Listen to the full podcast here.
Steve Crow:
Welcome to episode 15 of the RobotReport podcast, which brings conversations with robotics integrators straight to you. I'm Steve Crow, editor of the RobotReport, thanks for being here. I’m joined by Senior Editor, Eugene Dimitri. Geno, how was the long weekend?
Eugene Dimitri:
It was great! The weather was really nice here in Boston and it's nice to have a long weekend, even though it was still at home. I don't know that we have had normal vacations, but I do look forward to them.
Steve Crow:
Yeah. We actually camped out in our backyard a couple of nights a week, just to get some sort of normalcy back. We typically take a camping trip every summer, so try to get some normalcy fit in. So that was pretty fun. New episodes of the RobotReport podcast drops every Wednesday. You can find us on Apple podcasts, Google Play, SoundCloud, Spotify, YouTube, and anywhere you get your podcasts. Please subscribe, leave us a rating and give us a review. Gene, another packed show today, we actually just published your recap of August robotics investments. So we'll get into that in a bit. We'll also hear from Brady Watkins, the senior VP and GM at SoftBank Robotics America. But first, our conversation with Daniel Theobald the founder and CEO of Vecna Robotics and a founder of MassRobotics as well.
Eugene Dimitri:
Daniel has been a great friend of ours for many years now. And autonomous mobile robots are hot, right? All the companies here in Massachusetts getting $20 million, $40 million or whether they're getting these big implementations. And honestly the COVID-19 pandemic has increased interest and AMRs. And so when you talked with Daniel about trends, we talked with him about how intelligence and AI are actually making these AMRs more useful, not just for navigating around, but for coordinating and working with warehouse systems. And then we look ahead a little bit into latter half of 2020.
Steve Crow:
Here's our conversation with Daniel, enjoy!
Steve Crow:
We'll get to the robotics stuff in a minute, but I know you have a ton of interests and hobbies that are outside of robotics. You told Gene and I a story once about the solar powered mini bus that you built. I think the last time I actually talked to you, you were out on your family farm out in California. Tell me about the farm. I'm so interested in this. How did you become interested in farming? Have you made any robots to help you out there? Just tell us about that.
Daniel Theobald:
Yeah, farming and growing food has been a passion of mine my entire life and just a really amazing opportunity to help solve some of the problems in agriculture and farming. It's a weekend warrior type of activity right now, and the fun things that we're doing are numerous. I've been doing a lot of drone work. We've got a drone that's 3D mapping our entire 127-acre farm every week so that we can see how things change over the years. We have a solar-powered robot that’s navigating our fields.
The farm is called Twisted Fields because we are taking advantage of the topology of the terrain and having the rows not run straight as they normally would on flat land, but follow the the curves of the hill. And so we needed to build a robot that was able to handle that type of environment. A lot of the robots people are working on in agriculture right now don't have a lot of capabilities to deal with anything other than flat terrain.
So anyway, working on lots of things, the whole thesis there is to be able to grow a wide variety of healthy, organic crops locally. And the only way to do that is through technology, because labor is a massive problem in agriculture and it's a real limiting factor. And that's why you see these massive monocultures. Monocultures is where you have one farmer with one tractor that's farming the same crop for thousands of acres. And while that provides some some advantages from labor and cost and efficiency point of view all the other things are pretty much bad about it. It's bad for the environment.
It's not great for the crops, crops grow better in a biodiverse environment. And then you're also shipping these crops significant distances, and so we tend to create crops that travel well and keep well on the shelves rather than really maximizing their health and and flavor, the things that we should really care about with food. So that's a brief summary. It's pretty fun. My main focus of course is my day job and we're doing great things, but it's nice to have something else to focus on every now and then.
Steve Crow:
A future market for Vecna Robotics right there? Agriculture?
Daniel Theobald:
Oh, we'll see.
Eugene Dimitri:
So getting back to the robotics sort of core mission, obviously we know that supply chain automation is really hot right now with e-commerce demands spiking around the coronavirus pandemic and mobile robot companies have received a lot of investment lately. But a lot of challenges are sometimes overlooked. What are some of the biggest reasons why warehouse automation fails?
