00:44:11 · Steve Xie, Jonathan Stephens
[00:44:11] S0: Totally. Alright. [00:44:12] S1: Totally. Then moving forward, just kinda wrap this up. What what part of that stack do you think will look most different in, let's say, two years from now of those three, world or the world, have the behavior, we have eval, which one do you think is gonna really like change the most? What we're gonna see a big advance in? Or or is it gonna just have to be all three?
00:44:33 · Steve Xie
[00:44:33] S0: Totally. So I would say first, you know, we introduced the three layers, the work behavior, the evolve. But, actually, there will be much more connections among those three and also between real and also seen. We are actually truly building the flywheel. For example, you know, we started as purely simulation. Right? Then we found out that, you know, we need to scale simulation. We need to get real data. Then we get into the Ego centric data collection. And now we are building a really scalable pipeline from real to seen. Right? From Ego, from EgoSuite, from the behavior layer to actually the world layer, to actually the evolved layer, which is amazing. Right? But then the other thing that we are building right now is how can we best leverage our evolve layer? Right? Our evolve layer right now is we are providing the evaluation framework, evaluation products to the whole industry, to the academia, to help them understand where they are. But, actually, can we light our eval layer, to help us understand where we are with our work, with our behavior? So that's the thing that we are building right now, which is amazing. You know? We are we are actually leveraging our evolved layer to helping us to benchmark our EgoSuite right now. So meaning, leveraging a simulation evaluation to help us now... Then, basically, we will get this flywheel. Right? Meaning, basically, from, you know, the real behavior from the real EgoSuite. Right? We are getting into simulation. Right? But also from simulation, we are actually improving our real data collection from EgoSuite. Then actually, we are building the real flywheel. So this is basically the major focus I'm with right now, to building the real flywheel, a scalable flywheel, I would say. Right? From real to scene, but also from scene to real. You say that... Yeah. Go on. Go ahead. Yeah. Go on.
00:46:31 · Jonathan Stephens
[00:46:31] S1: I was gonna say in two years, you can't say one's... Like, one thing has stood out because they kinda all need each other.
00:46:36 · Steve Xie
[00:46:36] S0: They all know grow Exactly. Together. Exactly. Yeah. Yeah. And the second thing I I see is for these three layers that is going into a loop, there needs to be a center. Right? There needs to be a core. So what is the core? Right? So what is the core? I really see that the core is agent. The core is AI native. You know? So so, basically, you know, we are we are actually the infrastructure. We are not data operations. We are not simulation operations. We are actually the engineering infrastructure, the data infrastructure, the simulation infrastructure of physical AI. And in order to scale this infrastructure, in order to scale data, you know, we need to have a core agent that's sitting in the middle. Right? Mhmm. That's helping us to understand about the data quality, the simulation quality, and helping us to improve it. Right? So that, in my opinion, is another, most important thing. Right? Because, you know, we, so, so think about this. For EgoSuite, for EgoCentric data collection, in my opinion, in five years, there could be a 100,000,000 to 1,000,000,000 people collecting data for robotics. Why? Because there are millions of cars collecting data for Tesla, and it's it's just self driving. Self driving is mainly vision based. It is not that much physics based. You don't need to touch the world with self driving. Right? You just need to make sure that you don't hit the other cars. You do hit people, but you don't need to make sure that you can touch them. Right? But for robotics, you need to touch everything. Right? So the the diversity, the physical parameters, you know, everything is much more difficult. The the robotics is probably 1,000 times more difficult problem than self driving. And 1,000 multiplied by 1,000,000, you get a billion. Right? So I really imagine that in five years, you really need to have a billion people collecting data somehow, like, either positive, like, either active or negative, or or or, so either active or passive, for robotics. In order to drive this large data infrastructure, it's better that you have... You can automate everything. You can automate everything. Right? So I would say the agentic part is extremely important, and, also, it's our core focus right now. You know? We, rather than having technical artist, like real technical artist, we have technical artist agents. Right?
00:49:27 · Steve Xie
Rather than having, the data quality manager, we have the data quality agents. Mhmm. Right? And those are extremely important in our infrastructure, in our system. So I really see that, you know, in the future two years, the the one thing is to have this flywheel, right, to connect real and SIM, and also connect back SIM to real. Right? The second thing is to really stronger the core, to better agent to grab everything.
00:50:00 · Jonathan Stephens, Steve Xie
[00:50:00] S1: Now that that actually connects well with the the the first podcast I put out with. He's ex Amazon, and he built... He's building an operating system for, robots. And again, it's like that agentic layer. He's like, the only way that we're going to do this fast enough and it scales that we're not having a human sitting there doing the menial tasks. We're having an agent handle that and we just guiding it because there's enough information out there now. It shouldn't need that one person sitting there touching code necessarily. So that makes sense with, you know, I'm just, I'm hearing that from multiple people now in robotics saying we'll be interacting with an agentic layer. Now the ecosystem's being built. There'll still be a lot of people having to be building in the background, but as an end user, it will no longer be this Jonathan saying, Hey, I wanna, I wanna, I don't know. I need this kind of environment to test and train on this. Well, I'll just tell Lightwheel agent to glare that, and it will pull the pieces it needs, generates the backgrounds it needs, does the test and then does the benchmark. Right? There's not a human sitting there just moving files around all day long doing that. [00:51:08] S0: Totally. Yeah, I will say next time you can interview my agent maybe. [00:51:14] S1: Yeah, yeah, yeah. Sipping some cocktails somewhere on a beach while your agent's figured out. Exactly. [00:51:20] S0: Yeah, exactly. I'm looking for the day that I will report to my agent.