00:21:59 · Steve Xie, Jonathan Stephens
[00:21:59] S0: Yeah. Totally. Yeah. Thank you so much, Jonathan, for bringing up this word behavior and also evolve, framework that we use. Right? So I would say the... Yeah. Exactly. The word is basically how we leverage simulation, right, to build the physically accurate and also diverse words and also the similar AI size. Right? The behavior is, where we use the word to automatically generate data. And, the data should be the robotics trajectories. So for example, robotics, right, manipulating the fridge, the robotics opening up, you know, the oven, the robotics inserting cables, things like that, which is also extremely important. So behavior is actually where we generate the data. Right? And, also, we provide this data, to our customers in order for them to train their foundation models. Mhmm. [00:22:59] S1: So you need you need physics accurate data at scale, and so we're using simulation so you can, as you said, you are automating a lot of this collection, right? To go put a robot out in a data center and have it pull a bunch of cables and reinsert them. You have to either go to a different data center to get a variation or you have to physically move these several ton server racks, put different ones in. That's not efficient. So that so that's the... That's that's where this lays? Is this for efficiency that we're doing this all in simulation and not just putting out robots in real life?
00:23:39 · Steve Xie
[00:23:39] S0: Yeah. Totally. So we actually... We have two ways of generating behavior data, which both are very important. So, basically, there is this data pyramid concept introduced by Ego Drew from UT Austin, and, also, he is, you know, the co lead of GEAR from NVIDIA. So you you could point out this, data pyramid is, you know, that, the real teleoperation data, the the real robotics data is very limited. Right? So it's just the tip of this iceberg, which is a very small part. Maybe 1% or maybe even less than 1% of the data needed to train robotics. And the major data needed to train the robotic is, one, simulation data. Second, egocentric data, so human videos. And the Lightwheel, we really see the behavior as those two things there. One, simulation data. Two, egocentric, real human egocentric videos. And we are biting on those two things together. Right? So for the simulation data, we leverage, you know, the simulation words. Right? We leverage our auto data generation pipeline to generate synthetic data from simulation. Right? And this data is extremely scalable. It's extremely fast. Right? And also, it can be in different robotics forms, right, to provide it to our customers, which is extremely important data source. Right? The second data source is basically, we we have colliders worldwide, you know, by... Wearing, you know, the egocentric devices, for example, the glasses, cameras, and also the wrist cameras of, doing different type of jobs. You know? Like, a lot of time, they are doing jobs. Like, they are doing, you know, the house cleaning work, the housework, related work, chores, in their homes. Right? Others are doing the different type of works in factories, in grocery stores, in coffee shops, things like that. Right? And those data are extremely scalable, diverse too. And also, the other good part is is also from the real world. Right? And we provide those two scalable sources of data to our customers. We we check... We we think are extremely important.
00:25:59 · Jonathan Stephens
[00:25:59] S1: I I think so, especially now, not just to have SimTel ops, because I think if I feel like at least one paper or more a week I'm seeing of researchers finding that if they mix in a large amount of egocentric data and a little bit of real robot data, they're getting fairly good zero shot results. They're getting robots to do things unseen and they're not quite sure the mix, they know that egocentric data and what worse, what you're building is high quality. Right? So, you know, some of these teams are just scraping into, then they get VFX, which is not physics. Correct. Examples are getting cartoons. All these things we got to figure out are either filter out or just, taste the data. What we, what we provide, what Lightwheel provides is this high quality annotated physics plausible because it's it's captured in the right way. We even have, you know, the the the arm state data if you want. So you get this high quality data, then you can use this... The robot tele ops data, so they can get much further down the line. So I think I'm seeing, even know he built a robot in seven months and deployed it. So right now, it's all about how fast you can find the right mix. So us serving that... All that different ways for them to test, I feel like it's critical now. You can't just rely on one type of data.
00:27:19 · Steve Xie, Jonathan Stephens
[00:27:19] S0: I totally agree with you. I totally agree with you. Yeah. The first half of last year, we delivered tens of thousands of data. But the second half of last year, we delivered probably close to 1,000,000 hours of data. And for this year, like, for just, like, this few months, we delivered more than probably 2,000,000 hours data already. Right? So you can see the scale increases exponentially. Right? So I'm really seeing this year probably will deliver more than ten million hours data in total. [00:27:56] S1: I'm seeing more and more projects ingesting that, you know, they say, oh, we have forty four thousand hours of egocentric capture because that's all they could find. [00:28:04] S0: Right? Exactly. [00:28:06] S1: It's not that they don't want more. So if you go read their papers of this research, it'll say in the limitations, a limitation is we only had so much data that we could collect or we could find, or that was high quality. So the thirst for it, because it's, there's like infinite edge cases. We have an infinite complex world you need, You need as much data as possible to have a well informed under, you know, underlying model to do these tasks that we're asking them to do out in the real world. And so that's that's that's what we're doing. For exact structure, scale thing. Right? But not just scale. You're trying to hit quality. Yes. [00:28:42] S0: Because more deeply... Yeah. I I would say the quality and scale are the two most critical and probably the only two most most critical requirements that we need to hit. Yeah. Mhmm.
00:28:42 · Steve Xie, Jonathan Stephens
[00:28:42] S0: Because more deeply... Yeah. I I would say the quality and scale are the two most critical and probably the only two most most critical requirements that we need to hit. Yeah. Mhmm. [00:28:56] S1: Yeah. No. That makes sense. Okay. So then... So we have that we have... We went with the environment, the world, that's that's just your assets. And then we have... We do behaviors and that could be Ego centric, could be a SimTel ops. And then we need test that these models that we're training work. So, tell me about eval. So that's kind of like the, as I, at the start of this video, I did an intro of talking about the concept of those three. Can you break down then eval and why that is critical? And at this point, I don't see a lot of people. I've seen more people talk about it, but not enough people talking about it.
00:29:32 · Steve Xie
[00:29:32] S0: Totally. So I would say, you know, now we have the world. Now we have the behavior, which is basically the data we generate. Right? We are providing, you know, scalable exponential data growth to our customers. Right? And I would say now the bigger problem the industry faces is where we are. Right? Now we have the enormous data. We can train our models. But can we really understand how well our model is? That's the major problem. So that's the question that Evolve is trying to answer, right, to help the industry, to help the customers, to help our partners to understand where their robotics are. Right? What are their capabilities? What are their boundaries? Things like that. Right? So, you know, one of the typical problem of evolve for the industry, you know, is there are so many great open source evolves provided by academia. You know? There are behavior challenge. There is RoboCasa, and there is. They are amazing. You know? All the papers refer them, and they are actually the, like like, I will say the backbone of this commit... This academia for robotics. Because, you know, you have to take a test from Liberia. You have to take a test from RoboCasa. You have to take test from behavior challenge in order to showcase that your research is something providing insight to this academia. But the problem there is they are good for academia. They are not good for industry. The industry, we are running a much larger cluster of compute. Right? Mhmm. Many of our customers have tens of thousands of GPUs, to train their models. Right? The industry is getting much more data than academia. For example, academia probably have thousands, if at most ten thousand hours data. That... That's it. But for industry, you know, they are they are getting hundreds of thousands, if not million hour of data already. Right? Mhmm. And and the may... Many of them are from us. Then the question is, okay, where can I find the eval? Where can I find the benchmark that can help us to understand the robotics capability at this scale? Right? And for the academia benchmarks, you know, the the frontier robotics foundation models, they can take them on pretty well. You know? They are already re... Reaching, say, 99%, a 100% of post test. Right? So I was looking at a leaderboard