Daniel Theobald:
Yeah. I think, Gene, probably unrealistic expectations and overhyped selling continues to be a problem. It certainly is a problem in any new industry. We all have seen that typical curve. There's a lot of hype and then, there's the valley of disillusion and all that, but that occurs because I think people go in with maybe a Hollywood level expectation for technology, and don't realize that these are really hard problems and they just take work.
You have to get in, you have to start now and you have to learn. And there's going to be problems. There's going to be challenges. And if you don't go in with the right approach of getting started, learning fast and making sure that you are really committed to the automation of technology path to make it work. And then a lot of them just fail because people don't have that level of commitment and they don't really get to the ROI level. They don't really achieve the results that they're looking for. So I think that's a big part of it. I think we need to be careful to not set unrealistic expectations in the robotics community. And a lot of that has been around this idea that robots can do everything, and that robots should do everything without humans around and robots should replace humans.
Elon Musk made a pretty bold statement when he said, this was after the whole giga factory project running into problems. And, he said "I made a mistake. It's my fault I tried to over automate. Humans are vastly underrated." And I've seen a real trend now in the industry of robotics companies embracing that idea and realizing that humans are critical. Humans are really the point, right? Humans are why do we do everything, and when people take an approach of "let's just try and get rid of all the humans" in a process it usually doesn't work as well as you might hope.
Eugene Dimitri:
And we've talked before a lot about the idea of augmenting human capabilities with autonomous mobile robots, but given all the interest and maybe greater awareness of AMRs right now, do you find that you still need to do a lot of educating of customers and integrators?
Daniel Theobald:
Yeah, I think we're still early on. So the market has certainly turned from a push to more of a pull, meaning that customers realize that they need robots. They need automation if they're going to compete in this rapidly evolving global economy. So I think that has been a change. In terms of people understanding how to effectively apply robotics and making sure that they're introducing robotics in their operation in a way that is a positive for their staff. Something that's often overlooked is an area where we're still making a lot of progress.
Steve Crow:
We're talking with Daniel Theobald, the founder and CEO of Vecna Robotics. Daniel, you've called 2020 the year of orchestration. Tell us what you mean by that.
Daniel Theobald:
Yeah, really I'm talking about the decade of orchestration. I think last decade, there was a tremendous amount of focus on robots, the physical hardware and excitement around that. The interesting thing, and we learned this very early on, we've been building robots for a couple of decades now. Robots are just as good at wasting time as human workers. And robots are just a tool. Again, I think we sometimes ascribe human or superhuman capabilities to robots to just be able to do things by themselves and figure everything out.
Robots need supervisors, robots need effective tasking, and so the idea of making sure that you've got the right resource in the right place at the right time, and to be able to figure that out in real time, that's really the key to success - robots involved or not. And we've seen an incredible increase in interest around workforce automation, orchestration, to not just make sure that the robots are tasked effectively, but to make sure that the human-robot interactions, the robot-to-robot interactions, and oftentimes just the purely human worker tasks are being staffed appropriately. People don't realize a lot of times that the typical warehouse worker, or the typical piece of warehouse equipment is un-utilized or significantly underutilized from 30% to 50% of the time. That's factor number one. That really makes a big difference in realizing how much latent capacity there is too.
Factor number two, is that bottlenecks occur in any system you always have at one point in time, a single bottleneck that is limiting your throughput. And if you can figure out where that bottleneck is in real time and address it, then you can actually increase the throughput of your facility. Many times, what we see is people will have initiatives, to have this group keep this belt clear because, last week, this conveyor belt was a bottleneck. And so then this week, they throw a whole bunch of people at that conveyor belt. And they're super proud because they kept that conveyor belt cleared. But what they don't realize is that downstream from that conveyor belt, now they've just overwhelmed a different part of the system. And you've got you've got a bunch of backups in your buffer. So it's really about keeping everything in balance.
Again, right resource right place, right time. And you have to be able to adjust those plans in real time, because there's no situation where everything goes perfectly according to plan in any of these facilities or warehouses or order picking operations. You have to be able to understand what's going on now in the moment, and continually adjust. Send people, send robots, send other resources where they need to be. So that's really what we mean when we talk about orchestration. It's like the conductor up there, making sure that all the symphonies is doing their part exactly at the right time, and you end up with a beautiful result rather than cacophony.
