cost allowing everyone to make more money, including the customer. It'll simplify things for the customer and for Peloton.
Ali Tabibian: [00:20:30] It's interesting, the network operating center that we talked about, for me it's essentially an orchestration layer that has quite an independent future on its own. Most other companies in this scenario generally find that everything they offer is relevant for a long, long time, it just depends on which geography you're talking about. When I look at your investor list that's almost like ... It could almost be a skit in a comedy routine in terms of how many names are on. I got Intel, Mitsuye, NGP, [00:21:00] Okaya, Sandal Ventures, British Petroleum Ventures, Lockheed Martian, Schlumberger, Volvo, I'm out of breath, Den so, UPS, Omnitracs. How did you wind up efficiently reaching out to all these people, managing the diligence, and getting them involved in the company? There's the process part of it, of getting them all together. And then there's the process of managing all these disparate entities once they're your partners.
Josh Switkes: To a large extent, the [00:21:30] corporate and strategic investors, they all know each other. So a lot of our networking among them was referrals by existing investors saying "Hey you should talk to this potential new investor." I think in general it was fairly straightforward to get them on board because the ones we have are all valuable to use and we're valuable to them. Meaning we fit in some part of their current or future strategy. It varies. We have companies [00:22:00] with large fleets that are potential customers for us. We have technology companies. We have several oil companies. All these companies have in common that they see their industry is changing and they want to be involved with us to see how it's changing and then to adapt to it.
In terms of how we manage them ... One thing that's good about all these strategic investors we have, they all realize that they need to be careful with their [00:22:30] demands on the time of me and the management team. So they're all good about saying "Hey we've got this potentially interesting thing to discuss. If it's of value to you let's spend some time, if not that's okay." So we can filter in that way. Other than that we manage them just like any other investor based meeting we have board directors and board observers and we don't have all of those companies as board directors. That's important. We've been careful in [00:23:00] which ones we put on the board to avoid conflicts of interest.
Ali Tabibian: I think the strategics themselves undervalue what they're bringing to the companies that they invest in. Because people want their participation and generally, like you said, because they're operating people themselves, they tend to strike a better balance than the professional investor in a community. That's very interesting. When you talk about the various components of your system, and by components I don't mean hardware components, I just mean any aspect of your system, [00:23:30] you use word that we've used for a long time, hardware, software, and you don't use the words like data analytics or big data and artificial intelligence and all these monikers that I'm not sure exactly what they mean. What is your view on ... Is there something really grand and drastic that's happening or is it just more computing power bringing more solutions to bear with the occasional fundamental breakthrough maybe neural networks going back 10-20 years being an example of that?
Josh Switkes: [00:24:00] Certainly I think the techniques that people use effectively are advancing pretty quickly. I like to focus on the functionality and the value. This categorize have not changed and are not changing, meaning, you can look at the breakdown of expenses for a trucking fleet and those are the expenses that can be reduced. It's labor. It's fuel. It's crashes. It's [00:24:30] maintenance. It's the truck itself. There's no value in just having AI on a truck. The value is if that AI allows you operate more efficiently by reducing one or more of those expense categories. So certainly the new ... Some of these areas like computer vision has been completely revolutionized in the last three years basically by some of those new algorithmic approaches. That being said, that will allow, [00:25:00] for example, improvement in safety to reduce collision costs. It's the same old category of costs, a new way, new technology for reducing it.
Ali Tabibian: Interesting. When you talk about artificial intelligence in your truck, are you referring to what I would maybe just call machine learning or self learning? The software getting a little better at what it does over time, not really thinking for itself just having been programmed to take certain inputs and rework them into the factors that it uses.
Josh Switkes: [00:25:30] Basically yes, in general, the way we're approaching it is ... there's some people out there in the extreme AI camp of put the software out there, let it learn everything, let it do its own thing.
Ali Tabibian: Hard to do when you have a 40,000 pound vehicle driving 70 miles an hour.
