The best kinds of technology are often those created to make life better for everyone. This is what drives Daniel Theobald, harnessing the power of technology to revolutionize the world. He is a thought leader in the space of robotics, founding companies Vecna Robotics and MassRobotics. In this episode, Daniel sits down with Doctor Awesome to discuss the evolution of the field and what the future holds for us as it continues to grow. From humanoid robots to the recent surge in investor interest, Daniel shares his optimism about technology’s potential to solve global issues. One key area he highlights is agriculture automation, where he talks about his passion project at Twisted Fields and its open-source farming robot, Acorn. Daniel then further dives deep into the challenges in agriculture, enhancing human abilities through technology, and the importance of optimizing processes. Whether it is in agriculture or self-driving cars, robotics has undeniably shaped the world’s transformations. Let Daniel’s insights show you its significant impact not just today but also in the future!
—
Watch the episode here
Listen to the podcast here
The Future Of Robots – A Conversation With Daniel Theobald
We have a very special episode, another one of our in-person shows, which I always appreciate. We have a real thought leader in the robotics space here. Daniel Theobald is the Founder of VectorRobotics and the Cofounder and President of MassRobotics. The whole idea is that he’s doing a lot of cool things with robotics. Daniel, tell us a little bit about your background and what you’re doing.
I was very fortunate to have the opportunity to be born in a time and place where I was exposed to these revolutionary technologies at a very young age. I grew up in San Jose, California. It was before it became Silicon Valley. I remember when I first heard the term Silicon Valley. It was not until I was in high school at that point that I was very fortunate because I went to one of the elementary schools that had a computer lab.
It was the sixth elementary school in the world that had a computer lab. We had the Commodore PET and Apple-1 computers there. I was fascinated with this technology and what it could do. That started me on this path. The biggest theme of all of my fascination with technology has been, “How can we use this to make life better for everyone on this planet?” I had lots of opportunities. I ended up at MIT and was fascinated with the idea of robotics because I’ve been doing mechanical engineering. I’ve been building robots since I was a kid.
I was a little bit disappointed that MIT didn’t have a robotics program. That’s why I made my own. I might take classes in Course 6. Everything’s a number there. Course 6 was electrical Engineering and Computer Science. NASA took classes in Mechanical Engineering. I ultimately got a mechanical engineering degree. We are barely scratching the surface, even the technology that exists nowadays. Let’s say, for the sake of argument, technology progress stopped. With the technology up nowadays, we could do so much to revolutionize the world. It’s exciting. It’s filled with possibilities. Our only limitation is our imaginations.
I’m reminded of that old quote, “The future is now, but it’s not available to everyone.” I feel like it hasn’t trickled down how I would have hoped to and certainly not at the speed I would have hoped. I went to engineering school and Robotics was something that was available in a very rudimentary version at that time, but certainly not what you would see in science fiction, which is a great inspiration for me. I’m waiting for that day when I can have a robot butler. How far away are we from that?
It’s an interesting question because we all suffer from what I call the curse of ever-increasing expectations. If you go back a couple of hundred years, you had somebody who spent all their time washing your clothes and cooking food. That was their job. You had to wash every piece of clothing by hand. We have washing machines now. You could argue that we’ve got all kinds of great robot butlers already, but we take them for granted.
We all suffer from the curse of ever-increasing expectations.
The curse of ever-increasing expectation is like, “Now I want somebody to put the clothes in the washing machine for me, take them out, put them into the dryer, and then fold them and put them in my drawers,” because now we consider that the terrible chore of doing laundry. It will happen. The fascination with humanoid robots over the past few years is when it’s ramped up. All these things are cyclical. Remember, we get excited about something, a ton of work goes into it, and people get a little disappointed that it doesn’t accomplish everything that they thought it could. A self-driving car is a great example of that. People lose interest for a while until we get like the next school generation.
As far as I can tell, these things go in 4 to 8-year cycles. Somebody new comes along, like, “We could do all these amazing things with this technology,” and then it sparks a whole new fascination. The key thing to remember is that we need technology to make technology. The more technology we have, the faster we make new technologies. Despite your very real and valid point that adoption isn’t as quick as we might like, it is happening on an incredibly fast timescale compared to anything else.
The key thing to remember is that we need technology to make technology, and the more technology we have, the faster we make new technologies.