Steve Crow:
Vecna has made a couple of announcements, maybe within the last couple of years, around software systems and sort of artificial intelligence to create new efficiencies. How are those technologies helping your AMRs?
Daniel Theobald:
Yeah, it's significant. So in one research study we did, what we showed was that when the robots were just following a set schedule, how many of them are deployed. "Milk run" they might call it, or just moving things from point A to point B on a schedule. There were a lot of benefits. Injuries went down, the damage to the facilities went down, reliability went up, lots of good benefits, good ROI there. But this is when we realized that the robots were still not achieving their full potential. And so when we allowed our pivot orchestration software to now dynamically direct the robots to what they should be doing next, not based on only local information, and this is the key, but based on facility-wide information, they were able to increase throughput by 38% in this particular case. Your results may vary obviously. But the really exciting thing, I mean 38% is a big increase in throughput without adding any additional resources - just by better scheduling.
But what we found really incredible was when we now allowed our pivot orchestration software to also direct, basically give suggestions, to a shift supervisor around where the staff could be best allocated at any one moment in time. And we achieved 116% better throughput without adding any resources. So again, it's just this idea that right resource, right place, right time makes a huge difference. And sometimes people don't need robots. We had a customer come to us that was all excited about buying robots. And as we looked at their operation, we tracked what was going on. We did some simulation. What we showed them was they didn't need that physical device called a robot. What they really needed was better better coordination of their work and just by doing so it would allow them to achieve close to two times improvement in throughput without adding any additional resources.
Eugene Dimitri:
Well, Daniel, those results are actually really impressive. And I think that gets back to the point about robots plus humans are in some cases, when it's applied correctly, better than either by themselves. But we've spoken before about importance of standards for both robotics developers and for end users. What efforts are you working on right now, and why should the listeners care about that?
Daniel Theobald:
Yeah, this is a really important area, Gene, thank you for asking. This industry in a way is still very nascent to embryonic, that's probably the most correct characterization. And if you look back at some other industries, the computer industry, cell phones, et cetera, what you realize is those industries did not really start to achieve any kind of significant impact in the world, until there were interoperability standards. I'm dating myself a little bit, but back in the day, it used to be that a fortune 50 company would hire IBM to come in and build a computer for them. They would come to the facility, they built this computer and then they'd send an army of software engineers into program specifically for that customer.
One day, someone had this absolutely crazy idea, and it's hard for us to realize how crazy this idea was at the time, but this crazy idea to allow a different company to write software for an IBM computer. Who would ever do that, right? That's crazy. Because then IBM couldn't control quality, and you can think of all of the reasons why that'd be a really bad idea. Of course, that's exactly what happened and it changed the world. And we're still in this place with robotics where every system is really largely different. They don't plug and play like you would like them to. You can't just buy one and bring it home and plug it in.
There's a lot of work yet to do to get to the point where it's easy for a customer to buy a robot. They can buy a robot from a number of different vendors. Those robots will operate together. They will use the same infrastructure. And it just makes it easy for robots to actually make an impact in the world. So realizing this several years ago, Vecna Robotics has a community service program where we encourage our employees to spend 10% of their work week doing community service. Myself and a number of my colleagues here really felt like the industry needed an organization to help move robotics forward, to help try and create some of these missing pieces, and interoperability standards is a big one. So MassRobotics is now working on the MassRobotics Interoperability Standard. We have a broad coalition of robotics companies and end users involved in that process. And again, the goal here really is to take that first baby step in interoperability between robotic systems so that we can really make it easier for customers to adopt and see the benefits of this great technology.
Eugene Dimitri:
And we think that's a great effort. We've definitely talked with a lot of people about the need for greater interoperability, because we foresee in the not-too-distant future environments, and we're already starting to see this in warehouses, but lots of other environments where you have multiple robots, multiple vendors, all conducting different tasks. And ideally, as you were saying, orchestrated by some combination of humans and AI.
Daniel Theobald:
Yeah. And I would strongly encourage organizations out there that are considering investing in robotics that are trying to automate to make sure that the vendors you choose are committed to interoperability, are committed to open standards. That that's really, what's gonna make the difference in allowing your investment to really flourish.
