Dexterity's Bet on Robots That Can Handle an Unstructured World

[Music]

Cheers. Cheers, manufacturing coffee chat. It's another type of brew. Mhm, it counts. It would be hard to do every interview over beer with the frequency that I'm doing interviews now, especially now that you are a third time guest of this. Oh wow. Yeah, with all the people I interview for the first time, second time, you can imagine it would stack up. That would be a lot of beer. So what happens when I join the five timer club? I don't think I've gotten there yet, actually. I need to make, we'll come up with a good celebration for it.

All right. I would say your three time celebration though will be our party at Bear Bottle, Mara. It'll be long in the rear view mirror after this recording comes out, but we're recording this on Tuesday, January 9th. We are throwing a nice little shindig here in the Bay Area tomorrow at Bear Bottle Brewing, one of my longtime favorite breweries before I know it's, we're down here in Redwood City, which is about, I don't know, 30 minutes south of San Francisco. Have you been to Bear Bottle before? I've not, but when my colleagues saw that that's where we were holding it, they said great spot. Yes, and that shouldn't surprise me. Sure. But sure, well, it's a good size, it can hold a lot of people. You know, obviously I do love beer, and I kind of started all my podcasting and beer oriented podcasting when I was out here in the Bay Area, so it'll be fun to get back to an old favorite.

And it's funny, I was listening to your old episodes as I was getting ready for this. I can never remember the name of the brewery you suggested in the first one, because it's Mighty Squirrel. Mighty Squirrel. Mighty Squirrel outside of Boston, but I'm always like, is it Mad Squirrel, is it Barking Squirrel, like I go through all of those. I don't know why I think Mad Squirrel. There was a band when I was in high school called Narcoleptic Barking Squirrel, so I know why that one confuses me. Wow, that's a great high school band name, by the way. Just throw a bunch of crazy words together and throw an animal in there.

Anyway, we are off topic. We're here to talk about you, we're here to talk about Dexterity. And in the spirit of having a good mey conversation over a beverage, I want to go back to something we talked about in the past, because I've interviewed you when you've been with Dexterity, when you first started with Dexterity, when you were kind of at the end of your stint with Boston Dynamics. And what we've talked about before is how you love being part of transformational change. I think this has actually come up in both our conversations. You did drones when you were at DJI, you were working on Spot when you were at Boston Dynamics. So my first question is, what transformational change in the robotic space have you seen in the last two plus years that it's been since we've last had a conversation like this?

Yeah, you know, Dexterity, the thing that was so exciting about Dexterity is that we are at the forefront of this next big change in robotics and automation. With DJI and Boston Dynamics, a lot of that revolution was about mobility, that you were able to get automation into places that previously seemed inaccessible to robots, whether it's through the air or, you know, rough terrain. Dexterity is at the forefront of robots that are able to manipulate the world in a way that we never really thought was possible with robots. Manipulate meaning physically interact, touch, move things in the physical space.

The idea of dexterity, which is combining multiple levels of intelligence so that a robot isn't just able to pick something up and drop it off, but problem solve and manipulate, and when it fails it tries again and tries new approaches, and can problem solve by stacking things stably, that was still fairly new in the industry two years ago. It's only this year, this past year in 2023, that you're starting to see more and more companies starting to use that approach. Tesla, you know, you see some of the problems that they're trying to solve with Optimus. It's a lot of the same idea, where you have an unstructured world that a robot not just needs to get through physically, walking, but then also needs to figure out what plant to water and how much to water it. And those decisions are a lot more complicated than what robots used to be able to do until now.

Yeah, and we just took a tour of your facility here in Redwood City, which we'll talk about in just a second, where I think one of the things that stuck out the most is we were watching robots load a truck, right? And as we were watching that, you know, I was asking why are the boxes being stacked the way that they are, and you described what is it, a T? What is the official term that you used for it?