I did this TEDx Talk years ago. As part of that talk, I was talking about the power of technology. It’s this concept. I took the time to graph out the history of technology and the human race and show the amount of technology that existed. If ever there was a hockey stick or an exponential curve, this was it because the technology was essentially flat for millions of years, wherever you want to say the start of humanity was. You have this one thing, and that makes the next 2 or 4 things faster.
It goes up quickly. We’re at this point now where, with AI, we’re able to process information. I hate using the term AI because it’s a real misnomer. It confuses and alarms people unnecessarily. We’ve got these processing tools now that are able to handle such large amounts of data and help us accomplish things in a fraction of the time it used to take. I remember when I sent the AI Lab at MIT, there was always this debate like, “Do we even start this algorithm now or does it make more sense to wait two years when the next generation processor is out.” Starting it then, we are like, “Make-up for all this time that we could have been putting.” It’s this interesting thing. We will have robot butlers. There will be humanoid robots walking around doing practical things. What will those things be? We’ll see.
I feel like there’s a real schism in people interested in adopting humanoid robots versus other types of robots. I know that one of the companies that you’re doing is doing agricultural robots, which is like the second CAMP. How do you feel about the two of them? Where do you see the role of humanoid robots versus not humanoid robots?
The most recent humanoid robot fascination started or was the seat of it, which was the DARPA Grand Challenge. The Grand Challenge was the cars, which we kicked off that one. That time was very interestingly similar because there’s the DARPA Grand Challenge. That took a few years, and then suddenly, you had twenty self-driving car companies. Many of them have come and gone. There are a very small number of states, but it’s the same thing with the humanoid because four years after that was the DARPA Grand Humanoid Challenge or whatever it was called. The reason that they were doing that was because of the observation the as-built world was built for humans.
It’s about, “This scale was built for two legs. This was built for two arms.” There are ladders, tools and all this stuff. If we could build robots, they could operate in an as-built human world. How amazing would that be? It’s a perfectly valid thought and a reasonable thing to do. As we build these systems, we find that the human machine is an amazing machine that we vastly underestimate. I used to tell people, “Until you’ve tried to build a robot that does something that a human can do, you probably don’t have anywhere close to A full appreciation of how amazing humans are.”
I was doing a little analogy to somebody once about vision. Vision is amazing. What most people don’t realize is because of our experiences, we “see” not the world as it is. We see a model of the world that our brain creates in our head after lots of processing. I remember in one of my Machine Vision classes at MIT when this started to become clear to me, when Professor Horn was talking about discovering that they are multiple layers of neural processing in the vision pathways before even getting to the main brain or something like that. We’ve got edge and blob detection, different convolutions are called, and all these different processing steps happen.
Your brain essentially then takes all of these different inputs. It takes the same set of pixels that are colors, “This dot has this color. The dot next to it has that color. That next does that color.” Our brain then processes it in all these different ways and then uses it to build a very rich 3D model of the world. Oftentimes, that 3D model that our brain builds is wrong because sometimes you can misinterpret the visual data. Long story short, we see a robot that has a camera on it and we think it’s seeing. It’s not seeing, nowhere even close, not even remotely close. We’re working on this humanoid robot project with the US military. It was called the barrel robot. It was an experimental robot meant to help rescue a soldier who had been injured on the battlefield and was too dangerous for another soldier.
I’m meeting with the government representative who sponsored this project. He said, “I’d like it if you have a robot that is able to pick up that cup of coffee.” I said, “That’s a very challenging research project.” Keep in mind, this was many years ago. It’s a challenging research project. IT’S out of the scope of this. He says, “Why? It’s got a camera so it can see. Tell it to pick up the cup of coffee.” It’s understandable that he would say that.
It belies a lack of understanding of the complexity of vision.
I don’t know if you’ve ever had this experience, but I have had a number of times in my life where I experienced a visual ambiguity where your brain can interpret the scene in different ways. There was one time I was sitting at a bus stop waiting for a bus to come. There are all kinds of details that I won’t bore you with all of them, but the road has been paved. They’re no lines on it. There are all these perfect visual setups for what happened. I was staring down the road thinking and I saw this truck coming.