Josh Switkes: It's hard to do and I think people are finding that the pure approach is very difficult to prove a safety level or even measure how safe it is. Our approach [00:26:00] is more structured. Use learning algorithms where they make sense so in our platooning system it's a fairly small set of functionality where we actually use them because a lot of it we can structure, we can itemize out the hazard and how we handle them and prove statistically that we are safe. As we progress there will be learning out in types of algorithms on more areas, but [00:26:30] we see a need for that structure to persist in order to even know how safe we are, because otherwise you have to accumulate so much field data. If you're trying to show that you only have one collision every 100 million miles well you need a lot more than 100 million miles just to show that. So you know Google's been out there, I think they're up to 5 million miles now, which is tremendous. By far the most of anyone. It's not nearly enough to show safety in a pure sense [00:27:00] from just mileage accumulation. It needs to be, and they are combining with other approaches, just as we are, track testing, hardware in the loop testing, that type of testing
Ali Tabibian: Simulations
Josh Switkes: Simulations. Hand analysis. That's where the pure AI is very difficult to show safety.
Ali Tabibian: And it's interesting maybe that's a good segway into the world of regulation. Because certification is something that frequently relies being able to [00:27:30] present a formal view of what your system is able to do and not able to do. Let's talk a little bit about that if that's a good question. If it's not a good question fell free to discard it, I'd rather have what you want to talk about than my bad questions.
Josh Switkes: No, it's a good question. Let me answer it, then depending on how I answer it, I don't know if you'll want to rephrase your question. It was a good question.
Ali Tabibian: See we're [inaudible 00:27:55] this is a regulatory question.
Josh Switkes: So the reality today [00:28:00] is, for almost all car or truck systems, the regulations are not prescriptive on a validation test. There is a small number of exceptions to that. So like stability control, which keeps your car or truck from spinning, crash testing, fuel economy testing. Those may be ... There's probably a few more but those are a few of the only areas where there are specific tests required by law to show the safety of an [00:28:30] existing system. All of this is still being figured out for autonomous vehicles. So nobody really knows where it's going to settle but the reality is its very hard for the regulators to actually test these systems. So most of the certification is self certification. You say "I'm following the other 500 or 1000 safety standards, the ones that are not tested. I'm following them." And the main motivation there is, if there is some sort of collision or some [00:29:00] issue, and it's shown you weren't following one it's very bad for you. You get a lawsuit.
Ali Tabibian: Right. Right.
Josh Switkes: So in our care, we are the biggest decider of is our system safe enough is really us and our suppliers, you know, brake suppliers, truck OEM's. The regulators want to make sure it's safe but they're not testing our system for us. We're testing it. We're validating it. We're going through that whole process and we're following a [00:29:30] process called ISO 2626262, which is a standard for functional safety, a very rigorous process. The regulators don't check on that. They know we're following that. They understand why the system should be safe but then to a certain extent it's up to us to validate it. And we do that because we need to and want to, and our customers want us to and our partners want us to.
Ali Tabibian: What about the work related rules that are imposed upon the trucking industry? How does Peloton interact with [00:30:00] those rules?
Josh Switkes: Yeah so the trucking industry, for good reason, has very strict hours of service rules. So limits on how many hours you can drive. You have to take so many rest breaks and so on. For our platooning system, because we still keep the driver engaged at the wheel they're still steering, we are not planning on any changes to those rules when you're platooning. That being said those rules are a very strict limit on the operational efficiency of a fleet because it makes ... They [00:30:30] have to operate within those limits so it's hard to utilize their trucks fully, hard to utilize the driver fully. So we do see the opportunity for platooning and beyond. There's a logic that they should be able to relax those rules. If you're platooning its lower stress. You don't need to take a break as often, that type of thing. Certainly as we add more automation that lets the driver fully relax then that can also relax those rules. Those could potentially have a huge operational cost saving for fleets.
Ali Tabibian: [00:31:00] I have some of these semi autonomous features on my vehicle and by the most striking feat, for example, on the hour and 15 minute commute down here this morning, the striking thing about it is how much fatigue reduction there is associated with having the vehicle do just a little bit of the work itself. And what's always striking to me is how you never hear about that as a rational and I think that's one of the reasons on the passenger side, the uptake of these safety systems is very little when they're optional equipment.
Josh Switkes: Yeah. A couple thoughts on [00:31:30] that. Typically, the demonstration vehicles they have don't have the features. So for years you'd have to go and special order a vehicle based on a feature that was listed in an option book that you had never driven. So very few people adopt that.
Ali Tabibian: And an expensive option.