Yeah, it's a, you know, you do tacking basically. Yeah, so you have two boxes that are together like this, rather than columnar. Yeah, yeah, like box on top of box on top of box singularly. When a truck starts and stops, boxes fall over, right? If you've got a really heavy box on top of the smaller box, that smaller box is going to get crushed pretty easily. The nice thing about, you know, a nice T junction of boxes, kind of brick layering the box. I say it's like layering bricks, basically. Exactly, it's the same idea behind that, where you're distributing the weight, you're also creating a lot more stability, so that when the truck starts and stops the wall doesn't collapse in on itself.

This is something that anybody who's loaded a truck has been trained to do. It's part of the standard operating procedure for loading a truck. If you just give a robot a bunch of boxes and say stack this, you know, in the back of a truck as densely as possible, they might go towards a column or builds of stuff. But knowing those requirements from the customer, we're able to train the AI to problem solve in this specific way, so it actually fits seamlessly within the customer workflow.

Yeah, and one thing that has jumped out at me, having known Dexterity for a few years now, as we've been walking around your facility, this is a bit of an alternative startup location. Obviously the Bay Area is a perfect spot for a startup, but you've kind of taken over a strip mall and you've now expanded into what was an old Kmart for your headquarters, your testing and engineering locations. Let's just go over the different spots that make up Dexterity HQ right now, because when we interviewed two and a half years ago you had just gotten your Series B round, and I think it's always cool to see how companies have scaled during that time in terms of just your physical presence here. You started in an old auto shop, correct? And we're, I think, in an old dollar store where we're doing our interview right now, right? And then we were in a Kmart. Is there another spot I'm missing? Like, wasn't there a soccer store or something in there?

Yeah, this old soccer store where people used to get equipment specifically for indoor soccer, and it's kind of a bummer. You'll see families pull up and they'll be like, we're here to get shin guards, and we're like, wrong place, they moved a few blocks down. Yeah. The place that's right next to us used to be an antique clock repair shop. Okay. Which is kind of fitting, because we're tinkering on some of our old robots in that space. Sure. So yeah, I mean, the team has been expanding like crazy. When we spoke the first time, I think we were about 100 people. Yeah. We're now pushing 260 all in, with a wide variety of people across the country and across the world now, in both Japan and Taiwan. So, you know, the team is growing like crazy.

And one thing I do want to also ask is, it's like when I came up, it does not look like you're walking into a startup office until the door opens, right? Right. But you see the D on the door for Dexterity. It's like a startup speakeasy, I think, is what we were calling it out here, which is very cool. Now, you talked about the people, we've talked about the physical space. Yeah. What's scaling looked like from what Dexterity is doing? What's that looked like over the past two years? We've been covering our bases so far, but let's talk about the company itself and the type of work you're doing at this stage.

When we first started talking, we had initial success in bread manufacturing distribution. Mhm. We picked bread packing as our first task because it was one of the hardest tasks that we could imagine to stress the full scale of our robotic system. Mhm. So in bread distribution you're not just having to, you know, pick stuff up and drop it off. You need to pick things up delicately. You can't squish them. Like, robots love to turn bread into pancakes. Sure. And so our sense of touch is able to delicately pick up one, two, three loaves at a time, place them into a tray, and not just drop them but gently nudge them so that the tray space is occupied.

And then you have fleets of robots that are all working together to get this job done. It's not just one robot picking and placing it. You have robots on a rail that are going up and down, and they're coordinating with each other to figure out, you know, how they're not blocking each other, and they're coming together to do bimanual manipulation. When one robot can't pick something up, the two of them come together and they pick it up together. You know, all of those were complex pieces of the Dexterity AI platform. Okay. That was kind of our first build of all of these different vision systems, sensing systems for sense of touch, planning, stacking, you know, all of these different components together.

Now we've taken those individual pieces and have been able to say, okay, we're going to take this vision piece and this dual robot collaboration piece, and we'll use this to induct packages, kind of like what you saw, where we have two robots that are working together to grab these messy flows of packages and get them onto a sorter belt at 2,000 picks per hour. Mhm. You're using that same stacking and tray optimization logic, kind of the next version of that, to build pallets. And believe it or not, a lot of that same code is what we use to start figuring out how to load trucks. Yeah. So, you know, you built this core of a technology that's now spreading out into all of these applications that really cover everything from the inbound to the outbound of the warehouse distribution cycle.