The trucks were coming. It’s getting closer and closer, and then I start to get this weird feeling in my stomach. I start to get a little bit sick. My vision started getting blurry, and then suddenly, my brain’s interpretation of the scene changed from a big truck far away to a small truck close to me because it looked like a bigger truck, but it was a smaller truck. The way it was designed.
Everybody thinks we have stereo vision. We can tell depth and distance. Stereovision for my two eyes may give me some extra information when I’m doing fine manipulation, but for the most part, our vision is monocular. It’s almost more for redundancy than anything. Our brain interprets information to create this 3D model with cues and hints. It says, “I know that a person is roughly this tall. If I can see a person in the scene, I can then use that to try to scale this other thing in the scene.” It’s this incredibly complex thing.
That understanding varies between different people. For example, baseball pitchers and batters have developed this understanding of space and time in such a way that it’s very easy for both of them to understand that, but you throw in something different like a softball pitch and then all of a sudden, they’re striking out completely. Their algorithm changes from one to the next and I agree that the complexity of vision for me is the biggest obstacle to self-driving and robots. How far away are we from that being fixed?
When we ask ourselves these questions, a big part of the problem is defining the finish line. Self-driving sometimes works, but there are a billion edge cases where it’s not good enough. This is where my approach to robotics has been, “Find the right applications to start with.” It’s interesting because I can’t tell you the number of robotics companies that all started out building a humanoid robot.
I’m Vecna. I could probably name 4 or 5 other robotics companies that you go to their headquarters and there’s their first humanoid robot sitting in the corner was a cool project, but it’s not practical. It’s very expensive, complicated and not going to be reliable, for something to be reliable and cost-effective to have a return on investment, which is important at least in many applications. It’s about solving the problem in the most simple way possible.
The humanoid robot research is great. It’s awesome and amazing to move that forward. Eventually, if the interest remains long enough to continue finding this, you will start to find those applications where that technology makes sense. There is a legitimate advantage to having that humanoid form do this particular job that no other technology is going to be able to do as effectively for the same price.
It’s going to take time to find that. That’s what I would call the technology push side of the equation and the application poll, like we have this strong need to move pallets. Let’s say, for example, to give an example close to home for me. We could go out and build a huge humanoid robot with big arms and can pick up pallets and move them around, or we could use a forklift and automate it.
There’s always this play and back and forth. This is the fun of technology. This is the great opportunity that we have as a prosperous society. We’ve got the resources, the time, the interest, and all the pieces that can allow us to explore and play. That’s what’s going to ultimately give us the tools we need to solve some of humanity’s biggest problems.
As someone who’s in this space, you’re seeing the trends happening. Do you feel like more people are leading towards that automation of existing technology like forklifts now or are they reinventing the wheel and something else that was a robot that addresses that pallet situation because what I see is a lot more automation these days. I feel like that’s what I’m noticing. Is that accurate?
We are in a big robotics wave of enthusiasm. It’s interesting again to talk about how these waves occur because robotics was a bad word many years ago. Investors would not invest in a robotics company. There was a mobile video conferencing company called Vigo. Tom Ryden started it. He’s now the executive director of MassRobotics. He is an amazing guy. He talked to me about how, at the time, they couldn’t get anyone to be excited about funding this idea of a telepresence robot.
They called it a mobile video conferencing device because they didn’t want to have the word robot in there. Now, you don’t have the word robot and your new startup. You’re not as interesting. It’s like back in the dotcom day when if you put dotcom on the end of your company name, suddenly your value was 10X. It’s silly, but it’s human nature. It’s the way it works. Automation is good enough now.
If you’re a science lab government academia, I think we ask the hard questions, we try and do the stuff from the movies. It’s the right thing to do. If your goal, on the other hand, is to build a profitable business, then you need to take a different approach. You take a more practical approach. The amount of automation that will exist in that is going to be much higher than it was a decade ago because you can reliably do things that weren’t possible before.
How’s it going in the agriculture space? I feel like that’s something that has vastly changed over the past or however many decades we’ve been doing it. I grew up in Florida. It’s very agriculturally heavy. What I noticed growing up was that it was very much a labor-intensive process where you hired a cheap skilled lab. That was what got you to the finish line. Nowadays, large industrial engines are able to do the same things. What is the difference in the agriculture space?
