Latest / Freightvine / Eyal Cohen, Humble Robotics | No Cab No Driver No Problem
Transcript
- Chris Caplice: Hello and welcome to the Freightvine Podcast, your source for all things freight transportation. I'm Chris Caplice chief scientist here at DAT Freight and Analytics, and this fall I'm doing something a little bit different. I've picked a theme for the next several episodes. So over the next half dozen episodes of the Freightvine I'll be focusing on autonomous vehicles and freight transportation. This is an area I've kind of dabbled in with other guests over the last few years, but things have really started heating up over the last 12 to 18 months. technology is definitely moving beyond small pilots to actual commercialization. And the technology and general public acceptance of AVs have dramatically improved. So I'll have guests from carriers, AV startups, tech companies, OEMs, brokers, and others. And the goal is to provide a good overview of where the industry stands today for all of you and to see where it's heading in the future. And my first guest in the autonomous trucking series is Eyal Cohen. Founder and CEO of Humble Robotics, and he is the perfect first guest for this series. He's an industry pioneer. He spent his time initially working on vehicle electrification before joining Otto and eventually Uber in 2016 and working on autonomous driving. And this is when he was there. He helped execute that famous hundred-mile driverless beer delivery in Colorado for Anne Harris or Bush, you might all remember from 2016. Now he's at Humble. And he's merging these two decades of experience to build a cabless electric freight vehicle. And instead of retrofitting old diesel trucks with sensors, Humble is introducing a skateboard platform. It's a simple, flat, 40-foot motorized chassis that combines tractor and trailer into one highly efficient, camera-first vehicle. And in this episode, we explore how these vehicle platforms with a 200-mile range could revolutionize short haul freight, port drayage, and warehouse-to-warehouse moves. We talk about the vision language action or VLA models and other tech behind their driving system, how they plan to work alongside traditional truck drivers, and why the future of autonomous trucking will quickly shift from something miraculous to something being pretty mundane. Whether you're in logistics tech or just curious about this space, you won't want to miss this conversation or the whole series. So let's get started. And following this, following my conversation with Aisle, we will have a truckload update. Hi, Eyal Welcome to the Freight Find Podcast. Eyal Cohen: Thanks for having me. Chris Caplice: Yeah, this has been a long time coming. I think we established that we actually might have run into each other about, gosh, 10 years ago in the mission district as you were developing I don't know if it's generation one of autonomous trucks, but a long time ago. But it's long time been wanting to have you on. So I'm so glad you're able to join. you've had a pretty interesting path. Can you just give us a quick rundown of your path to your current position as CEO of Humble Robotics. How'd you get here? Eyal Cohen: It's been quite a journey. thanks for having me on the podcast. I'm I'm always excited to talk about Freight and Humble and i you know the it's cool that we intersected ten years ago and I'm having that experience a lot actually, people who saw, you know, some of the work that we did. but I'll Chris Caplice: Yeah. Eyal Cohen: give you the I'll give you the whole the whole story. so for me now it's 20 years in the Bay Area. I've been working on very cool, exciting projects now for for twenty years. the first ten years of my career was all about electrification and picture like kind of the early days of Tesla kind of time frame, right? So this is very nascent technology. And so we were trying to figure out motors and batteries and and how to package them. And I got to work on a wide variety of vehic vehicles, buses, motorcycles, military tanks, but all of it was electrification focused. And you know that technology in some ways, you know, w we started a lot of it in the Bay Area and it sort of graduated, you know, you started finding it more in in in it certainly has gone to Asia a little bit and also in the Midwest. and along that route I joined Apple. I can't talk about too much what we were doing there, but around that time I got into autonomous driving, let's say. And Chris Caplice: Okay. Eyal Cohen: and you know so I was coming at I was coming from this position of being an electrical engineer working on electrification, but autonomy really caught my eye, and you're like, okay, this this technology is is is the future, it's exciting. and I got introduced. To a startup called Otto, and this is now where our story intersects. Chris Caplice: Yes. Eyal Cohen: And Otto was trying to automate long-haul trucking. And this is way back, OTTO. This is way back in 2016, right? So this is like Chris Caplice: And it was auto O T T O. That yeah, auto, yes. Yeah. Eyal Cohen: ancient history for, I guess, the way tech tech moves now. but in 2016, you know, we started working on autonomous trucking and you know, we did some pretty cool things. We did a demonstration of of a beer delivery over 100 miles in Colorado with nobody in the front seat. That got people really excited, right? Like, hey, you know, this this technology is right around the corner. But it was such early days for the tech, there was no supply chain, there were no partnerships, no commercial activity. Chris Caplice: Right, right. Eyal Cohen: And the tech itself was very basic. yeah. Chris Caplice: So looking back, does that scare the heck out of you that what you did in twenty sixteen, a hundred miles with no driver in the seat? Think of how far Eyal Cohen: I think it Chris Caplice: the tech's moved. Eyal Cohen: I think we did it safely, but it but it's still a lile demonstration that we did. Yeah, yeah. It was it was you know we did it at night, we did it with we did it in partnership with Colorado DOT. So you it was it was done in a way that was that was as responsible as I guess we thought we were being in 2016. Chris Caplice: Yep, yep. Eyal Cohen: But it was I look back at the technology that is very basic, very, very basic technology compared to today. And and that just you know kicks things off for me. So I I I worked at auto, I worked at Called Auto was acquired by Uber, and so I continued working on trucking there. There was a company called Pronto, recently acquired also. I worked on trucking there. I was at Wabi for a couple years, they were doing autonomous trucking. And you know, for me, all that has led to what we're doing at Humble, which is combining electrification and autonomous driving, and starting to think a little bit about: okay, autonomous technology is here, it's coming. There's no doubt about that. If you're in San Francisco, you can take a Waymo ride, it is safe. And so you see that this is where things are going. And you know, it took a long time for the technology to be ready and safe. I saw Gaddock made an announcement today on a fundraise, right? It's like there's a lot of momentum in this space. And so I started asking the question where does this all go? And we started thinking about what does the future of commercial vehicles look