Although there is some robot research I've been working on that I alluded to in the last article, I thought of something else that was pretty important and related to robotics that I wanted to write a short article about.
Just FYI, I'm doing this article using voice narration just to give my wrists and fingers a bit of a break from the keyboard, but I don't expect it will affect the content too much.
So, the question is, why don't we see more robots in our daily lives? It may sound like a silly question, but it's something I was thinking about just recently and I have some ideas, so I'm going to go ahead and talk about those in a little bit of detail.
What robots do we see?
First, what kind of robots do we see in our daily lives? Well, as many of you know, probably one of the most popular robot genres is the vacuum cleaner robots. For example, Roomba, and the one I have now is a Shark, because I think Roomba is not selling well lately for various reasons. But those are pretty common. I've even seen a few websites that are just called something like robot ratings, that are just about vacuum cleaners. So it's pretty clear that they're very popular. I don't think everybody has them, but I've definitely been to a handful of houses where people have them, and they seem to be pretty commonly used. I don't think they're perfect. There's definitely some issues. But for many types of houses, I think they're totally worth the cost. So that's one type of robot that we do see, at least relatively commonly.
Another one that maybe is a little less common to actually see in daily life, but I think at least at some point was selling pretty well, is drones. You can go somewhere like Best Buy and see a bunch of these drones. Now, those are generally pretty expensive, in the hundreds of dollars. You can get cheaper ones at other places like Walmart and Target, probably in the $40, $50, maybe $100 range. I haven't actually seen too many of those in daily life, just once in a while. I do own a drone, and I use it once in a while. I have lost one and broken a second one, and there's some challenges there. But I think they're a lot of fun. But even having said that, I just don't see them that often in daily life.
What else? Keep in mind, the definition I'm using for a robot here is just anything that's basically a machine that kind of moves around.
One example that just came to mind is a sushi place kind of near me. It's a Japanese company and they have a robot that rolls in and brings your wasabi and your drinks, while playing some funky music. It's not very expressive, but I guess it is kind of cute. It's slow, but it does navigate fairly well, and I've never seen it make any big mistakes. It'll wait until you take off your products and then eventually it will go away. So that's another fair example, I guess.
I'm struggling to think of other ones. I've definitely heard stories of organizations that own things like robot dogs and use them for inspection or patrol or something. But honestly, I've never seen any of those. I'm sure YouTube is filled with them, but in my daily life, I just don't see them. Maybe I should get out more.
And in my travels as well. I've been to a few places, Japan a few times and a few other countries here and there, and I just really never see many robots. I did talk to a friend recently who said he saw some pretty well-designed robots in Disneyland in California, and I saw a picture of them. They looked pretty cool. So I think there are some rare scenarios like that, like theme parks, where they have enough money and they really care about the atmosphere.
That does remind me, there is one robot that I saw in Japan several times in stores, Pepper. I saw him a few times. And there was another, a robot hotel called Henn na Hotel, that I went to once in Japan. This was three or four years ago, and the theme was that there's nobody in the lobby, there's just robots, and there was a dinosaur there. And again, they weren't very expressive. I think they went out of business, or at least the location that I went to went out of business. I don't think it's there anymore.
And that reminds me, even from decades ago, there's the whole technology area of Epcot Center in Florida, and they had kind of a robot theme. If I remember, they had different types of robots or droids. But again, that was something more just to show off, kind of like talking about the future. So I don't really count it.
So yeah, I can't really think of much else in terms of actual robots.
Why is this the case?
So I guess the next question is: why is this the case? Robotics is a field that clearly has been going on for a long time, and from what I've seen, there's some pretty advanced robots out there. Just on this website, I've categorized over 600. I went through a lot of those descriptions myself and edited some of them to get a feel, and it's pretty amazing. I mean, just in the last 20 years, there's been some huge leaps in terms of humanoid robots, and even things like cooking robots and window robots, pool cleaners, that kind of thing, things that just weren't around, or at least weren't popular, two decades ago. So there's been huge advances, but it just seems that there are a few things that are stopping them.
Shut-in robots
One of my theories is that these robots actually are relatively popular, at least based on some of the sales that I've seen during my research, thousands or millions of units sold in some cases. And frankly, they just sit in people's homes, and maybe they don't use them that long. They buy them and they get bored of them and they throw them out, or they throw them in the garage. I've definitely done that myself with one or two robots when I was younger, but that would help explain where the sales came from and why we don't see them.
Even if they're using them at home, they probably don't take a lot of these robots outside because frankly, they're not made for outside, like a lot of the more toy types of robots or rovers and stuff. I mean, there are clearly ones made for outside use, but many are inside only.
Fear of loss or damage
I think one of the other concerns is that, frankly, with an expensive robot, even if it is made for outdoor use, the owner is going to be concerned about it getting lost or damaged. Even if it's in theory waterproof, the nicer ones are going to be thousands of dollars. So it's going to be a real risk.
I mentioned my drone earlier. The first one I had was a DJI Spark, and it just flew away one day, and they ended up giving me a free drone. Maybe it was because I was near metal or it was a bug in the code, I don't really know. But that's a few hundred dollars, and since then I've been much more careful with how I drive it. And frankly, I'm just afraid to send it out very far.