What's really exciting, though, is that, you know, that bread system is very bespoke to one customer and their particular challenges. The most exciting thing is starting to see these solutions being deployed at multiple customers who represent multiple industries. So we've got the same system with a retailer, a third party logistics provider, and a manufacturer. Yeah. And it's, you know, our palletizing and depalletizing system. It just drops in at the customer site, gets up and running in about 48 hours, because it doesn't require a lot of integration. It just kind of gets up and running and works and learns with the customer's system as it starts going into production, and it's ramped, you know, very quickly, within a month, to their full production capacity.

So you went from just bread packing to basically covering, what you said, the entire inbound to outbound portion of a warehousing and distribution operation. And this was over the course of, I mean, this might have predated you a bit, but what was the time frame from beginning to where we are now? Rough is okay too. Yeah, I think the first bread system was installed, I think, in 2020. Okay. So yeah, it's been about three and a half years in total. Still pretty quick. Yeah, that's a lot of progress over a relatively short period of time.

But you know, Chris, that is going back to that transformation question that you had. That is the transformation of using a software first approach to these problems. Yeah. So, you know, the hardware in a lot of ways is there. You have ubiquitously available six degree of freedom robotic arms, you have camera systems that have been, you know, somewhat harder to get a hold of more recently, you have GPUs that are incredibly powerful, incredibly affordable, suction pneumatic grippers, force sensors. All these components are out there, but using one software framework, unlocking them, really powers a huge transformation in a really short amount of time.

Yeah, and another thing I wanted to ask you from one of our old conversations, because I do want to continue to talk about where Dexterity has gone, but I also want to go back in time a little bit. When we talked about your role at Boston Dynamics, you talked about how Spot, the famous quadruped dog that everyone in industry knows, you were originally betting Spot was going to be for utility applications, and it turned out it was good for manufacturing inspections. Sure. The reason I bring this story up is, has there been something that surprised you over these past few years here at Dexterity about the technology you've been developing that's opened new doors? We just talked about some of the directions you've been going, from bread packing all the way to where you are now, but are there other new things that came up along the way as you built out this solution for logistics and warehousing that maybe you didn't expect early on?

Yeah, I'm trying to think of all the things I can say. But in terms of, so I think there's a few things. One is the number of verticals where this type of technology is applicable way exceeds our expectations. You know, we were originally thinking, oh, some manufacturers doing end of line manufacturing, maybe third party logistics providers that are kind of that middle chunk, and then retailers who actually get stuff to customers, that's the full story. But as you start looking beyond the horizon, there's a wide variety of verticals where something like this can really help companies with their material handling challenges. So I'll leave that a little bit ambiguous. As we are in the heart of Silicon Valley startup world, I'm well aware we can't share every application, every customer that you're working on. It's part of the game.

But one of the things that's really interesting and exciting for me is seeing the transformation from a single task focused piece of automation, let's take our, to what we think of as a full task automation. So let's take palletizing and depalletizing as an example. A single task piece of automation would be depalletizing. You grab something from a pallet, you move it to a conveyor, grab something from a pallet, you move it to a conveyor. Okay, we've added this thing so that you can ensure that you're barcode scanning as you're depalletizing. That's pretty cool, that adds another piece of the value equation there.

Late last year we introduced something that we call our print and apply solution, which is a task in a warehouse that requires one, two, three people depalletizing something. They put it on a conveyor, or they just have a little gun and they'll, you know, attach a label to it, and then they'll repalletize it. So that can be one to three people, or three different tasks that a person is doing: depalletizing all these things, applying a label, and putting it on a pallet. Yep. We now have, because of our system's flexibility to do both depalletizing and palletizing and do some of these other things in the middle like barcode scanning or whatever, we've now created one robot system that depalletizes, applies a label, and then repalletizes, and it does it at basically human parity in terms of speed.