Agriculture is an interesting environment because it’s still very non-uniform. Most farms have significant differences from other farms. There hasn’t been a real standardization yet. There’s an important distinction between large farming equipment and labor. Most of what we would call specialty crops are still done with manual labor.
Raw vegetables, strawberries, fruits and a lot of that. We’re now trying to figure out how to do those things with automation, but it used to be that back in the 1920s or so, 73% of the population worked in agriculture. It took 73% of the population to produce the food that we needed to eat to sustain a population. The latest number is like 2% or something like that. A lot of that was the invention of the tractor.
What that means is that we figured out how to grow wheat, corn, and soybeans on a massive scale, and we’ve made them incredibly inexpensive. It turns out that eating wheat, corn, and soybeans in those quantities is not particularly healthy for us, and it’s caused a tremendous number of health problems. You could probably make a very strong argument to say diabetes IS the direct result of figuring out how to grow thousands of acres of corn very cheaply.
My background is in Craniofacial Surgery. Craniofacial growth has been affected. Our jaws and faces were entirely different pre-agriculture. Nowadays, a lot of biological anthropologists will credit the Industrial Age causing our jobs to shrink. It causes a whole number of different maladies, diabetes being one of them, but because we don’t have to chew as much exactly.
That’s interesting in our culture. There are a lot of cultures that still prefer food that has some bite to it. They want to have the bones in there. They want to work at it.
There are still non-agrarian societies that have a full complement of tea. They have no need for braces or any of these other lifestyle diseases. The point is that I agree with you. I think that industrial agriculture has been this whole other industry with pros and cons, pluses and minuses, but I feel it has been very technical. When I think of the biggest pieces of technology, I think of construction and k industrial agriculture. You have these giant things that make or grow other things.
It’s important to keep in mind that there are pros and cons to all of these. It used to be that 70% of us worked in it. Now, it’s 2%. Food is cheap. We’re not eating as healthy. That’s a con. The pro is we’ve now got an extra 71% of the population that can focus on other things like creating this technology. That should help us solve these very problems. It twisted fields. We’re building an open-source farming robot called Acorn. The whole idea there is, “Can we now use this much smarter, much more capable automation because the automation you see on large tractors is decades old at the end of the day?”
This is much smarter automation to try and solve some of the labor problems by growing good, healthy food locally without pesticides and too much fertilizer that run off into our waterways and destroy our environment, without spraying things that are killing our honey bees. The idea is that we should start using technology to empower small local organic farmers, for example, to be economically viable.
It always comes down to this at the end of the day. Can we grow healthy food that tastes good and is affordable for everybody? A large segment of the population can’t afford to go to Whole Foods and buy groceries there. Organic food is priced out of most people’s budgets. It’s a real issue that we need to figure out.
What is the open-source robot look like? Is it more of an automated tractor type of setup, or is it more humanoid?
In this case, it looks more like a Mars Rover. I worked on the Mars Rover program when I worked at MIT. This is a VW Bug-sized Mars Rover-looking platform. We’ve done some unique things. Our goal is not to solve every agriculture problem with it. It is to provide a platform that other people can use to solve every agriculture problem because, as I was talking about before, farming has been done differently. Our goals are to keep it incredibly low-cost, repairable, easy to maintain and then easy to build new apps for.
It’s a lovely shape, but we’re trying to basically provide an iPhone-style platform where the app builder doesn’t have to worry about navigation, safety, power management, communication and all of these things. They can say, “All that’s built already. Now I can take that and program it to pick up the food from the grain bin and deliver it to the chicken coop without having to build my own.”
Think about cell phones. If that teenager in his bedroom had to build the cell infrastructure, the actual cell device, the user interfaces and the battery, it’d be impossible, but we’ve now created this ability for people to build on top of existing technology. It opens up all kinds of new possibilities. That’s what we’re trying to do.
I like the idea of taking mundane tasks and making them automated. That’s something that I think is going to provide a lot of benefit for a lot of people because the admin requirements for modern living, I feel, is getting out of control. I don’t think that’s something that is isolated to me. I don’t think that that’s my gripe on modern society. Lots of people have that feeling. I have a money tree in my house. Even watering it every day, maybe I forget when I’m going to go on vacation. Even that, I want to be automated. I know that there are systems out there. I want these general-purpose robots to be able to accommodate that stuff. I like the idea of an app being able to set that up.
