Going Direct Conversations
00:32:28 · Michael Cho
[00:32:28] S1: Yeah, you can imagine, you can use a simulation engine to, let's say, simulate a robot frying egg, let's say. You can imagine in your head maybe an infinite way to set up that scene, that task. Different lighting, different robots, different eggs and whatnot. And so it will take a crack researchers like Chris to drastically shrink that search space into something that is tractable with the compute budget that he has. But as researchers do that, well, so, you know, they are missing a lot of actually potential edge cases that may happen in the real world that they could have simulated, but they just don't know. I mean, they would have messed up. And so I think that's a big part of it. So if I look at it that way, actually the bottleneck to a simulation, a big bottleneck is having more people like Chris, honestly, because that process is very curated. Like it's very handcrafted. Like you need, like Chris will have to come up with very like a man, like handcrafted, like the rewards functions for the reinforcement learning, things like that. It's all very handcrafted. So, which means actually the human in the loop, and unique researchers to do it. Like, you know, it's a very, very specialized skill set. That's a big bottleneck. World model, on the other hand, all this video generation, I think it's also very, very interesting. I think they potentially can move the needle a lot. But you can imagine to build a world model, you also need a lot of probably like video data and tele-op data too. So you still need a lot of data, a huge amount of data to build a decent world model, right? Yeah, so it's hard to pick one. I feel that tele-op data, video data, simulation, all are needed. Like we, we need all. And there's a lot of papers that we've come across, you know, that we've interviewed that do co-training of various sorts, right? They'll, they'll, they'll, they'll, they'll scale it using some video or like, you know, do the real to seem the real thing.
00:34:24 · Jordan Wolfe, Chris Paxton
[00:34:24] S2: Yeah. So, yep. Sure. What's your thought, Chris? If there was kind of what, if you have one thing that kind of stands out in your mind, if someone who's actually in there doing, you know, working on it every day, [00:34:34] S0: I mean, I think that if you were going to wave a magic wand and had infinite money, then I think you'd just have lots and lots of robots doing. I think that they worked. But practically speaking, I think Michael's correct. It seems like you need a little bit of everything, because they're all good at slightly different things. And they all have different pluses and minuses. Sure. Yeah. So the real solution is going to be a mix. So, Chris, I actually saw a recent tweet that you put out that caught my eye. [00:35:07] S2: And I don't have it in front of me, so I might butcher it a little bit. But basically, the premise was the reality of it is if we want low-cost generalized robots, we need a bunch of robots in the field learning and actually doing stuff. That's,
00:35:07 · Jordan Wolfe, Chris Paxton
[00:35:07] S2: And I don't have it in front of me, so I might butcher it a little bit. But basically, the premise was the reality of it is if we want low-cost generalized robots, we need a bunch of robots in the field learning and actually doing stuff. That's, [00:35:23] S0: yeah, that's increasingly kind of what I, my opinion on a lot of this is just like, it'll be solved with that. Like we have, and that's one of the things is that like the robots are so much better. Like, so I, I've, I've been doing this since taken one form or another since like 2014. Right. So like I've been doing robotics for a long time. And when, when I started, we paid like $40,000 for a UR5 industrial arm that failed like every few days. And like we had a, we had another, like this really, really, not so great robot arm that like, They had hand-tuned PID controllers where one of the grad students stayed up and slammed into the wall and broke it into a million pieces and that kind of stuff. And so this was the reality. And that shuts down the lab for months, many months. So that was the reality for a long time. And now you can buy a $20,000 unitree robot from Walmart. Or you can 3D print one of those SO-100 or What's it? Vector's robot. I forget that. The robot. The various little robot things. You can 3D print. You can build these things. We're so much closer to that world with infinite robots everywhere, I think, than we used to be. So that's...
00:36:42 · Jordan Wolfe, Michael Cho
[00:36:42] S2: Yeah, it's cool. It's cool seeing it all come together. And as someone who, I guess, without that kind of background in robotics and the more folks that I speak to, especially on the research side, and someone always thinking about like market, market application and how to get adoption and useful application of these different robotic systems. Yeah. It's very clear that it's such an exciting time. It's very clear that there's a lot of movement and it's moving fast, but yet I'm challenged because most of the robotics companies that I see, I think the path to any type of adoptability or market adoptability is very challenging. still at this case. And so I don't know if you guys think from like a market viability standpoint that the vast majority is still not viable or do you actually see kind of a window here where it's going to start becoming really viable with any type of scale? Well, I think it's absolutely viable if you look at it as an entertainment thing, right? [00:37:40] S1: So Unitree is putting up... I love it. [00:37:43] S2: He's pulling in the entertainment plug. Give it to us.
00:37:43 · Jordan Wolfe, Michael Cho
[00:37:43] S2: He's pulling in the entertainment plug. Give it to us. [00:37:46] S1: No, no, for real, for real. Like Jordan, I absolutely believe in it. Like, you know, if you look at, Unitree is going to start shipping the R1, which is a humanoid plastic key. Yeah, sorry. I mean, Chris, you saw it too, right? In Korea. So $6,000, $7,000, something like that. So at that price point, I imagine every bar, every disco in the country is going to have a few of these. Yeah. so that patrons can, you know, every bar is going to have their own. It's like having a karaoke system, maybe cheaper than a karaoke system, let's say. You know, like, you know, all these bars in Texas, they have all this, like, you know, the pool thing, the mechanical pool. It's like that, you know, so people are going to have that. And so, and so that's, honestly, this is how I think we'll get to the first 100,000 huminoid. It's just bars, entertainment venue, buying a bunch of these, and then... And then that will get some of this supply chain, you know, more honed up, you know, you're talking about the actuators, the batteries and all that. And once you get to that run, you know, you get a lot more instances of how these things may fail, right? You know, a lot of these actuators may fail, especially on the shoulder. So you just can't really know all this until you apply in the real world. And I think honestly, like the entertainment, everything is the first use case. 100,000 units at least. Not the application that most people are thinking about, but I do think that's realistic.
00:39:18 · Jordan Wolfe
[00:39:18] S2: Well, you know, I think what I like about the perspective, I actually have thought a lot about this because I just started seeing some of these robots showing up at conferences and events and trade shows and not like robotics trade shows, just like generalized, right? And you have like the Pepsi branded humanoid and whatever it might be. I think that that kind of perspective is nice because it's kind of like you take kind of the seriousness out of it and it's kind of like playful and it's playful and fun versus like trying to be so overly, I don't know, just like intense about like... We need to solve general labor and we need to do this and this. It's just like, you know.