Realbotix prioritizes autonomy and useful tasks over walking

Transcript

Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.

1. 55:32

   and people are like but here's the thing all of those robots that i've seen in any of these events or remote controlled. There's no AI, none. So there's a guy driving the robot. And like he said, there's always security and there's falls, there's accidents. So there's this huge sort of separation between the reality of what's going on and what people think is going on. Yeah. They're pre-programmed, right? Like here's 20 tasks for you to remember within this environment, but there's no real AI there. No. And here's the biggest question is what is it good for? Like you see, you see a robot, you know, like walking and you see a robot waving. Like, what is that good for? Yeah. Did you say busking? Yeah. Yeah. Yeah. Cause a lot of these robots, I mean, there's the, I think, I think they're the novelty and we're the real thing. You know,

2. 56:28

   A robot that can do karate chops and has a 30-minute battery life, what is that good for? Right, right. Okay, and then coming here and seeing these for the first time, I mean, the uncanny valley question that you brought up in the beginning, that's always a big thing here. I mean, there's parts of this that are extremely realistic and a huge advance beyond things that I've seen before, and then there's parts where, if I'm being totally honest, you know it's a robot still, you know what I mean?

3. 56:49

   So, like, do you guys feel like you've hit, that you're satisfied with where this is? No, no, no. We're working on this stuff every day. You know, we're hiring new people. We've just hired a new COO from Agility. We're working, like, we've got huge plans for this company. We're not stopping here. This is version one. And, you know, I will always hang our hat on the autonomy. I was telling you earlier about when we were in Times Square, there's a store called Tomorrow, high-end electronics store.

4. 57:17

   we sort of taught the robot about the products and the floor layout and then we're like let's open up the doors and have it facing you know one of the busiest intersections in the world and we sort of watched it did thing and then we were like let's go for dinner you were going to a show with your wife and you know we left the robot there with a security guy sort of a few feet away and just let it do its thing for several hours yeah we come back and we look at it it was talking, you know, German, Swiss, Hebrew, whatever, to different people, people taking selfies, it's waving, it's doing things. We just left it there. And we did that. We left it there for six days. Like eight o'clock in the morning till eight o'clock at night, store hours. We let the robot do its thing. Nobody else is doing that. You don't see anybody else doing that. A couple of weeks

5. 58:04

   later, there was the, the optimist was there for like 10 minutes serving a drink or something that was tele-operated. There's no tele-operation here that we need to do. We can do that as well, but these robots act and think autonomously. And just to touch on your question about improving, and you can still tell it's a robot, but that's on purpose. So could we go further in terms of, let's add some more skin pores, let's add some tiny veiny details. And I think that's that line that you cross, and then you're in the uncanny valley. That's what I wanted to ask.

6. 58:36

   He's like, you've seen robots, I'm sure. Yeah. Ishiguro does it. Yeah. Where the face looks like literally was peeled off of a cadaver because the detailing is so high. It is, but even when I'm with him, I just wanted to talk this out with you guys because when I'm sitting with him, he's like, you can't tell this from a human. I'm like, yes, I can. Yes, you can. Right. And I can't tell if he really believed.

7. 59:01

   Right. No, but I want people to be able to tell it's a robot. You do. Okay. Yes, absolutely. Until the underlying mechanism and control systems are able to move in such a way that they're indistinguishable, there's no point in making the skin indistinguishable. Because when you do that, one thing looks super real and the other doesn't, and that causes humans to recoil. It's built into us. We react, some people react to it as if they were looking at a dead person.

8. 59:33

   Yeah. Right? Or something creepy. Often what we do is we put the robots in vests. And the reason why is I want people to see the hands. And when you see sort of the hand articulation, which is, you know, we put actuators in the same place where you have knuckles. I want people to see that this is a robot, just a very friendly robot that draws people in. And Matt, you know, we talk about this. CS, our robots bring in massive foot traffic.

9. 59:58

   It's like traveling with a rock star wherever we go. After CS, we normally get 30, 40 invitations all over the world from people who are like, can you please bring the robots here? Because nobody else has something like this. But when you're talking about, this is the heart of my question, if you're talking about this being a hospital, in a hotel, in a casino, do you want it to look as lifelike as possible?

10. 1:00:19

    Is that what it would take to make it super mainstream? Or you think it should be distinguished from? Like I said, until the underlying mechanisms can move realistically, There's no point in pushing the outside to that point. So is that something you guys... They have to work together, right? Is that something you guys would work on? On both, yeah, absolutely. So what's on our roadmap for the next couple of years is hands that can articulate a lot of things now but are capable of small tasks.

11. 1:00:50

    So maybe feeding someone. We're working on the ability to type on an iPad so they can transcribe things. So small tasks like that, opening a pillbox and giving someone medicine, things like that is sort of on our roadmap. Walking is not on the roadmap. Not right now. At some point when it's perfected, which I still think is years away, we'll find a way to replicate that, but it's not on the roadmap today. But we want to work on the AI, some of the things that might be useful to somebody. Like even if you were to have it in your house in the kitchen, if it could actually lift small things and articulate certain things, that I think is far more useful than a robot sort of clunking around your house and bumping into things because it doesn't even have a real vision system.