There’s no fundamental reason why we won’t have an assistant that’s a robot. You can say, “Can you buy me a money plant and keep it alive for me?” We’ll get there. My hope that’s important to me is that as we make more time for us, we can spend that time doing more important things, and this interesting experience where early in the days of Vecna Technologies, we’re real early leaders in a lot of different robotics areas.
This group from Japan came over. It was a combined group from both the Japanese government and Toyota. They came over to meet with us. They’re talking about how there’s this huge problem coming because there was the graying of the workforce. They were going to have far more elderly people than young people available to do the work. They said, “We want to build robots to take care of the elderly.”
It’s partly because of the sparrow about rescue soldiers thing. That’s when it struck me. I thought, “This doesn’t make sense. Why would we build robots to care for people so that other people can work in a factory? Let’s get the robots to do the assembly so that people can take care of people so we can be more human.” We learned this very quickly while working on the bear project. A robot that is trying to interact with the human body is hard.
Getting a grandpa in and out of the bathtub, in and out of bed or carrying them on the toilet. That’s a recipe for a lot of lawyers making a lot of money. That’s when I thought, “We should be focusing on automating the mundane so that we can be more human. We can take better care of each other.” I know that’s very general, but it’s vague. It’s been a guiding principle for a lot of what I’ve been focusing on.
We should be focusing on automating the mundane so that we can be more human.
I’m appreciative of that fact because I do feel like that’s something that I look forward to with a lot of the emerging technologies because things like artificial intelligence, all these other things that are coming down the pipeline, I looked at them as taking away more than monotonous aspects. Even in my field, a lot of focus is on taking a lot of the admin work out of being a doctor. You could focus more on the human aspect. I can talk to my patients. I can interact with them. You know what’s going on in their lives because I’m bogged down. It’s too much. I appreciate that philosophy.
I do want to talk with you a little bit about not so much human replacement or automation but human augmentation because I do think that the word robotics in my field is used as an opportunity for some telepresence. I’m using a joystick that makes this other thing go nuts or moves things around. Personally, I feel like that’s the most exciting part of Science Fiction for me. When I think of like mechas and giant robots we can step in that make a stronger and more capable, that’s something that excites me. What’s going on in that space?
A lot is going to happen in that space over the next decade or two. That’s one of the next big frontiers. There have been exoskeleton projects for your sarcos as a great example and trying to do exoskeletons forever. It comes down to the same challenges I was discussing before, like interacting with the human body isn’t very difficult. We think, “I can put on this suit. It’s got actuators. It’s all going to work.” There are many details why it doesn’t.
The list is long and challenging. It’s about this idea of finding those applications. You’ve seen a lot of exoskeleton companies that have focused on injury prevention rather than augmentation, but I loved your example of the surgical robots because this is a great example of human augmentation where the human is still the key. You’re not replacing the human. You’re making the humans better.
This idea that I could remotely operate a surgical instrument and then achieve better precision. I could achieve better safety because I can make sure that I put a no-go zone so that I’m not accidentally going to Nick an artery or something like that. I can make it so that my hand motions are big and translate to smaller motions. All these things create empowerment and superhuman abilities that let us do more than ever before. Let’s be better than ever before. For me, that’s exciting. I’ve always looked at it as technology should empower humans to do more and be better. As soon as we start talking about replacing humans, I get less interested. It turns into more of what people see in a dystopian future.
Technology should empower humans to do more and to be better.
One of the most interesting things for me when it comes to the idea of the human interface is for me for it to feel natural. There’s a focus on haptics and all this other stuff.
I worked in the vision and touch-guided manipulation lab and MIT. It’s a big thing we did. This was decades ago, but you still don’t find some of the stuff that we had on there in common practice. It’s amazing. A young engineer there named Thomas Massie was working in the same lab as I was. Our advisor was Ken Salisbury. They were developing this very smooth cable-driven transmission robotic device. It’s called The Phantom. It was cool. I remember when I first got to the lab, they had the slower button. You could put your finger and a thimble that was connected to the robot. You close your eyes, and then they’ll say, “Feel what’s on the table in front of you, and tell me.”