like? And how do I combine everything that I've done in my career on the electrification side, first 10 years, Thomas trucking the last 10 years. Years and kind of bringing that together to form a kind of a new enterprise. And so that's that's how Humble was born, where I'm the founder's day. Chris Caplice: So let me ask you question. So y you've built, you know, from Apple to Auto to Uber to Pronto to Wabi. Was this the plan? Was this the plan or did the things just kinda happen? Okay. Eyal Cohen: There's no plan. No, there's no plan. I was always just I was just driven by what I would say two things. One is one is what's exciting to me and I and I the freight really captured me. I mean I think I did not go into f we I went into freight originally because we thought robotaxis were gonna be too hard. Okay. So I was like, I don't know anything about trucks, but they seem easy, you know. I mean it was like a very like extremely naive I sort of laugh about it today with with with friends but a and partners in the industry, but we were very naive in twenty sixteen about the the freight industry, very naive. And just we were driven by the technology side, but it grabs you, logistics grabs you. It is a very exciting place to to work. I I think you the the industry is just a fascinating industry. There's great people in it. You know, freight starts to kind of pump in your bloods a little bit. You see you just you just see all the all the action happening and you think about trucks all day long. You know, you drive on the highways you stare at trucks, right? It's a yeah, yeah. It grabs Chris Caplice: Yeah, you start noticing them now, yes. Yeah, yeah. Eyal Cohen: you and and so I this became my industry. I just love it. but it w there Chris Caplice: Yeah, and so Eyal Cohen: was no there was no plan no. Chris Caplice: But you did you what was did you learn some key things along the way? which which job do you think prepared you the most for what you do at Humble now? Eyal Cohen: That's a great question. I had learned something from all of them. You know, Chris Caplice: Yeah. Eyal Cohen: you know, I look my my parents stayed in their jobs for forty years in the same place. You know, you never you never anticipate sort of moving around. But every new, every new opportunity that you do is a new opportunity to learn. It's new people around you. I think I've learned something from all of them, but I think that first introduction to freight, you know, coming into an industry cold and suddenly being like, hey, I gotta learn everything there is about Chris Caplice: Yeah. Eyal Cohen: trucks, about how they're regulated, about how they're built. what what does a truck driver do? You know, I think people, you know, again, being naive ten years ago, you think the driver is mostly driving. A truck driver does a lot of things, right? They do a lot. They're involved in many, yeah. They inspect the vehicle, they they're they're there at either end to be to be you know a service point. I I mean, I think truck drivers do a lot, right? so you just learn so much about what you don't know. It's always that pie chart, like you don't know what you don't know, right? And then suddenly you got Chris Caplice: Yeah, yeah, yeah, yeah. Eyal Cohen: so I think that first that first that first job working in freight really prepared. Chris Caplice: So did you watch Smoky and the Bandit to really understand the industry? To understand the movie? Yeah, that's that's the way to do Eyal Cohen: I've learned everything I know from Smokey and the Bandit. Chris Caplice: it. So Humble is different though than all the previous companies that you mentioned. There's there's some similarities. I mean, you're in the same space certainly as Wabi and the others, but tell us a little bit about the whole the major difference, which is the cabless feature of the hardware. Can you talk about that? Eyal Cohen: Yeah, yeah. So yeah, so I started looking at where where A V tech is and and where it's gonna get to, and it's pretty clear that this tech is is now deployed, it's commercially viable. and so a lot of the ways that we have worked on it in the past is we take existing trucks and we we add sensors to them, right? That's that's sort of the mechanism. and sometimes it's in a more integrated way, sometimes in a less integrated way, but that's the basic thesis, right? Eventually, once the technology works, you start asking the question: what should commercial vehicles look like when autonomous vehicle technology works really well? And and I always I always tell people, you know, that we're not gonna replace the class A trucks that are out there. This is an augmentation, this is a new piece of equipment that you would you would add to the toolkit. But you know, you start to wonder, okay, do you do you need a cab as a good example, right? What are the some of the challenges that I've seen with existing trucks? For example, when I try to back into a dock with an existing truck where we've retrofitted sensors, it's really challenging to do without being able to see behind you. Okay, how do we see behind us? You start to ask all these questions about the existing equipment and start to imagine what future commercial, commercial free equipment could look like. While we were like looking at that, kind of like, hey, what could the truck be? You know, there's this electrification side, you know, a little bit nascent are in our industry. There's not a lot of electric trucks out there. but electrification not class eight. Yeah, class Chris Caplice: Not not class eight. You're you're thinking class eight, right? Yeah, big truck. Eyal Cohen: eight. No, not in class eight. a little bit more, yeah, in in in in in other other classes. But you know, we if if you're doing autonomous technology, electrification kind of goes hand in hand really nicely for a lot of technical reasons, but ultimately, you know, autonomous technology, you're trying to get the the vehicle to be hands-free. It's it's a lot easier to imagine charging a vehicle hands-free than rolling up to a diesel diesel station, right? so electrification kind of goes hand in hand. But electric vehicles, you know, the especially in class eight freight, that really serves shorter moves. Long haul electric trucking is very challenging. I think as the industry has realized. And so to to to sort of like sum it all up, we're looking at these these these autonomous freight movers, these simple skateboards, we're looking at shorter moves, and we're looking at making it electric. That's humble in nutshell. Chris Caplice: So then, yeah, so if you look at I like to think of the whole freight transportation system as three main segments. There's in the yards where it's totally controlled. You're in a controlled environment. then you have the first mile, last mile, where you're kind of collecting on the arterial roads, and then the you have middle mile. And so it sounds like what you're looking at is more of w well, in the yards could be, but I'd would you use this as a as a yard dog, as something to move? Okay. But then Eyal Cohen: Potentially. Yeah, potentially. We're looking at those applications, yeah. Chris Caplice: but then Look, cause a lot of the other companies are talking about hub to hub on highways and you know y Eyal Cohen: Mm-hmm. Yeah. Chris Caplice: essentially it's mimicking the intermodal