Full Transcript Below

John Mauldin: Raff, I’m really excited to be here with you today. We get to talk about a topic that I’m a hobbyist—but you are the professional—and it’s one that’s very, very interesting to me, and that’s robotics and automation. Give us a little bit of background of about what you see here, first of all, then a little bit about what you’re working on right now and some background on what you’ve done and what you see coming down the pipe.

Raffaello D’Andrea: Right, thanks for having me here today virtually, through our Skype meeting. I’ll tell you what I’m working on right now. My lab here in Switzerland is focused on creating machines that do things that have never been done before, and in particular having machines that learn, adapt, and improve their performance based on experience, so the more they do something, the better they get at it. We’ve done a lot of work with flying vehicles as of late, but we also do more traditional things with say, robot arms, but I would say half the focus has been with flying machines.

In terms of what I’ve done in the past, from an entrepreneurial perspective, I am one of the cofounders of Kiva Systems, the company that makes robots for distribution facilities. Imagine having hundreds if not thousands of robots loose in a distribution facility, moving material around, still having people involved in picking orders, but the majority of the movement being done by machines. Kiva was bought by Amazon last year. So that’s what I’ve done in the past—and just as an aside, although I don’t think it’s relevant for the discussion today, I am also the creator of Dynamic Sculptures, and some of my work has been exhibited worldwide and is in several permanent collections.

John: We’ve had a pretty phenomenal run in the robotics and automation world in the last 10, 15 years—almost six or seven times outperformance over the regular stock market. What are the big changes we’re going to see and notice?

Raffaello: I’ll answer your question, but first I’m going to suggest something that your viewers and your followers can do. It’s a simple exercise that they can do themselves to figure out where the market is going with this, and that’s the following: In your everyday life, think of something that’s repetitive, something that you see, something that maybe affects you directly, but even something that you observe, something that is repetitive, and on top of that, something that is boring. If you have those two things, it’s likely that in the near future that will be replaced by a machine. The repetitive part is a necessary ingredient, because that’s what machines are very good at doing. That’s where they excel, especially with new algorithms being developed to improve performance the more that they do something. And the boring is an indirect measure of how cognitively simple something is. If we find something boring, it’s likely that a machine, a robot, can be programmed to do it. So in general, look around. Something that is repetitive, something that is boring, it’s likely that it will be automated.

Now, specifically, what do I think are some markets that are ready to be automated? I would say some of the big ones are agriculture—I think it falls into that category of something being repetitive, and driving a tractor around is probably very boring. I think mining is also another example of something that can be replaced. Manufacturing, it goes without being said—moving things around on a manufacturing floor, assembling parts, the same thing all the time. I think those are three simple examples.

John: Let’s talk about Google for a second. They’ve been acquiring robotics companies and information companies. They recently acquired Nest for what seemed like the outlandish sum of $3.2 billion for a $100 million a year company. Interestingly though, I’ve just installed Nest thermostats in my house, and they are way cool. They actually program themselves by seeing how I set the thermostat, so then they basically figure out my lifestyle, my rhythm, my patterns, and then program themselves. And they do it for a third of the money that other thermostats would cost to put in, so it’s an even cheaper product, and it’s all wirelessly hooked in so I can control it from my iPad. That’s kind of the wave of the future. I remember when Google bought YouTube for, I think it was something like $1.6 billion or something like that, and everybody said, “Wow, that’s a lot of money.” Here we are four of five years later and YouTube is said to be grossing $5.6 billion a year. In terms of looking at the robotics investments that Google is making, what is that telling us that they see?

Raffaello: Well, obviously there is no substitute for actually asking them directly. I’m actually going to visit the folks at Google in two, three days, so maybe I’ll have an answer for that, but of course, it may be that I wouldn’t be able to share it. But, just taking a step back from that, I think that when you look at robotics and automation, really you can think of it as five things. It’s sensing, computation, actuation, communication, and algorithms.

If you take something like the thermostat, the example that you just talked about, here we have the ability to sense things in your house, to actuate the furnace, and then there’s the algorithm, the intelligence, and the communication. So what’s interesting is that to have something learn your behavior, that’s not outside of the realm of existing algorithms. This is something that people have been working in research for quite a while now, so that’s just a good deployment of those algorithms.

Where it gets really interesting is, what happens if you have that information from thousands of homes? Then how much do you really need to learn of that specific user, and how much can you simply learn from the collective of users and then after only a few iterations based on what thousands of other people have done, the system can get a pretty good indication of what you’re going to do. So this collective learning, I think, is very exciting. I think this ability to collect information from not just one user but thousands of users and intelligently using that information to make decisions is very exciting.

I could say the same thing when Amazon bought Kiva for $775 million. A lot of people said, “That’s a lot of money for a company that’s only a couple of hundred, three hundred employees.” But people looking back at it now are saying, “Wow, that was a really good deal.” So it’s hard to say how this will play out. I think that Google is making some serious bets in robotics and automation. It’s going to take time for some of those bets to play out. Some of them will play out immediately, but when you’re dealing with the physical world, I think the timeframes are longer, simply because it takes a long time to really flesh out the technology, either for safety issues or robustness issues—things breaking down. There are a lot of things that have to be figured out, but they are making a bet that within five, ten years, a lot of those issues will be resolved.

John: How difficult is it to keep up with the pace of change in the robotics world?

Raffaello: That’s a good question. I guess I would have to compare… I’m trying to think of what to compare it to. Say something like medicine. I think it’s much easier than medicine, where there there’s just so much knowledge that is anecdotal, that doesn’t carry over into other realms. I would say with robotics and automation, there are a lot of constant themes. You develop a certain algorithm, then that algorithm will be used in many [ways]. Let’s say you develop a really good vision algorithm for recognizing something. Then that will get a lot of traction, and then you can use that algorithm for mobile robots, for flying robots, it doesn’t matter… you can modify it.

If you’re in the space, of course you will have your own certain area of expertise. Mine, for example, is dynamics and control. I don’t think it’s that difficult to stay up with the advancements. I think that what is difficult is to figure out how to deploy those advancements. What are the incarnations of all that technology that will actually be successful in the market? That’s very difficult to keep up with.

John: We’ve touched on this, but let me give you the chance to talk about the ways, specific ways that robots and automation are changing business habits—going to in the next few years, what are your personal favorites that you see coming down the road?

Raffaello: One that has been getting a lot of attention lately, and I think deservedly so, is agriculture. I think that has a lot of potential for automation. But I think that there are also other things that robots and automation can do that people cannot do well, and I think we’re going to see a lot of those innovations, and I think the best examples of those are in surgical robotics.

I have a colleague that I just recently visited in the Netherlands, and they are going to spin off a company that does ocular robotics to aid surgeons in eye surgery. And basically what they can do with this machine a surgeon just cannot do. They can do procedures that a surgeon could never do before. So I think that’s also very exciting. New capabilities, to be able to do things that we could not do before simply because the things are too small or we don’t have the strength or what have you. So I think that those are two exciting possibilities of how the technology will be deployed. I think manufacturing is also an obvious one.

John: Well, thank you, Raff. I appreciate you taking the time, and I look forward to catching up with you again sometime in a few months—and maybe you can tell us what you learned at Google.

Raffaello: Okay, thanks a lot. Have a good day.

---

« Back to Mauldin Economics