You use your finger and feel around these objects. It was clear to me that there was a cube, a triangle, or whatever. You open your eyes and there’s nothing there. The way that they were able to do that is to have very smooth transmission or force feedback on the robot. The harder I push, the harder it push back. It could have the right compliance, the right springiness to materials, and even the right texture because if you rubbed your finger over something, you’d get the vibration. These technologies all exist. We haven’t gotten to the point where they’ve been implemented into enough of the day-to-day.
It’s nice to know that because, as someone who’s an end user, I don’t have that experience. The robotic surgery applications for me are not natural. It’s like playing the crane game at those arcades.
They’ve tuned that thing to drop it. It’s not surgical.
It takes a long time to set up. It’s not human augmentation right now. It’s like this interesting novelty. It hasn’t replaced the ability of a surgeon to do his job. I feel like that is what I’m interested in because I do think that technology is all about making humans better. I think it would be great if we could be more efficient, especially with the healthcare system being what it is like as an access to care issue. My point is that I do feel like technology should make things better. Now, it’s not happening. It’s nice to know that the technology is out there hasn’t been tricked out.
One of my startups is called mechable. Mechanizable, but not to mecha. I have always wanted to build basically take Earth moving equipment and put modern robotics. These things are happening. As I drive my tractor around, the hydraulic controls are clunky. It’s probably very similar to the surgical robots at a different level. You’re controlling this joystick. It’s moving something. I should be able to have a much more natural interaction with those devices.
It is exciting what we can do now and it takes some time. The thing that is important for people to realize in the adoption of robotics IS that robotics is a physical product, whereas in software, you can hit a button and get a download of a new version of the software. I don’t know your Facebook app or whatever. Somebody has a great idea and they can code it up. It’s worldwide. Robots take some time.
The hard work component is a lot more difficult than the stuff and expense.
It will take longer. At the same time, you will get to this point where once the hardware is capable enough, things do move at the speed of software. We were real leaders in this space because we, in this case, I’m talking about robotics because we were one of the first ones to push the idea of robots as a service. You buy a piece of equipment, but that piece of equipment gets better and better the longer you own it.
That’s not normally the way the world works. Normally, if You buy a car, it’s good at first. That wears out and it gets crappy. If you buy a forklift, it works well and then performance degrades over time, eventually throwing it away and getting a new one. It’s quite the opposite now with our products because they’re software-enabled. You buy this product and it works well, but it works even better next month. It’s smarter, faster and safer. The month after that, it’s smarter, faster and safer. It’s happening. It’s exciting to be able to also get that data.
The key is the data coming back. When you sold a product, it was out the door and then you had to do a survey and ask the customers if they liked it. Most don’t answer, and they wouldn’t get any good feedback. Now, you’re monitoring the actual performance of your product around the world in real-time many times. Now we’ve got AI and the data processing tools, the massive databases to make sense of all that data. It’s this flywheel of innovation and an acceleration of our ability to improve and innovate.
I look forward to that day. Before we get into the final closing remarks, my big question is why do you feel like it’s taking so long? I’ve been watching the Boston Dynamics videos for decades. I see how far they’ve come and it’s impressive, but it hasn’t permeated past this prototype phase.
One of the things that’s important for people to understand about videos that you see, and it’s not always the case, but oftentimes, that one clip you saw where it did the backflip perfectly quickly and landed might have been 1 of 100 videos they shot and 99 times, it didn’t work. They caught it once, and it worked. Maybe it was the technology. Maybe it was almost a random chance. There’s no doubt. It’s absolutely amazing technology, but we often see something and think, “That’s solved. It works.” That stuff isn’t solved and doesn’t work until it works reliably. That is a long-tail self-driving car. It’s a perfect example. Having a GPS-guided vehicle that could stay in a lane we had that many decades ago.
That wasn’t a hard problem and that’s still most of what we have, but now people are trying to deal with all these edge cases and safety, pedestrians and roads changing. The list goes on and on and on, and the weather is a huge one. The key is understanding the difference between having a prototype work in a one-off laboratory environment versus having a reliable product that can be deployed safely and work reliably and be done for a money that makes sense.
Robotics may very well end up being decades apart despite the acceleration in technology. Going back to where we started the conversation, humans are amazing. I would encourage everybody to try and build something. Try and figure out how to build a machine to do a little task that’s easy for a human. You start to get an incredible respect for how amazing human beings are.