model. Are you thinking of more of drayage for this? Is that one of the primary Eyal Cohen: Dray it Chris Caplice: aspects? Eyal Cohen: yes, yeah. So drayage, shorter moves, warehouse to warehouse moves that that you find even in the class eight space. I think it in your in your model it'd be more akin to that first and last mile, even though there's still class eight freight that does that, right? I think I've worked Chris Caplice: Yeah and Eyal Cohen: on the long haul problem with those other companies for many years, right? I want to look at these sort of shorter moves and see what we could do there. Chris Caplice: Cause it seems like that, I don't think we're gonna see like you know, going through downtown Boston doing multi-stops with a class eight with that, but you're talking more kind of feeding into a an LTL coming from a a port, I assume. This would probably fit very well in California with that has more of the different standards that are required. Eyal Cohen: That's right, yeah, yeah. you can imagine intermodal's a great a great place for us to be participating. Chris Caplice: So w by being not having a a a human in the cab, you're you're lighter, I assume. Are you is it smaller as well? Does the does the dimension shrink as far as the length? Eyal Cohen: What we've tried to do is maximize the dimensions that we can take, given where the state law is. But Chris Caplice: Okay. Eyal Cohen: one way to think about the vehicle that we've, the first vehicle that we've made, because we have a family of vehicles, we have a concept, a concept that we're gonna present to the world soon. But the first vehicle that you see is the skateboard, right? It's just a platform. And that platform Chris Caplice: Mm. Eyal Cohen: is designed for moving 40 foot. standard containers, right? So it's so the vehicle itself is roughly forty feet. but by combining effectively what that skateboard does if you think about it, combines a tractor and a trailer into a simple platform. Okay, so so so Chris Caplice: I thought I so we're not even articulated for this. You don't have a cab, a tractor, and a trailer. You're actually replacing combining that to a a motored trailer or w what do we want to call it? Is a skateboard. Interesting. Eyal Cohen: Yeah, the first vehicle is just a skateboard. It's just it's it's like a straight tr it's like a straight truck, is one way to think about it, right? but it's just a just a platform that moves. Chris Caplice: So you would need some kind of c a crane to load or unload. Does say so drayage would be natural. Eyal Cohen: At least have yeah. That's right, yeah. Chris Caplice: Okay. And would that be able to go on the highway? Is that the thought? Is that regulatory okay or is that a hurdle that we need to talk about? Eyal Cohen: It's a hurdle. I I mean I think, you know, what the regulators wanna see, and we work hand in hand with them right from day one. It's it's it's Chris Caplice: Sure, sure. Eyal Cohen: one of the you talk about lessons learned over the last ten years. You know, before we even launched the company out of stealth, as as as we say, you know, before I was public about what we were doing with Humble, we had already talked to the regulators. And you know, it's it's kind of like a walk with a crawl walk run or wh whatever whatever whatever phrasing there, right? But we we wanna, you know, you you start, you start simple. You work with the regulators, you show them what you're doing, you show them how it's safe and you and you progress all the way to highway. Chris Caplice: So that makes sense that you're describing because I know of some of the other things I was reading, it has a a top speed at this point, around fifty-five miles an hour, which makes sense for local drayage. And then the two hundred mile range. At first when I read that, that seemed you wanted to be longer than that, but if you're just doing drayage, you could get two round trips in a day, easy with i as far as a charging cycle. Eyal Cohen: Yeah, I mean we found some use cases where class A trucks are moving, you know, single miles, you know, back and forth all day long. And intermodals. Chris Caplice: Fioda Chicago, the you know, connecting the intermodal, you see him just going back and forth, back and forth. Eyal Cohen: Exactly, right? So those are great use cases for a simple vehicle like this. and it's the first vehicle that we're doing, but there's there'll be there'll be more, there'll be articulated versions, there will be more tractor-like versions, but they all kind of use the same technology and and and fit that same profile. Chris Caplice: That's really interesting. So that I didn't I I didn't envision that when I was reading through because I didn't see any pictures. Cause you you just came out of stealth mode, what, three months ago? April? Eyal Cohen: April, yeah, yeah. Chris Caplice: What made you decide to come out at that point? Eyal Cohen: well we had built the vehicle and I just thought it was you know what you see on our website is a lot of renders of the vehicle, but we had built that exact vehicle. It drives. It's it's a real it's a real it's a real real piece of equipment as as some of our partners have seen. And so once we build something very cool, you know, you just want to announce it to the world. So that's why we did it. Chris Caplice: So can you tell me a a little bit about the partners you're doing? Are you running some commercialization pilots or tell me more about that? Eyal Cohen: I can't say too much yet because we're we're we're Chris Caplice: Okay. Eyal Cohen: early days, right? So the Hubble's been a company for about a year. and we we have signed we have agreements with a number of customers to kinda explore the technology with them. Chris Caplice: Okay. Eyal Cohen: Can't say too much yet, but more soon. Chris Caplice: And then the the diff big difference, like we we talked about a little earlier, was that a lot of these other companies going long haul, they're going hub to hub and then have a human take it the last mile. You specifically call it dock to dock. What do you what do you mean by that? Does that go hand in hand with the drayage side of things? Eyal Cohen: Yeah. Yeah. So so on one side doc, you know, that the the the on the drage side, at least on a port, it's it's it's where the containers are landing, right? But we also it's something that I ran into when I when I was working on some of the other use cases, we wanted to go right into customer depots and even go directly into, you know, backup into the dock and and and and do an unloading. You know, the the vision that I have for freight for some of these shorter moves is you know, you load a piece of equipment with freight, it it drives automatically to another another location and unloads or loads, you know, it's like that all is completely hands-free. and so So we started thinking about, okay, what does that look like? How does that work? How do you do that? Well, you need to back up a DDOC, you need to navigate these depots, you need to kind of work with the guard shacks. Like there's all this sort of other piece of technology we have to kind of explore to make that happen, which is what we're doing at Hubble. Chris Caplice: So that's one of the fascinating things I found about trucking, because I came into this backwards as well. I was a civil engineer designing