What's cool about that is it's moving, now that might sound like a pretty boring example, but what's cool about it is that it's three tasks that one robot is doing. So you would typically think of robots plus a conveyance system plus all these different pieces of equipment trying to solve this problem together, where it's, you know, one robot and a labeler, and that's your solution.

Well, I think it goes back to your earlier comment about, hey, there are a lot of physical robotic solutions out there, but the reason the software development is so important right now is that you're now able to do something like three different tasks with one robot, where in the past it would have been, you know, three different robots doing that same one task over and over, maybe even more. To this point, when we were looking

...inside the truck earlier, you have robots that are leveraging an algorithm to figure out where the optimal spot to place a package is. So there's some decision making going on there as well.

Yeah, absolutely. And you know, if you think about that task specifically there, you can solve that with a lot of quote-unquote dumb automation, where it's just doing what you tell it to do. And you could do that by saying every box is going to be this size, shape and weight, figure out a way to stack it. You can do it that way, but that requires a high amount of upstream optimization, where you know every package for the 3,000-some-odd packages there going in the back of a trailer.

That's great if you have that, and some companies are investing into that type of full end-to-end automation transformation, but the vast majority of companies don't have that capex to make that transformation. So how can you have robots kind of slot into your existing workflow without having to drop tens if not hundreds of millions of dollars?

Yeah, well, let's talk about this a little bit more. I was going to ask you later, but in addition to your appearances on Manufacturing Happy Hour, your CEO and founder Samir has been on the show before, and he was talking about how part of his mission has been to democratize robotics. So let's talk about how, you know, you do work with a lot of large organizations, but how are you seeing robotics get further democratized, since we were just on that train of thought a second ago?

Yeah, great question. So a lot of our initial customers, or customers that have been speaking about our work together, our collaboration, include some of the largest names in parcel distribution and retail. But the exciting thing for them is the same thing that's exciting for the industry as a whole, where you have a very flexible piece of automation that can slot into their existing workflow without them having to change what they do to accommodate automation. And, you know, if they need it in one site one week, another site the next, or even within the same shift, it has that flexibility and mobility to get that done.

What does that imply? That implies a much shorter time of deployment, yes, but it also means a lot less upfront investment in all the infrastructure and software integration and all these other things that are typically necessary to get a traditional piece of automation up and running on a customer site. That means also that we can start going to the brownfield sites for much smaller companies that don't have the resources to invest tens of millions of dollars into an automation project. They can have a point solution that helps them with the single most painful task in their warehouse, and they can get that deployed within 24 hours, 48 hours, rather than having to gut their warehouse and put something else in.

So I'm hearing quick deployment as well as robots with flexibility are two of the big things that are making it more accessible to everyone. I'm going to totally switch gears on this conversation for a bit. Putting on your marketing hat — you're wearing a hat today, so no need to switch the actual hat.

I just got off a plane, so my... You really did. I mean, for context, I got here like 5 minutes before you did. As I'm signing in, you're walking through the door as well, so I appreciate your level of ambition to jump on a podcast after flying from the East Coast to the West Coast. You have had a full day. It's been worth it. It's worth it.

Well, the reason I want to switch gears is because you're putting on your marketing hat, because you lead marketing here at Dexterity. How have you seen Dexterity develop more into a brand name in the industry since you started in 2021? That was one of the things you had talked about before, where it's like, hey, I've worked for some companies where the brand name's already there, but with Dexterity you were excited about going into the new frontiers and the applications. But I hadn't heard of Dexterity when I first talked to you; that was my first intro. I'll share some of my thoughts on Dexterity branding here in a little bit, but this interview is not about me. I want to hear how have you seen the Dexterity brand grow over the years.

I think your opinion on this is much more important than mine. Like, Chris, I don't know if you experienced this, but sometimes when you put something out there, the worst thing you can do is flood your channels with information about it, because what you hear back is your own echo. Sure. Right. Yeah. So, you know, I have a perspective on the Dexterity brand and what we're hoping to achieve and how far we've achieved it so far, but more importantly, it's what do people... your perception of it, I think, is even more powerful.