Josh Switkes: And expensive, right. But now on the fatigue reduction, for many years there have been studies by psychologists and human factors researchers [00:32:00] that show basically ... They show the task workload curve. The basic idea is that there is an optimum workload that keeps you at peek performance. If you are overloaded, like if you were driving down here and you were doing a crossword puzzle and eating your breakfast and shaving and talking to someone on the phone, you would not do a good job at some of those. It would be up to you to prioritize but you couldn't do all those probably. That's overload.
If you're under loaded, [00:32:30] that means you're not given enough workload. So, for example, if you're in your car and the car is driving itself and your job is just to watch it, but to take no action unless something goes wrong. Studies have show humans are terrible at that. They're not good at supervising without a task because it's hard to pay attention. It's hard to stay engaged. That's under loaded. In between, fortunately there is a broad area of good workload. Still steering for [00:33:00] example, but not being able to control the gas a brake, like you have with adaptive cruise control or with platooning. From what we're see that is a good workload. It lets you focus on the part of the driving task that's important. It lets you not be stressed out so you can do it for a long period of time, but when you add too much automation without enough capability it gets dangerous because you get under loaded and you're no longer able to supervise properly.
Ali Tabibian: The aviation industry has a lot of interesting experience [00:33:30] with how long it takes for a pilot to reengage and I think the statistics I've seen applied to highway driving is you need around 850 feet of look ahead at around 60 miles an hour for the system to be able to say "Pay attention driver" for the driver to be able to reengage, then make a decision, then be able to bring the vehicle to a halt, as an example.
Josh Switkes: Yeah there have been a couple of good studied on that and, part of the challenge is, it varies a lot. So sometimes it's even longer than that. Some times it's very quick. [00:34:00] So the challenge is, depending on the situation, if it's "Okay the system just had an issue and there's not imminent danger on the road," maybe you have that much time. If there's an imminent danger on the road and that's why the system didn't have the capability anymore then you're in trouble. Trying to retake control
Ali Tabibian: Right, lateral movement into your lane, that sort of thing.
Josh Switkes: Something like that.
Ali Tabibian: Josh I want to be very respectful of your time because it's always a pleasure- I could go on forever, it's always a pleasure to be here. Is there anything else that we haven't talked about that you'd like to [00:34:30] point out? What kind of milestones allowed you to raise this wonderful capital?
Josh Switkes: Maybe I'll touch briefly on that and on the regulatory part I can add more there. So I talked a little earlier about doing the most relevant parts of development as we progressed as a company. So what we did is we said "Let's show the fuel economy. That's what's really going to get the attention of the fleets and it's easier to measure than the safety benefit early on." So we built a prototype. [00:35:00] We measured the fuel economy with a major fleet so we got the brand name, brand recognition of that major fleet. That really got the attention of other fleets. They said "Wait a minute. This fleet not only was willing to test it, but the test went well and shows 10% and 4.5% fuel savings for the two trucks." That was very powerful for them to move on to the next stage. Then the key really was show milestones around production viability and production. Show that it could be safe. [00:35:30] Later show that it is safe, or close to being done, validated to be safe. So we just kind of moved along the milestones in that way and got more investors on board as we went.
One more comment on the regulatory side, we have found the states very supportive of platooning and for use the regulations that matter are mostly at the state level. Each state has their following distance laws, tailgating laws, and so we've worked state by state, and the good news is states understand that trucking is a core [00:36:00] part of their economy. Almost everything moves by truck so it's basically important for all physical aspects of the economy. States are eager to do anything that improved trucking in their state because that, you know ... If you're deciding where to open you're new warehouse if trucking is easier and cheaper in one state than another that's one of the factors you're going to include in your decision. So we've gotten good support from many states. I think we're currently at nine states that have commercially approved platooning about another dozen where we can do testing [00:36:30] and then a bunch of others coming on board, some for commercial approval, some for testing all the time. We've got a lot of good momentum there.
Ali Tabibian: Any questions for me as they usually say?
Josh Switkes: I guess I'll just say thanks Ali it was a pleasure to speak with you today.
Ali Tabibian: It was truly my pleasure Josh, you know I always love coming here, seeing you and giving Alex a pat as well. That's one of the treats.
Josh Switkes: Thanks Ali.
Voice over: Tech, cars, machines, big trucks. Keep your [00:37:00] engine running. Click subscribe. Subscribe with the little button in your podcast app, or click the three dots in the little circle, or visit us at techcarsmachines.com and GTKPartners.com, where our subscribe buttons are much bigger.