bridges and stuff. yeah. Yeah, it didn't, it did it wasn't Eyal Cohen: We all tripped into it, it's amazing. Yeah. Chris Caplice: really here. But the thing that I find most one of the things I find really fascinating about it, it's not done in a vacuum. manufacturing, you can you know control your factory here. You're you're dealing with at least three entities, right? From some receiver to someone, you know, the the shipper itself, as well as some maybe an intermediary. So The interface between, say, the receiving dock. So as you come in, a lot of companies will, you know, they're they're not open parking lots. You gotta go through a gate. And so how do you get the company to get comfortable with there's no driver, no one to show the bill of lading to or any of this kind of stuff. So it seems like this there's a lot of little touch points that are used to be done with humans. So how are you handling that and coordinating with the warehouse? Eyal Cohen: Well that's that's a great question. I think that that's why we sort of engage with these customers so early, Chris Caplice: Yeah. Eyal Cohen: right? I mean, you know, the The typical way this has gone before is that you you know you s you you start the effort, you build all this technology, and you start then going to customers and saying, hey, how do we deploy this, right? But you know, now given that we've been working in trucking for 10, 12 years and and working with a lot of customers, we want to start those conversations on day one. Okay, you have a piece of paper that you're handing to a driver. How would we address that, right, with a a vehicle that doesn't have a person potentially? You are backing into a dock, how do we address that? You have a pool of trailers that you use, how does that work with this vehicle, right? So We're basically exploring all that with these customers because, like I said, the the ultimate vision that that gets us excited is you know, you you Call over the truck, you know, it it shows up. digitally you ask the truck to be there, it shows up. We load that vehicle with freight, it automatically latches and closes the door, delivers it to the customer. But there's a lot, like you said, there's a lot of touch points on that process. And so the that's why we're engaging with the customers early. And there's a lot of people working on various aspects of that tech, right? Like you'll Chris Caplice: Mm-hmm. Eyal Cohen: find automatic automated gates now, so people working on that technology, right, to streamline that. you'll find automatic automated load. Loading and unloading of trailers, right? That's there's companies working on that. Chris Caplice: That's one of the la that's a hard one. That's that's one of the last frontiers, yeah. Eyal Cohen: Yeah, I mean, even, you know, you there's been a lot of work. you know, our vehicle is unique on on how we designed it, but you'll find people trying to automate hooking trick tractors and trailers, right? Figuring out how to get the airlines hooked. There's there's all these details, right? One of the things I love about trucking, it's all these details, but also like we have to figure that out. and so it's it's in it's just why it's why you start these conversations early as a company. And you know, even as we're doing the the the fun stuff, we're building the vehicle, developing this this software technology, it's really important. and double engaged with customers from the day for the first day so we we work in the use cases correctly. Chris Caplice: Yeah, so you raised the point that the trailer pools. i the skateboard model that would be pretty expensive. 'Cause you'd have to the pool you're it's kind of married, isn't it? That unless you're 'cause a trailer pool, you're not gonna have a gantra a crane to move things around. is this more I I'm trying to think how that would work. Eyal Cohen: Yeah, so so it really depends on the customer, right, and what they're trying Chris Caplice: Mm-hmm. Eyal Cohen: to do. and you know, like I said, we have a family of vehicles that we're working on. The first one was that skateboard, Chris Caplice: Yeah, yeah, yeah, yeah. Eyal Cohen: right? So some some companies, you know, the what they're excited about is that skateboard just moving back and forth between two facilities all day. Like that's you can imagine almost like a box fixed to that skateboard and all it does is just go A to B back and forth, right? Chris Caplice: And that would make sense if it's the company owns both entities. So I'm thinking Lt Eyal Cohen: Yeah. Chris Caplice: L, parcel, maybe some manufacturing to D C with a local D Cause those moves Eyal Cohen: Yes. Exactly. Chris Caplice: the the sad thing is drivers love those those moves, right? 'Cause they can just drive all day. but that's Eyal Cohen: Yeah. Yeah, I think I think there's there's some great material handling use cases for the vehicle. you know, I think Chris Caplice: Yeah. Eyal Cohen: drivers love those drivers love those use cases, but d drivers aren't going anywhere. I mean those use cases will be served by drivers for a very long time. Chris Caplice: Sure. Yeah. Eyal Cohen: but you know, this this augments this augments that kind of work. You know, i I think drivers are always gonna excel at the dynamic, right? The That w you know, we w maybe today we go back and forth, maybe tomorrow we go, you know, A B and maybe there's a third point, right? I think where this kind of technology excels is repetitive baseload kind of work just every day in out. But we're also working on vehicle concepts that are articulated, we have vehicle concepts that are for trailers. It just really depends on what our customers are trying to do. Yeah. Chris Caplice: Sure, sure. But they they'll all I assume what you've created is a platform. And then the physical manifestation of that can differ slightly, but the platform it's it it's mainly the same. Yeah. Eyal Cohen: Exactly. So you can imagine a single concept where the hardware and the software are the same, right? And where sensors are located, how the how do how do we interact with gates, how do we do backups, all that we standardize, right? And then the actual way the equipment looks, the length, the size, that's where we sort of can can play with those concepts. Chris Caplice: So another big difference that I had and I don't know the difference here, so I I I I'll need some help is the vision language action models, the VLA models versus Eyal Cohen: Sure, yeah. Chris Caplice: a lot of the rules based autonomy stacks. But d do many of the current ones have are they all rules based anymore, or is that kind of a a dying approach? Eyal Cohen: Everyone's evolving, you know, and and and the technology looks radically different ten years ago than it does today. often what you find is that people have made their bets on some form of technology and then they just kinda wanna work on that and get that working, right? Yeah. It's it's that classic yeah, it's a classic innovators dilemma, right? It's like it's like, I have a thing that Chris Caplice: And it's hard if it's if it's working to switch that once you make your bet. That that's true. So tell us about BLA. The inner absolutely. Eyal Cohen: works, what do w what do I why why do I do something new, right? so Chris Caplice: It's a it's an ATM. I'm making money, so why should I change