I'll bite. I mean, I think it's great, because, like I said, I hadn't really come across Dexterity much in 2021, and granted, Dexterity was pretty young back then, so I think there's no surprise there. But I feel like I keep seeing you around the industry on a regular basis, and I made the comment before about feeling like I was walking into a startup speakeasy, because you have such a defined... you know, your letter D is pretty defined. And obviously the logo and the fonts aren't the main thing behind a brand, right? But there's all these things at play that make me continue to feel like Dexterity is a key player in the robotics and warehousing game right now. That's my impression so far.

I'm thrilled to hear it. Yeah, you know, it has been a challenge, because both Boston Dynamics and DJI are the easiest cold call in the world. Sure. You're like, I'm from Boston Dynamics, do you want to talk about robots? The answer is yes.

That's what it was like when I got the email saying someone from Boston Dynamics wants to be on your podcast. I'm like, yes, I will take that all day long.

So we're not in that space, and you know, I've been thinking about this, like, is that important for Dexterity? And considering that we are very focused on executing with the resources that we have, the amount of investment that we have — we were talking about this before — you know, we're trying to be very judicious with what we have. It's less important to be out there in front of everybody; it's more important to be out in front of the right people. Agreed.

And so we are trying to be a little bit more focused on building very custom awareness within the industry to the people that, at least for now, are going to benefit the most from our technology. You know, the hope is that that success then elevates the brand as a whole. Like, we're doing really, really cool stuff; we want the world to know about it. Yeah. But in the interim, it's more important for us to have the people that could benefit from this technology now get excited about it.

Wholeheartedly agree. It's about the right audience versus the biggest audience. You always want to be talking to the right people. And I think, as someone that is a branding fan myself, I like that Dexterity is a one-word name where it gives a pretty good impression of what you do when the right audience is hearing the name. It's like, oh yeah, you know, Dexterity for robots, right? You have a robot that, with the work you're doing around software, allows robots to have more dexterity, very simply put, right?

Well, I appreciate that perspective, and if you talk to our engineers, that's how they think about it. But from my perspective, what I hope we can engender is that we don't sell robots or software; we sell outcomes, dexterity for them, so that in their operation they have the flexibility to do whatever task is most important to them. And that's the business outcome that we want to give our customers: the dexterity to get the job done.

I like that. If I had been smarter with my comments, I would have said business outcome as well, rather than just what the tech is doing. But let's put this in a wrapper for the manufacturing leaders that are listening out there, particularly the ones that are marketing-minded, work in marketing, etc. Can you share some advice or tactics that you feel have been helpful as a marketing leader yourself in continuing to build the brand of Dexterity over the past couple years? I think your comment earlier about targeting the right audience is some great marketing 101 there. Anything else on your mind?

Chris, there's a revolution going on in AI that is not just impacting warehouses, but it's impacting every single industry, and marketing and sales is no exception. The powerful thing that I'm seeing — and again, if you're thinking about being targeted and focused in execution — it's no longer the spray and pray, let's put out a message and just hope it works, right? It's really understanding your customer and creating a custom experience of your brand for them. So, you know, we can create messaging, materials, microsites, emails, internal newsletters for specific customers that give them information about how we can help them and help them solve their problems. We can empower our internal champions at these companies to go tell the rest of their organization what a great partner we are to help them solve their problems.

You know, if you had asked me this, I'd say, five years ago, a lot of the distribution arms were there. Like, distributing information is no friction at all, but creating that individual piece of content was very hard. And now you have the flexibility to say, here is how this is applicable to you, let's show this in your workflow, let's create a custom site that shows you everything that's most important to you as a company that we offer, and stack rank it based on your priorities, and let me give that to you so that you can be empowered to go tell our story.

Yeah. So I wanted to ask you about AI in general, more specific to Dexterity, but I'll ask you around marketing as well. Have you found AI tools to be helpful in your marketing efforts?