It’s very difficult, but it’s still exciting to watch. We appreciate you coming by here and telling us a little bit about everything that robotics has to offer for the rest of humanity. I feel like the time flew by. I did want to end with the same three questions that I ask all my guests. The first one being, where do you draw your inspiration from? For me, it’s science fiction. I look at the utopian visions of Science Fiction, like Asimov, Star Trek and all those different interesting science fiction books, novels, movies and TV shows. That’s something that I hope that the future looks like. You’re building something that is very futuristic, which is robotics. What do you draw your inspiration from?
Most of my inspiration comes from playing. The reason I bought this farm in California was because I didn’t want to be an engineer who tried to solve a problem I didn’t understand. I want I do things conventionally. I went to the strawberry farms in Salinas. I spent time picking strawberries out there in the sun with the workers. That is hard work. I go to my farm and I’ve got my conventional tractor. I do my conventional farming. That, for me, is incredibly inspiring because I’m doing these tests. I start to understand them at a deep level.
That’s where the inspiration comes where I see, “I can’t automate this whole thing, but I can automate that piece right there. It’s going to be helpful. It’s going to save money and the environment.” Play is an incredibly valuable source of innovation and inspiration. I would always tell my engineers I’m like, “I got to go out. You got to do that thing. You’ve got to play with it. You have got to have fun with it. You have to be excited about it. Don’t jump in trying to design a tool or machine to do it.”
You got to go live with it first because the physical world again is complicated. I can’t tell you the number of times I’ve designed something and I’m like, “This is going to work for sure.” Some weird thing happened and I realized, “The friction here is not.” It’s all because we create very simplified models in our brains of how the world works. All models are wrong.
I’m reminded of the Elon Musk philosophy of treating life as a video game. That’s certainly something that’s very difficult for a lot of people to achieve, but I’m glad that you’re able to live your life like that. The second question that I have and I ask everybody this, and I feel like we’ve covered it a little bit. What do you feel like the robotics future is going to look like in ten years? How do you feel that robotics are going to be ingrained in our lives in ten years?
You’re going to see a whole lot more of them. We’re at the point now where we’ve shown and found in many cases where robots do provide value. We’re starting this adoption curve. You’re going to see that there are going to be hundreds of thousands of robotic forklifts out there because this is a job that’s ideal for a robot to do. That’s freeing people up to deal with the exceptions, things that robots aren’t great at. Are we going to have the robot butler in ten years? Probably not.
You are not an optimist. You don’t think so?
Probably not in ten years. If you’ve got a lot of money and you want to get an early prototype of a robot. We have a robotics vacuum cleaners. I don’t like to try and predict the future. There are surprises all the time. Somebody’s going to come along and some smart person is going to find this perfect formula of, “I can have this um bipedal robot that does this great job.” The thing that’s pretty amazing. We can thank Boston Dynamics and MIT for doing a lot of the groundwork for this, but you’re seeing the price of quadruped robots and bipedal robots plumbing. One of those robots would have cost $1 million to build until recently. Now, you’re seeing humanoid robot companies pop up all over the place. Are they going to be around in exactly what they will be doing, or will they be mostly novelties? We’ll see.
I hope that technological deflation keeps happening. I was talking with this guy. I’m definitely one of the first adopters. If I can even have a robot that waters my plans, washes my dishes and folds my laundry, I’m sold if it can do those three things. I’ll pay any price for that. I do look forward to the technology being more affordable and accessible.
Last question. There are many different and interesting technologies that are happening. As you were saying, we’re on that hockey stick path as a species. What other than robotics excites you because I feel like there are many articles that I read. I could tell you AI is the lowest hanging fruit, but what is interesting to me is also genetics longevity. Robotics is something that interests me, but what about yourself? Being this thought leader in the robotics space, what are you most excited about?
To tell you the truth, one of the things I’m most excited about is what I would call simulation and optimization. You might not hear this often, but I think our huge opportunity as a species is to start optimizing In ways that we haven’t been able to before. We have a tremendous amount of waste right now that occurs in our economy. We get packages delivered. We’ll sometimes get five different delivery trucks at the farm in a day. I’ll be delivering packages down this very long road that all wasting fuel and oftentimes, it’s one little teeny package. There is much waste going on right now.
Our huge opportunity as a species is to start to optimize in ways that we haven’t been able to before.