it? Yeah. Yeah, yeah, yeah. Eyal Cohen: hundred percent, yeah. So so a VLA is it's in a family of what's called end-to-end models, right? But it's basically this idea that tr we're gonna try to simplify it, that you know we have we have a lot of data, we have a lot of data that we can put into a vision model. We use cameras primarily at Humble. That's that's the primary technology that we use. There's a lot of vision data out there. There's not just driving data. There is YouTube, there's movies, there's podcasts like this one, right? There's all this data out there that we can train a vision model. Kind of like you see with Just kind of like you see with language with an open AI today, you can use a chat Chris Caplice: Okay. Eyal Cohen: GPT, that's been trained on works of literature ranging from like Shakespeare to you know internet forum posts, right? It's just been like given this huge corpus of data, right? And it's it's you try to train that model on understanding the world as much as possible. There's equivalents now for vision. Right, where they get trained on all all this data and then they you know they try to understand the the world visually as opposed to language. so VLA kind of starts with that and then uses a little bit of language to decide what to do. what what's unique in the technology side, hope so hopefully hopefully that's a clear explanation. There's there's a lot of details there, but there yeah. Chris Caplice: Yeah, let me just ask a a question. So the do you use any LIDAR to this or is it is it all camera? Eyal Cohen: So the primary driving mechanism for Humboldt is camera. But we Chris Caplice: Camera, okay. Eyal Cohen: have a lidar and radar on the vehicle because I my my belief, pretty strong belief, is that you need this vehicle to be as safe as you possibly can be. And you know, you can't you can't make like a decision ideologically about sensors. You just need all the sensor, as much data as you possibly can to make that decision correctly and be safe. So we use cameras, we use LIDAR, we use radar. Primarily the driving decisions are made off camera, kind of like Tesla. Chris Caplice: Yeah. That's what I was gonna say. Tesla kind of took that approach too. But it sounds like it's a belt and suspenders kind of approach. You just want anything out there to help you make that decision. Yeah. Eyal Cohen: You want to be as safe as you can. This is a these these are these are heavy pieces of equipment carrying a a very large load and it has to be as safe as it can be. Chris Caplice: And is that a hard sell to people to like when you talk to your partners or or do they even care about the technology? Eyal Cohen: I think they're all very curious, you know, and and I think they see some of the the debates that have played out on the technology side. So actually I get asked Chris Caplice: Right. Eyal Cohen: about LIDAR and camera a lot. which is good because ten years ago, you know, you said lidar, people are looking at you like like you're crazy. Like, what do you mean? What's a lidar? And so I think there's a lot more understanding of the technology. Radar you know, radar became a very important sensor for automatic automated emergency braking, right? So that that's been used a lot in the industry already. and you know I think it gives everybody comfort that we use all three sensors, that we use a combination of sensors. Chris Caplice: Yeah, it's it's actually amazing. When was the first DARPA challenge? Was that 2006, 2007? Where where Eyal Cohen: Yeah, it was a it was a long time ago. Chris Caplice: none none no one finished, right? None of the none Eyal Cohen: Yes. Chris Caplice: of the vehicles finished. And it's only it hasn't been that long. It's pretty amazing how fast, even though a lot of people thought, you know, it was gonna show up next year for the last 10 years, but it's really, really moved. let's talk about the unit economics, because w a lot of times people will justify autonomous vehicles for long haul. Because of hours of service, right? It can go a thousand miles in a day, it doesn't need to rest, all that kind of stuff. That doesn't apply to your approach, right? So where where where's the economic value here for the for the humble approach? Eyal Cohen: Yeah, so you know, i if you think about the skateboard model, right? It's like this it's like a very simple vehicle, ultimately. It's it's very, very simple. But what what we can do with that vehicle is we can take night shifts, we can take we Chris Caplice: Yep. Yep. Eyal Cohen: can take we can take the hours that drivers don't wanna do, right? Chris Caplice: Yes. Eyal Cohen: we can take the routes that drivers don't wanna do. And so we can augment I've said that word a few times, but it's it's we can add we can basically add to the supply chain in various ways we can fill in the gaps. I think that's a very important piece of the puzzle, right? It's like if you can't find enough drivers to move containers out of a port at three in the morning, great, you have a mechanism to do that, right? If you have some some back and forth work that's not very exciting, or in faraway place or you know, there there's a lot of permutations of it, you can have automated equipment do that. and so the yeah. Chris Caplice: So yeah. Yeah, no that that makes sense. You made me think of something talking about the you the three am shift coming out of a port. How do you think the longshoremen will feel about this? Eyal Cohen: I think I think automation in in general is is is somewhat of an inevitability, right? Chris Caplice: Yeah. Eyal Cohen: and you know there's there's today that we have you know for example you go down to Long Beach, Port of Long Beach, right? There are there are some automated terminals and there's some non-automated terminals. And the automated terminals, you know, are are more competitive with other ports and other terminals out there. So so we'll see we'll see automation coming. I think the the right way to think about it is it's it's just technology that we are adding and to these areas that makes them more competitive. how that evolves on the labor side, you know, I think it's T V D in all aspects. Yeah. It's gonna be interesting, yeah. Chris Caplice: Going to be interesting. Yeah, because for trucking, which used to be heavily, heavily unionized, obviously, until the 1980 Motor Carrier Act and all that kind of loosened some things up with LTL, there's only one or two left. So it's probably not going to be a hurdle for the actual operations. But I'm curious for the, like we said, the touch points. And so there are some workforce unions that are very anti-technology or automation opposed for that. But we'll see. We'll see how that. How that evolves. Eyal Cohen: Yeah, we'll see we'll see how it evolves. I I I think I think what we would want is our ports and terminals to be competitive, as competitive as possible too, right? So that there's that you know, and we see automated ports springing up all over the world, right? So why n why not in the US too? Chris Caplice: That that's a question many people have been asking for a long time. but you gotta watch on the waterfront to understand that one. That's that's the movie that helps there. Let's talk about regulatory. What's Eyal Cohen: Mm-hmm. Chris Caplice: the biggest hurdle when you talk to regulators for this? you talk about