I have the blinking cursor problem. Yeah. Which is a blank page, and you're just staring at it and you don't know how to get started. Yeah, you know, I think with AI you get a great — not even a great first draft; like, sometimes you just rewrite everything. Yeah. But you start from a standpoint of, I'm switching to my editing brain rather than my creating brain, right? And my creating brain and my critical brain are often too much in conflict to get things done quickly.

If you're trying to scale something like, you know, custom messaging for people, or heck, even writing a press release, right, that blinking cursor problem can take up a lot of time that can otherwise be used actually getting something done. Yeah. And I think that's one core piece of the AI puzzle that's been most beneficial to me. I don't know, have you used it much?

I have. I've been playing around with it. I look at it as an idea generator. I think, you know, if I think of a podcast episode, sometimes I'm just staring at the blinking cursor — I've never heard that specific term before, but it makes a lot of sense. That's going to be a simple thing I take away from this. But I have the blinking cursor issue when I'm trying to think of a podcast episode title a lot of times, especially if it's one that I recorded three weeks ago and, my own fault, did not just go and write all the stuff immediately following the interview. I can say a lot of the time when I do that, I usually have a fresh idea then, but when I'm trying to recall something three weeks later... Long story short, yes, I do have the blinking cursor challenge from time to time as well, and I've been leveraging AI tools to help get past that a lot quicker and come up with ideas.

And you know, I think you brought up a great point. Sometimes you have to go through and edit the whole darn thing again, because it does feel a little robotic — I'm being intentional about using that word in this interview — but it feels like it's kind of just very standard and things like that. You've got to go through and touch it up, but it is much easier to do that editing than creating in some of those cases.

For sure. And yeah, it helps you at the end of the day do more and create, at least in our world, very custom experiences.

So let's talk about AI in the context of Dexterity a little more directly. How is that undoubtedly playing into your solution and development these days?

Yeah, you know, it has always been a core piece of Dexterity's stack, where our website is dexterity.ai for a reason, and that's because we're using artificial intelligence to solve a lot of problems. But importantly, you know, if you talk to Samir, you talk to our founding engineers, they'll all tell you that it's not a silver bullet in and of itself. It works really well in conjunction with some traditional controls work, with some very standard computer vision approaches. You know, there's all of these different pieces that, when you start pulling them together, can solve really complicated problems. Like, you saw our palletizing robot, where you have random boxes being presented to the robot and it's needing to figure out how to stack them into a nice pallet, but it doesn't know what box is coming to it until it arrives, and it's able to take big boxes, small boxes, and create a nice, neat pallet.

Now, part of that's a vision problem, where you're trying to dimension the box and dimension the pallet and figure out where there's empty places. Part of that's a force control problem: when you're picking something up, understanding how heavy it is, how much force to put it on the pallet with, how much force to apply when you're kind of nestling boxes together, making sure you're getting nice density — all these things that the human touch has traditionally been very helpful for. 100%.

And you know, if you just use one of those approaches, which is kind of how we've seen AI deployed until fairly recently, where it's just like, well, I've got an algorithm for figuring out how to stack things, but then you don't have the vision component or the sense of touch, or you're not talking directly to the motor controllers on the robot arm so that you can translate from the conveyor to the pallet with the smoothest motion... You need all of these pieces in concert to get the job done in a way that's fast, efficient and, you know, at human parity.

So let's say we're at be bottle having a beer, when someone asks, how are you leveraging AI in your robotic software and solutions for warehouses? What is your one-sentence answer? Man, I just want to put a summary around it. I don't know if I can do a good job myself.

That's a good question. You know, I think we use multiple approaches in AI to solve the sense, think, move set of problems that a robot needs to do in order to tackle very complex problems.

That's getting turned into a quote, by the way. That was a great one-sentence answer, and I feel like I might have cut you off, because I feel like you might have had a follow-up to that, but that made a lot of sense.

That's perfect. You know, from our perspective, sensing: you're looking at the world, you're touching the world, you're getting additional sensor information from other sensors that might be distributed in the warehouse. You have the thinking: you've got to figure out how to stack these things and problem-solve this, looking for empty slots. And then moving: figuring out how to translate this as smoothly and efficiently as possible. So those are three areas of intelligence where, if you just tackle one, you're probably part of the way there to getting something useful done, but if you're really looking at human parity, a human level of flexibility, you really need all three to be working together.