Eugene Dimitri:
And you mentioned MassRobotics and we're certainly familiar with the organization. But how, in terms of initiatives, obviously trying to nurture the startup scene here in New England and Vecna is a neighbor of mine, but in terms of what you see going forward, this pandemic has obviously affected all industries, but wondering what sort of initiatives that you are looking forward to as with MassRobotics.
Daniel Theobald:
Well, I think it's so great that MassRobotics has provided resources for robotics, startups and IOT and AI. One of my big motivations was that starting a company is hard and starting a robotics company is probably one of the hardest. There's a lot of unique needs that robotics companies have, in terms of hardware, machine shop, sensors, sensor expertise. Robotics is in some sense, the culmination of all human engineering we take advantage of electrical mechanical software. We bring all of these things together in complex systems that need to operate reliably 24/7 in the real world. I mean, this is in many ways, the most challenging engineering that the human race has ever done. And that's a team sport. We need to help each other as an industry. And MassRobotics has just done an absolutely phenomenal job, and I'm so, so proud of the team there in trying to help reduce the challenges for new innovative robotic companies and not just in the New England area, but really around the world. They make lab space available, they provide expertise, they provide access to equipment. They bring in investors and big companies to help make those really important business connections.
A big part of it too, is helping the innovative, oftentimes just out of school, talented roboticists, understand what are the problems that are actually worth solving. We saw a whole spade over awhile of companies that had really cute robots, really technologically challenging ideas, but the economics just weren't quite there. They weren't really solving the problem that the customers needed solved, and that step when you have a business and you need to make sure that the customer's needs are being solved, and you can't do that if you don't understand who your customer is and what their real issues are. So MassRobotics' ability to bring these customers in, and connect them to the robotics startup, I think is probably one of the biggest value adds that they brought to the industry.
Steve Crow:
We're talking to Daniel, Theobald the founder and CEO of Vecna robotics. He's also a co-founder of MassRobotics, and Daniel just sticking with the startup theme here for a minute. I go back to, I think it was last October, which is a lifetime ago now, you have your new headquarters in Waltham. You had a great ask me anything panel, give a lot of startup tips, advice for startups. It was a great panel of other Boston robotics experts, just for the startups who are listening. What's some advice that you can give them, maybe some of the challenges that you've experienced in your career?
Daniel Theobald:
Advice number one is really don't reinvent the wheel. I can't tell you how many times I've seen talented young people try and do everything. They want to build the new, the best motor controller. They want to build the best vision system. They want to build the best you name it. So you need to stand on the shoulders of people who have gone before you if you're really going to have an impact. Figure out where your unique contribution is, and focus on that. Don't get distracted by the plethora of things that you could do. There is some narrow focus that is going to allow you to really make an impact, so I'd say that's number one.
Number two, you have to be customer driven. If you are not working very closely with the number of customers, stop everything you're doing, that is your priority. Until you have that, you're not working on the right problem. I guess number three is get out in the real world. And if you're not trying the technology in the real world, you're probably focused on solving the wrong problems.
No one is smart enough to predict where the issues are going to be. There are going to be issues. There are going to be problems. But these systems are so complex and the world is so complex that until you get the technology out there and start learning you're not going to be making much real progress. And that goes to the big theme here at Vecna Robotics, which is, our goal is to learn faster than anyone else, and the way you learn as you try stuff and you make mistakes and you fail and you try again and you improve. And getting on that to very rapid learning cycle is something that has allowed us to accomplish a tremendous amount in a relatively short period of time.
Steve Crow:
How else does Vecna plan to differentiate itself going forward? I mean, obviously supply chain, logistics, and automation is the market right now, and for the foreseeable future, very competitive, and in our own backyard in Massachusetts. There's a lot of those companies working in that space. So, how do you make sure that Vecna Robotics stays innovative?
Daniel Theobald:
As companies go through their life cycle, there's various stages and over those different stages, the focus changes. Early on, we had a tremendous innovation engine, I would say that it was a cranking, and that allowed us to really attract and meet the needs of a wide range of customers early on. And then you go to the scaling phase, and this was the feedback we got from a number of our customers a couple of years ago, which is, "wow, we love your technology. Your stuff works. It's great. Our only concern is, are you going to be able to scale with us? Are you going to be able to provide the number of robots we need? Are you going to be able to integrate with our warehouse management systems, and pass our security audits, and pass our safety reviews?”