the skateboard. I imagine they their head kind of exploded the first time you proposed this. Eyal Cohen: Actually it wasn't so bad because you know there there are this is what you would call a purpose-built autonomous vehicle. It's a it's a vehicle designed for a world where we're autonomous completely. And on for the regulators, you know, we're talking about NHTSA and and FMCSA, just you know, the people that are involved in in freight, you know, what are they concerned about? They're concerned that this is a safe operation, right? And There have been other purpose-built autonomous vehicles that have been out there. For example, on the Robo Taxi side, there's a vehicle manufacturer called Zooks. You they have a they have a vehicle with no steering wheel, right? And so they they have worked through a lot of the qua challenges, I would say, on, hey, if we do have a purpose-built vehicle, how does that, how will we deploy it? How do we get it in safely, right? And you know, the for the regulators, what they want to know, you know, they're they're committed to to innovation as well, right? And they want to see us, they want to see us progress. If you look at, if you go to China, for example, you will find skateboard platforms are kind of ubiquitous or becoming, right? Yeah, Chris Caplice: Really? Eyal Cohen: yeah, yeah. Yeah, they're they're they're delivering a lot of freight on skateboard platforms, not class A freight necessarily, but but certainly smaller freight. And so you know they're excited about promoting innovation provided that we are doing it in a very safe way, that we're testing in private closed courses, that we are that we are managing the the technology rollout in a responsible way. And you know what I always say with the regulators is they don't like surprises. You know, you don't want to surprise you don't want to show up their doorstep with the front, you know, with something on the front page. And so we just work with them every step of the way, which I think they appreciate. Chris Caplice: Now you one of the things you said earlier that you realize how much drivers do beside just driving. And one of the things if there is a a breakdown, they put out safety triangles. And I don't Eyal Cohen: Mm-hmm. Chris Caplice: I know there was a recent FMCSA ruling, I believe, where the idea of having neon signs or something on the truck itself was not sufficient and that they required you to have triangles. Are you familiar with this? Eyal Cohen: I am, yeah. I've been I've actually been involved in that effort, previous efforts. Chris Caplice: Can you give us give us some background 'cause I kinda butchered the explanation. Eyal Cohen: Yeah, so so you know, as you said, you know, there's a r there's a requirement by law that the drivers deploy triangles. there are i if you are if you are automating a tractor Typically what other companies have done is they don't touch the trailer, right? So they Chris Caplice: Yes. Eyal Cohen: say the trailers can come from anywhere in any condition. And speaking of learnings, trailers come in all kinds of conditions. That's that's something that I think Chris Caplice: Kind of like pallet. Eyal Cohen: it's you know, it's like it's an interesting, it's an interesting challenge for the drivers and non-drivers alike, I would say, is that trailers come in all kinds of ways and shapes and forms and conditions. and so drivers would leave the vehicle and deploy safety triangles behind. in an automated tractor where you don't have a driver, there's no way to deploy those triangles because you're not touching the trailer. and so there were there have been proposals. I think the original one actually came from Waymo when they were doing trucking, but there have been proposals there have been proposals to use lights that are high up and flashing and of certain luminosity to replace those triangles. and I think there's been rulings sort of back and forth and And it's not, as far as I know, I might be miscor incorrect here, but I don't think it's been codified in a while what the right answer is yet for an automated Chris Caplice: Yeah. Eyal Cohen: vehicle. one of the reasons I liked the skateboard design is because we could deploy triangles if we needed to, right? Because now you have access to the back of the vehicle, which means Chris Caplice: You right. Yes, yes. Eyal Cohen: you can see behind you, right, which is important, right? You can see behind you and also you can act behind you. And so we wouldn't at Humboldt with at least with the skateboard design, wouldn't necessarily have that challenge. but it is a challenge for an automated tractor. Chris Caplice: Yeah, it it seems like the there's trade offs because there's something very positive about having an autonomous tractor that can deal with any tr with any trailer, because that gives it more flexibility and that, but then you lose that trade off. So that that's interesting. So will you you are you're faced with that challenge then when you develop an articulated cab, to an articulated tractor, then so Eyal Cohen: That's right. Chris Caplice: you'll have that that problem. Eyal Cohen: Which is, you know, that and that's why that's why we're working on a family of vehicles, right? Because there's different use cases for different customers. You know, the the the the back and forth skateboard design works really well for certain use cases, you named a few, right? For customers that are really dealing with a trailer pool and different trailers, then a then a tractor, an automated tractor concept works a little better. and so we're just servicing, you know, what our customers' needs are. Chris Caplice: So isn't it as you get further into it? I know when you first start, you think a truck's a truck's a truck, every lane's the same, right? All this stuff. And as you get deeper into it, you find that there's so many different degrees. And it's such a nuanced industry that seems so simple at first. but there's so many different use cases for it. and then the other big thing is we talked about the touch points between shipper, receiver, and all that, but then you've got, you know, governments, state governments, and they have very different stands on different things and acceptance on things. So can you tell us a little bit about the geographic regions that you're focused on for this? Can you t give us any info on that? Eyal Cohen: Yeah, so we're focused on Texas and California. and and the and the yeah, yeah. No s Chris Caplice: Well that that's a shocker. That's a shocker. That's surprising. Eyal Cohen: no surprise. and and maybe the states in between, right? And you know, for for a lot of the autonomous players, Texas has become kind of the the the focal point for a number of reasons. One is that it's I believe the largest freight state. You most most amount of freight moves through Texas. the weather is generally cooperative, generally. I've seen some extremely large hail hailstones coming in Texas, extremely large. Chris Caplice: Yeah. Yeah yeah. Eyal Cohen: but the weather is generally cooperative and there's a variety of use cases. There's ports, there is highway, Chris Caplice: Yeah, yeah. Eyal Cohen: you know, there's there's a there's large growth there. Chris Caplice: And Eyal Cohen: yeah. Chris Caplice: yeah, the the Texas triangle itself, what is it? Dallas, Austin, San Antonio, just constantly has freight moving. Eyal Cohen: Houston, yeah. Yeah, Dallas, Houston, San Antonio with Austin. It's kind of like a stop in the middle there. It's just like, yeah, this Chris Caplice: Yeah, yeah. Eyal Cohen: this this constant this constant when when you're there, you know, the scale of the infrastructure is impressive, actually. So Texas is great. and and we know the regulators there. yeah, it's flat, which is which is Chris Caplice: And it's it's flat. Yeah, yeah. Eyal Cohen: good. California, California has become an interesting freight state. I they're very excited about electrification of trucks, right? There's good incentives Chris Caplice: Yes. Eyal Cohen: for that. you know, there's there's a lot of desires sort of to get the air cleaner around the ports. there's challenges too, right? There's challenges with charging and and finding the the right the right amount of energy for the vehicles. Chris Caplice: But but that's a really good point. Sustainability is is that a a big play for your market? 