Similarly, somebody who has a robot, even one that, let's say, runs in the grass, or a robot that walks around outside, probably doesn't want it to get damaged or stolen, or who knows what. So I think that's one of the reasons we don't see them that much in daily life.
Good enough, and worth the money?
And another one might just be that they're not safe, or effective enough. For example, humanoid robots doing anything that's advanced. Let's say there was a chef robot, and it was able to cook skillfully. The question is, would we put that robot in a restaurant? Would that be cheaper than a real chef? Would it be better, and safer than a real person? In cases like that, I think even now the real chef is probably going to be way better.
That does remind me, there's two places I've seen in Portland, one in a mall and one in downtown Portland, that had a coffee-making robot. It was literally a robot arm with a lot of joints, and it was pretty cool and relatively accurate and kind of fun to watch. It was slow, but still pretty cool technology. One of them was in the Washington Square mall, and it left a long time ago. So I'm guessing they just didn't make enough money. There's obviously a lot of coffee shops that are competing. The other one, I forget where it was, but it was a Japanese store in downtown Portland. I think that one's still there. So that's another kind of rare example, but I guess the question is, are some of these robots really good enough? Are they worth the money?
For example, I read when I was researching robots that the Portland police department had bought one of these super expensive robots. I think it might've been Spot, one of the $150K robots, which they bought through some grant, but they only use it for really special occasions, like a bomb threat or something really serious. So I guess you're not going to see it just going on everyday patrol, probably for the same reasons.
It reminds me, I was watching a video of Spot. It was a different version, I think the $80,000 version. This is one that I think is designed for inspection of oil places and such, and it uses custom parts, so it's extremely expensive. It's because they haven't mass produced the parts, but anyway, it looks really cool and it moves great. I saw somebody doing an unboxing and demonstration of it, and I think he was controlling it with a joystick or controller type thing, and it fell. I don't know how many hours he was using it, but in the first hour of the video, it just fell. It was walking on a place with some elevation changes, and it just fell and it couldn't get up. So that was a little embarrassing, and a little surprising, I guess, because it's an $80,000 unit and it can't do something as basic as getting up. And I've seen $500 robots getting up.
Now, again, that might've been a rare situation. My guess is when they're actually used in practice, it's more the robot's AI itself than a human, and it's probably very careful about where it walks. Whereas if you're doing it yourself on a controller, it's going to go wherever you tell it to and probably not stop. I don't know, maybe he even hacked it or did something that you're not supposed to do. So it's not a good sample, but nevertheless, it makes me wonder about even these really expensive robots. How good are they? How well can they handle a variety of real-world use cases?
Safety
Obviously you look at something like a car, the self-driving car, which they've been working on for many years now. It's similar to a robot in many ways. I guess you could kind of call it one if you want. And obviously that's something where if you screw up, you're in serious trouble, because you could have many people going to the hospital and getting injured.
So that's an extreme example, but even a robot that's able to carry heavy items, or even everyday items, is going to need to have pretty strong servos or hydraulics or whatever it uses. And that's enough force to potentially injure somebody. So it needs to have good safeguards, maybe somebody watching it. It has to be very careful. So I think that's one of the integration challenges with getting a robot in the real world. Not only does it have to be robust and worth the cost, but it has to be extremely safe. Especially if it's a place where there are kids and the kids are going to walk up to it, you have to be super careful not to hurt anybody, or maybe even not to scare anybody. So there's definitely a lot there that needs to be fulfilled until it's popular.
Programming and testing
I'm not sure if there's any other reasons. Like I mentioned earlier, the cost is kind of a big factor. But having said that, as I mentioned in my last article, I think the introduction of strong AI like Claude and some of these other platforms makes robot programming a lot easier. It doesn't mean it's good robot programming. That's something I'm personally working on, to see what level of quality you can get. I'm working on a room navigation project now, but the question is what level of quality can you get using something like Claude, where you're probably not looking at any of the code, or most of the code. It's mostly written indirectly via prompts.
And then of course, there's going to be a lot of weird use cases where you don't really know what will happen until you try it, because you've never even seen the code. It's probably too much to read or too hard to read, although I could in theory read it. So instead you need a huge amount of testing. Fortunately there are simulators for a lot of the robots. Even for this stuff I'm doing, I have an ad hoc simulator, but that only goes so far. There's a bunch of things that you just can't capture, at least with the simple simulators. So you've got to do simulation, get as many bugs out as you can, then put it in the real world.
And you can't easily speed that up. I mean, you could buy a hundred units, but you've got to do real tests that take time. If there's a problem, you've got to fix it or return it to the charger yourself or whatever, which is something I'm working with. That's another reason. The actual testing takes a lot of time compared to something that's virtual, where it's all software. We have such fast machines that if you have software, you can just test the heck out of it if you buy a fast machine. Whereas if you have a robot, especially a lot of these advanced ones, they don't move very fast. I think it's probably for safety, and maybe for other reasons as well, like the way the servos are designed, but they generally move relatively slowly compared to a human. And that means it's just going to take that much longer to simulate a single test. Just do some action and see how it works out, then repeat.
Anyway, I'm planning to do a lot of studying robots in the next few months, and I hope I can give a better, more informed opinion on this stuff, because at this point it's more of an opinion piece. But I did want to get these thoughts down, and I hope you found this useful.
Thanks for reading.
—J.D. Wisgo