There's a lot of I dotting and T crossing that goes on. And so things swing that way a little bit, but recently Vecna Robotics reinitiated our advanced development team. And this was because we realized that dealing with all of this really important stuff that we needed to do to scale was getting a little bit in the way of us staying at the leading edge and keeping our innovation engine really humming. And so we have created a new division in the company as of last year called Advanced Development. And we're doing some really exciting work, and we're engaging directly with a number of customers around new platforms, new projects, new capabilities, as well as of course continually improving the robots we have.
One of the things that I think does make us really unique in the market is that our product is constantly improving, getting better, and learning. We've got this approach of making sure that the robots are feeding data back to us and to the customer, providing us insights into the customer's operations, which helps us help the customers understand how they can increase their throughput and efficiency and lower costs in their operations, but also allows us to continually tune and improve the robotic systems. We had a deployment to the large retailer where, day one, the robots the system was exceeding the customer's ROI metrics that they had set for success. A few weeks into it just through the natural course of our learning process with both machine learning and our system learning here at Vecna Robotics, the system was performing 17% better throughput than it was before. And this process just continues and continues. Typically robotic systems, AGV systems, they work pretty well when they're first installed, but then, as the environment changes, as the workflow changes over time, the systems may be grade and you have to go in and sort of reconfigure or recalibrate things. Our system is much more continuously learning and if something doesn't go well yesterday, it's probably gonna go better today and it will definitely go better tomorrow.
Eugene Dimitri:
Daniel, Vecna Robotics won a 2020 RBR 50 Innovation Award for the fifth year in a row. So congratulations for that!
Daniel Theobald:
Thank you so much. I really want to just express gratitude to the whole team here at Vecna Robotics. We have an absolutely amazing team. It's a great place to work. One of the biggest pieces of feedback I get from people is how enjoyable it is to interact with the people here at Vecna Robotics. We really work hard to create a workplace that is productive, safe and a lot of fun. And I've gotten that feedback that we're doing a pretty good job of it. Always room for improvement, of course, but we continue to make that a focus and I think both our customers and our employees really appreciate it.
Eugene Dimitri:
So Daniel, we recently covered the announcement that Daniella Rose, who is a noted scholar at MIT CSAIL has joined the robotics board. Can you tell us a little bit about that? And also the women in robotics initiatives?
Daniel Theobald:
Thank you. Daniella is a great addition to the team, we're really excited to be working with her. She is not only a world-leading roboticist, but also is really passionate about the idea of creating work quality and engineering and technology and robotics. We're working with her on a number of initiatives and I'd like people to keep an eye out for upcoming events with women in robotics. It's a project that we're doing here now. We've had one event already, which went really well. And it's really just to raise awareness of the robotics industry as a great career choice for women engineers, and to encourage young women to pursue stem education and get excited about robotics. We are underrepresented right now in this area and are working hard to try and fix that. But really excited about Daniella joining us and helping us in this important initiative.
Eugene Dimitri:
Yeah. Diversity in technology in general, but especially in robotics, as well as those areas. I know you mentioned before Vecna's devotion to social good, and what better way to lead by example than to do this?
Daniel Theobald:
Yeah, it's important that we make sure that we're really building the kind of world that we want to leave to our children. That's something that we really are all about here. Empowering humanity through technology, and I think oftentimes people can forget that that includes everybody. I'm really proud of our team here for taking the initiative, creating that organization and working hard to help make this career not only accessible, but an attractive option for a wider range of people.
Steve Crow:
Daniel Theobald, founder and CEO of Vecna robotics and a co-founder of MassRobotics. Thanks for joining us, whether it's a family farm, at Vecna's headquarters or MassRobotics, Gene and I hope to see you soon. It's always great talking shop with you.
Daniel Theobald:
Thank you. Appreciate the time.
Steve Crow:
Gene, when all this COVID stuff ends, which will be soon, we to get out to this farm and check it out. Sounds pretty cool!
Eugene Dimitri:
I think it would be a lot of fun.