'Cause it doesn't seem like it's something it seems like a ancillary benefit. It it's not like Eyal Cohen: Yeah, sustainability is is great. I mean I think you know if you can make the air cleaner with your vehicles, amazing. It's actually not the primary motivation for it to be electric. primary motivation for electrification on the trucks for me is that that's what really makes sense for Thomas vehicles hand in hand. That's the that's the idea. Chris Caplice: Yeah. Yeah. And then the the other thing is for the ports is sound. I know that's a major issue for a lot of people. Just having trucks idling for a long time is just annoying. So s that the EVs are are pretty much silent. Eyal Cohen: It's a wild thing to see our skateboard design driving. and it is aside from springs and air compressors, it's silent, which is a it's it's Chris Caplice: Now is that a problem? Is that Eyal Cohen: it is actually an issue and we have to introduce sound into the vehicle because of that, Chris Caplice: Yeah. Eyal Cohen: yeah. Chris Caplice: Dude I d I'm I'm trying to think for any kind of autonomous EVs for cars, do they add some kind of sound because you don't want sneaking up on on people. Eyal Cohen: Yeah, the EV providers add sound. They do it a variety of ways. Sometimes mechanical, sometimes like with literally a speaker. But it's it's one thing to have a car that's silent, it's another thing to have a a class A truck silent. You know, you're like, well, Chris Caplice: Sure. Yes. Eyal Cohen: okay, we need to hit we need to know that this is here. It needs to make its presence known, right? Which in the past wasn't an issue. Chris Caplice: Yeah. So let me ask a just a long term question. I I haven't asked people this for a long time. How many years in the future do you think it'll be or how many generations before someone talks to their kids and their kids were shocked that we actually drove our own vehicles? Eyal Cohen: So we thought 10 years ago that our kids wouldn't have to get driver's licenses. Okay, so that was 10 years ago. But they're still getting them. I know it, yeah. They're opting out. They're opting out. They're Chris Caplice: Well now the c a lot of these kids are not getting driver's license, right? They're they're they're it's Uber. Uber and Lyft. Yeah. Eyal Cohen: opting out, but not not for autonomous reasons. so look, I think I think it'll be there what there's something like three million class A trucks out there, right? And who knows how many cars are out there. It'll be a long time before automation has reached every single vehicle. I do think automation will come for every kind of vehicle in class type. I think that's sort of inevitable given where the technology is going. But our our kids will get driver's license, surely. there's just so much freight that moves out there. There's so much, there's so many cars that people want to drive. People have different reasons and purposes. the generation after, though, I think that's a good question. We'll see. Chris Caplice: Yeah. Now you mentioned some other usage. So of was it Pronto that was bought by John Deere? What was Eyal Cohen: n I worked on a company called Spark AI that was bought by John Deere. Yeah, cool. Chris Caplice: Spark AI, okay. And so was that developing the same type of was it A V or was that E V? Eyal Cohen: It was A V Tech. We were working on we were working on remote driving. That was what the concept of that company was. we, Chris Caplice: Okay. Eyal Cohen: you know, it's one aspect of autonomous vehicles that typically almost every autonomous vehicle, this is the case, there is a command center somewhere where they are watching over the vehicles and doing some remote driving if if need be. And it's actually it's actually where you know we expect some of the jobs to move over time. It's just people being responsible for the vehicles. Chris Caplice: And will that happen for humble as well? Do you s do you see for the skateboards Eyal Cohen: Yeah. Chris Caplice: where you need a human now will that be something the company, if I'm I don't know, I'm an LTL carry that I will have my own room running these? Or do you view this as something that you'll provide as humble? Eyal Cohen: We will provide yeah, we will provide it because we understand we will understand the limitations of our vehicles and what to do about them, right? But over time over time the the way that this will evolve is it can evolve in a variety of ways. I think you'll have probably command centers and at every carrier, you know, to manage the remote fleet. Chris Caplice: Well this this raises another question I should have asked earlier. Are you are you gonna become a carrier? Or are you gonna become an OEM provider? Or Eyal Cohen: We're we're open, you know, we work with our customers in a variety of ways. You know, sometimes the a shipper wants us to move freight directly. Mostly we're looking at working with existing ecosystems, so carriers working with the carriers. Chris Caplice: Okay, so a carrier would procure X number of skateboards, but they would operate them? Eyal Cohen: We would operate them in tandem with them. We Chris Caplice: Okay. Eyal Cohen: would you we understand what the robot can do and what its limitations are, right? But ultimately we're providing technology. That's what we do. Chris Caplice: So I guess it's two different types of they they would control loading. I mean not loading, but utilization. Like I need to go from A to B to C. You would be providing oversight mainly for safety and maintenance. Is that the idea? Okay. Eyal Cohen: You got it. You got it. Yeah. So we so we'd be making sure that the operation is safe, that the vehicles are doing what they're expecting to. It's like Chris Caplice: Okay. Eyal Cohen: it's like we build a robot, right? It's a big robot, but it's a robot. And if we are if we're putting the robot in the hands of somebody else, we have to make sure that they understand all the limitations of the robot and that we need to be responsible to make sure the robot is doing what what we said it would do. Chris Caplice: Okay. So when you go home for Thanksgiving and your your grandparents come to you, how do you explain what you do to them? Is that is that a hard thing or is that an easy thing at these days? Eyal Cohen: Yeah, I it is it is a it is a it's a