Yeah, I think that's an excellent way to put it all together. And, you know, for context, I'm really starting to ask all of my guests how AI is impacting their business from a practical standpoint, because, you know, I know it's going to evolve very quickly. Technology is going to continue to change, AI is going to continue to get better and better, and

better, and I try to keep some element of timelessness to each podcast episode, but I can't ignore where we are in this particular point in time, and I love how you simplified that for the audience out there. As we start to wrap up, what are you excited about? What are some of the upcoming launches, developments that you have on the table here at Dexterity that you'd like to give the audience maybe a little preview to as well?

So we have a few things cooking that you'll see from us throughout the year. I think one of the key things is our palletizing and depalletizing system. It just continues to grow in its capabilities, and we'll be showing some really cool new capabilities of that system at MODEX this year. It's truly phenomenal what the team has put together.

But earlier, when we first started the podcast, we were talking about the breakthroughs in mobility and the breakthroughs through manipulation, or touching things, and I think the next big breakthrough that you'll see more of, and we've started teasing that with our truck loader, is mobile manipulation. So one robot system that's able to do a wide variety of tasks within the warehouse at different points, whether it's station A, station B, station C, or maybe station A is one task whereas station B is another task, and that same robot can handle some of that challenge. I think that's really where the next big breakthrough in our industry will be, and Dexterity's approach here is game-changing.

Well, I'm glad we got that part covered in today's conversation. I'm excited to see what's coming out throughout the year. I'm sure a lot of people listening to this are catching it like mid Q1 or so of 2024, but at the rate we're going, you'll be on the podcast in another year too, so we'll have plenty of time for an update at that point. I've got to ask, is there anything you wish I would have asked you that did not come up in today's conversation?

No, I think that's it. I guess the other big change since you and I spoke last was we were a little bit cryptic about some of the deployments that we had with our customers. But now that the cat's out of the bag, FedEx announced some of the work that we're doing together on our truck loader. Together, Sagawa, which is one of the largest parcel companies in Japan, announced that they're doing something very similar with our truck loading program. UPS announced their work together with us for small package induction and sortation, and VF Corp, which makes Supreme and The North Face, is using our palletizing and depalletizing system. That's kind of just the tip of the iceberg, though, with these deployments with partners and collaborators in the industry.

The other big change is, so we've been focusing on deploying with these large customers and working directly with them, but as you were talking about democratizing the technology, one of the challenges that we've had is making sure that we as Dexterity can give this white-glove, totally immersive, soup-to-nuts experience of getting set up and deployed at a customer site. In the last year, a big focus of ours has been how we can enable an ecosystem of partners, traditional system integrators, companies like Dematic, DCS, JR Automation, MH, who already have the experience working together with companies in the space, and they can get up and running with a Dexterity system on a customer site without necessarily having Dexterity need to be there shoulder and arm with them together. What's really powerful about that is that it expands the number and profile of the companies that we can work together with and help enable with our technology, and that's really exciting.

What I like about that is, I mean, partnerships come up a lot, whether we're talking about robotics, whether we're talking about building your smart manufacturing ecosystem, everyone that's a part of it. I love hearing it from the startup perspective of how we talked about scaling throughout this conversation, right? This is going to be one of the next things that allows you to scale, by having those players in the warehousing space, you just rattled off a bunch of them, that are able to start deploying solutions on your behalf, if I'm saying this the right way. That's a super cool thing to see play out. I know when I go back to listen to this episode for our follow-up, whenever that may be, we're going to be talking about that, so awesome. Hey, I appreciate you taking the time. It's good to hang out here in the Bay with you. Thanks for doing this right after making the trip. Hopefully you get some shut-eye before our party tomorrow.

Absolutely. Cool, Chris, thanks for having me again, and I'll see you at the fourth time. Cheers.

Cheers.