With the data that we’ve now got available, we’re starting to get to the point where we need an environment to try it out in that simulation, tracking everything with the technology to process that data. We now have the processing power to build what many people call digital twins of the world of a warehouse or of a transportation system or whatever it is. If we can get those digital twins to high enough fidelity that we can run realistic models on them, then what that allows us to do is basically A/B test the future. A/B Testing where they’ll say, “Let’s design the web page. Let’s design it this way.” Have users tried both and said, “This one is better than this?”
We can’t A/B test the future right now. You can’t say, “What if I ask this question or that question,” and then choose the future that works out best. Simulation allows us to do that to some extent. I’m excited about this. What robotics means is when our robot is trying to accomplish something, if you have this full model of the world, I can then be like a Grandmaster chess player.
What do those chess players do? They see the state of the board. They understand maybe how their competitors think a little bit, but they see the state of the board and then forward simulate 100 different possible moves and then pick the move that gives them the best chance of success. That’s what we can do in simulation. We’re in a warehouse or factory, and I need to try to make a decision, “Do I send a robot to loading dot A or B?” Those things are typically fairly hard-coded based on what
If I can say, “Here’s my simulated warehouse. I’m going to send one robot to A and I’m going to send one robot to B. I’m going to see what happens.” We’re basically going to cheat and play time forward faster than accelerate time. You’re not guaranteed to get the right answer, but the chances of getting to a solution that is most likely to provide the outcome, like in the chess game, is much higher. Simulation allows us to optimize, which allows us to further accelerate and make better use of our resources than ever before.
We did this very interesting study at one of our customers where we installed robots and had the robots do what the humans did on a standard, pre-programmed schedule. It was great. The robots showed up for work. They were reliable, all of that. What we found as soon as we said, “Now we’ve got robots that can respond to changes in commands in real-time, what if we now make this so much more dynamic where the robotic system and the humans involved working together could respond in real-time to what’s going on upstream or downstream in the warehouse or factory?
We found that if we could do a better job of optimizing in real time what was going on, we could get twice the throughput 2X productivity in the same space without adding any resources. You’d think, “That doesn’t sound realistic.” but it is realistic if you realize that in any system, the throughput is always limited by one bottleneck. There’s always one limiting factor.
Humans are terrible at figuring out what that bottleneck is in real time and dealing with it. With these electronic systems with access to data, we can now say, “Things are slowing down on conveyor about C. Let’s fix that immediately. Let’s not wait until it brings everything to a grinding hall.” By achieving that next level of optimization, you get this exponential boost in your productivity across the entire system.
That level of productivity boost can literally happen across our entire world economy. As we integrate more, simulate more and take more advantage of AI processing of massive amounts of data, you’re going to see massive changes in that area. There are about butlers who will come and they will be great. I think the thing that changes the trajectory of humanity is our ability to optimize our research utilization.
The thing that really changes the trajectory of humanity is our ability to optimize our research utilization.
I’ve never heard anybody talk about that before, but I could see the dividends that it would pay. This has been an interesting conversation. I appreciate you coming by. I’d love to have you back if you’re ever doing anything cool. We’ll be happy to talk about it. For those of you who are reading, check out Daniel on all his social media. Check out what he’s doing with MassRobotics and Vecna. For us, please like and subscribe. For those of you guys who are following us on a regular basis, we will see you in the future. Thanks, everybody. Have a good one.
Important Links
- VectorRobotics
- MassRobotics
- Daniel Theobald – LinkedIn
About Daniel Theobald
Daniel has decades of experience leading research scientists and teams of engineers in developing cutting edge robotics technology. He has 67 issued patents and over 30 patents pending. He achieved a perfect academic record and was chosen by the faculty at MIT as the engineering student most likely to make a significant impact on engineering and humanity, He also received a prestigious Hertz Foundation Fellowship. In 2018 he founded Vecna Robotics, now a global leader in autonomous forklifts. In 2014 he co-founded MassRobotics, a non-profit dedicated to the global advancement of the robotics industry. His passion project is agriculture mechanization at Twisted Fields research farm on beautiful Rancho San Gregorio. He is currently focused on helping large organizations successfully automate through his latest startup Mekable.com. He is dedicated to the idea that technology can be used to improve life for everyone on the planet.
0 Comments