great question. It is it is not an easy thing to explain that autonomous vehicles are present for people that haven't seen them. One thing that Chris Caplice: Yeah. Eyal Cohen: happens is that when people often get into an autonomous vehicle, they go from wonder to boredom very quickly. Chris Caplice: I the economist had a great phrase. It goes from what miraculous to mundane in like two minutes. Yeah. Eyal Cohen: It's a great way to put it, yeah. Yeah, you you forget almost immediately that you're dealing with an automated vehicle. and so but you know to a person observing, you know, they see, for example, automated tractor without something in the front seat, that is a that is a sight to see. Chris Caplice: Yeah, we're but we'll get used to it. We'll get used to it. Like in you said in San Francisco, you you get used to seeing Waymo's just tooling around without a driver. Is it it's pretty common. In Eyal Cohen: That's right, yeah. Yeah. We are we've we've become we've become here in San Francisco kind of a I we've probably even lost our sense of wonder. You know, you're like, you know, Waymo, get out of the way since I can cross the street, you know, it's like that kind of thing. Chris Caplice: Yeah, yeah. If it in Boston it we would freak out. We just aren't we are so we don't think you know, we're different. You'll never figure out the streets here in Boston. Eyal Cohen: You're the second person to tell me this week that Boston would be a place that wouldn't see automation for a while. And I just I t I'll tell you what I told the person this week. I was like, yeah, it's coming, it's coming for you too. Chris Caplice: It's it's coming, but the question is when. we don't we're not really f you know, for MIT being here, we're kind of a tech and biotech kind of a a a big leading area here. We don't like to autonomate automate a lot of things. It's it's weird. We'll see. We'll see. We especially on the roads. we'll see. But AL, thanks so much. I I really learned a lot about Humble and now that you've come out of stealth, you're still kind of stealthy. we still know who you're partnering with. What what questions should I have asked you that I didn't? Eyal Cohen: I know you asked great questions. I think you covered the gamut. you know, you Chris Caplice: Okay. Eyal Cohen: we talked about the vehicle, the software, the the the the business model. yeah. We're Chris Caplice: I didn't ask you how much it cost. I know you you probably you're probably sh figuring that out. Eyal Cohen: we're driving that cost down every day. That's what our engineers do. Probably stressed because I told them we have to get the cost lower. So yeah. Chris Caplice: Yeah, yeah. Yeah, but but but you're probably hoping for the price of diesel to go up, right? That makes it the economics so much better for you guys. Eyal Cohen: Yeah, you know, I I I I think you know, I wanna see freight I wanna see freight succeed. You know, it's I think it's great that we're out of this recessionary moment in freight, right? And it's sort of taken off again and I don't wanna see diesel go off. I w I just like to see the industry succeed. Chris Caplice: So it's funny here at DAT, we work with shippers, carriers and brokers. So when you know, one man's down market is another person's up market. So the shippers were were loving the last three years. And now the carriers are loving these Eyal Cohen: That's true, yeah. Chris Caplice: times. So it's always someone's always happy, someone's always sad. Yeah. Yeah. Eyal Cohen: Someone's always up, someone's always down. But you know, like not everything has to be zero sum, like, you know, rising tides and all that. Chris Caplice: Hopefully. Hopefully. Hopefully. Well, thank you so much for joining me today. Appreciate it. Eyal Cohen: Thank you. Thanks for having me. Chris Caplice: All right, everyone, stay tuned for the truckload market update. Hello, this is the truckload market update for 17th September 2026. I'm recording this in the hotel at the annual shipper summit in Atlanta. So excuse the backdrop of the hotel. but let's get started. So in Dryvan, we saw the change in active contract rates increase 0.4%. We saw spot rates also increase 2.4%. And the current level of replacement rates are a whopping 14.6%. That's almost double what it was. In the previous two weeks. The market gap between spot and contract is 24 cents a mile. That's increased. And the year-over-year change in spot rates is about 45.6% for dry-van and for contract rates, it's 14.6%. Okay, temp control. We saw contract rates drop 1.4%. Spot rates also dropped a point three percent, small amount. current level of replacement rates for temp control is 9.1%, meaning new rates. Contract rates coming in are about nine percent higher than rates that are leaving. the gap between spot and contract is four cents a mile, where spot is above contract. Year-over-year change in spot rates is twenty-three percent for a reefer and eleven percent for contract. So, intermodal, we saw that the contract rates increased two point five percent. Spot rates, although there's not much of it, increased one point five percent. The current level of replacement rates for intermodal is four point two percent, and that's a big switch from two weeks ago where it was a negative. And the market gap between spot and contract is negative four cents a mile, but again, not a lot of spot intermodal. Year-over-year change in spot rates is 9.6%. Year-over-year change in contract rates for intermodal is 10.6%, about the same as it was last period. For flatbed, we saw contract rates drop 2.2%, spot rates went up 0.6%, and the current level of replacement rates is about 9.1%. The market gap is about 10 cents a mile between spot and contract, spot above contract. Year-very year change in spot rates is about 28%. And for contract rates, it's about 13.1%. All right, so what are we seeing? Same kind of bouncing around, drive-in and intermodal contract rates increase slightly, temp control and flatbed contract rates had slight decreases, one to two percent. Replacement rates for all the modes are positive, four to fourteen percent. That 14% is a big jump. We're seeing this a lot in the new rate differentials for new bids coming in. is what we're seeing for and the year over year increase is about ten to fourteen percent, which kind of maps, right? The replacement rates and the change in contract rates. Those are kind of very highly correlated. spot rates increased one to two percent, except temp control, which drops slightly. We did see over the last month or so kind of a cooling off period after the fourth of July and in through August kind of cooled off. We'll see what happens in September if the rates start peaking up again. And the gap between spot and contract is anywhere from five to twenty-four cents per mile, except intermodal which dropped. But again, Intermod's an outlier for that. So that's the truckload market update for 17th September, 2026, as we head in to the wrap up rather, the third quarter of the year. And we'll see what happens for the remaining quarter to see if this market continues to be tight or if we're seeing some green chutes where thick capacity might be entering and we might see some easing off. I'm kind of banking that we're not going to see any easing off until Q two of twenty twenty seven, but we'll see. Thank you.