The Robotics Stack
00:31:35 · Oskar, Varun Nair
[00:31:35] S0: What are the hardest type of deployments that you see? If you could categorize them a little bit? [00:31:40] S2: I think there's a company called Persona. It's one of the very latest age companies that does robotics welding. I think that's pretty hard for a regular policy to do. I'm not sure how they do it, but welding is definitely on the harder side of things. Anything that's very precise and dexterous is just usually very hard. The more margin of error or what we'd say is the tolerance of a certain task is easier to do. So if I just have to put a box in a package and the package opening is very big and I can just put it anywhere and it just goes in and I just flap it open, that's easy to do. But if I have to put a very precise part, like a screw, into a very small hole and put that in, it's much harder to do because precision is just more important there. And on that spectrum, a lot of tasks sit. The lower the tolerance, the more precision is required, the harder the task is, usually speaking. Or if there is a lot of tool use, that's another place where a lot of models fail today, they're not good at using tools, they're good at using their end effector. They're good at using their hands or their end effector paddle jaw grippers. They're not good at using other tools to make something happen, and I feel like that is definitely a big part of what the frontier will look like when it starts using tools. [00:32:49] S0: Interesting. And simulation specifically, maybe something that changed your mind recently?
00:32:49 · Oskar, Varun Nair
[00:32:49] S0: Interesting. And simulation specifically, maybe something that changed your mind recently? [00:32:53] S2: I think with simulation, the biggest thing is it's not like I've changed my mind on this. It's just a new... Decent learning is you wanna just keep recalibrating your sim. And part of that is recalibrating with new objects, new scenes. But a big part of it also, like, if you can, changing up the physics engine, is really hard, but if you're using Isaac Sim, obviously you have no say in that, but having the recalibration needs to come from scene and asset generation, which is easy to do for anybody using these softwares, but you cannot recalibrate a whole engine. So if there is a way to recalibrate the physics engine, in my opinion, that would be game changer, but I don't know how that is done. So I need to do a little bit more digging and understanding that. There are open source engines that must exist, but SIEM is just such an annoying software to use. It's not very intuitive and it requires a lot of compute and it takes a while to set up. That's just 10% of the work. That's not even the actual work. It's just set up just like getting it to like work well. Imagine changing up the entire physics engine, like I can't imagine that's easy for even a team of people to do. [00:33:57] S0: And did you take a look at what Niantic is doing with their geospatial model? [00:34:01] S2: No. Do you want to tell me?
00:34:01 · Varun Nair, Oskar, Leo
[00:34:01] S2: No. Do you want to tell me? [00:34:02] S0: Yeah, they want to create a AI model that is pretty much grounded in physical data. So they combine all sorts of different data sources. They will also work with Mentor, I believe, from satellite data. It's like a big mess of data that they're trying to bring into this, like, one model, but, it's super interesting. [00:34:20] S2: I actually have heard of the whole Pokemon GO world model, like, project that's happening. [00:34:25] S0: Huge company because basically they worked before on Pokemon GO and these games, and then at some point they just decided, alright, we're not doing games anymore. We're selling our games division, and we're just doing AI and this geospatial model now. They also changed the CEO, so definitely something worth looking into. Nice. [00:34:42] S1: There's been a famous article on X, famous in our space, of course, from Sam Padilla, who was the founder of Avea, a data acquisition company that they closed down after two years. And he basically said that robotics data is a broken business because there are too few buyers and they don't exactly know what they want. So I'm curious, what's your take?
00:35:05 · Varun Nair
[00:35:05] S2: Time is correct. It is true if you are not someone who owns these relationships or a top dog in the field. If you are not one of the bigger bigger players who have these relationships and the understanding of looking at data the right way, it's just a hard business. A lot of people in my circle have tried to become an ego data business and like it's not lucrative at all. I mean, it can be in the short term, but like who are going to sell to? Do they even know what they want? And this kind of goes back to what I was saying earlier about like, researchers are just trying different things with their models to understand what works and what doesn't. They don't know. They're not data scientists. They're not people who understand robotics data. Nobody does. Like, there's the ops muscle in data companies, and then there is folks who understand algorithms really well and are trying to understand data. It's not like there is a recipe that exists. They are extrapolating from LLM successes and other data. So they're coming up with the best strategies, but at the end of the day, don't know if it works for sure either. So it's worth knowing that a lot of things Sam is saying has truth in it, but that doesn't mean data companies are not winning. It's just that the winning companies are already there. They are already figuring this out with the labs. Their job is to demystify the uncertainty. Their job is to build that relationship.
00:36:15 · Leo, Varun Nair
[00:36:15] S1: For sure. What company do you think right now is interesting or is the best place to make it in the current space? But also it made me think of an analysis I read recently saying that even though physical intelligence has a pretty strong team and a pretty strong model, they are going to be overtaken by all the others like skilled generalists, Dyna, because they lack deployment data. [00:36:39] S2: I don't think I'm at liberty to talk about what these companies are doing, but that being said, I do think there are going to be multiple winners. It's not one winner take all. The world of Atoms has always had that. It's a positive... Definitely a positive sum game, and a lot of these companies, if not almost all of these companies will be winning in their own ways, in their own niches. So I wouldn't say that one of them is better than the other. It's a race that keeps everyone accountable and makes everyone work hard to do better. And I'm excited that there are such brilliant researchers and such brilliant, like, minds sort of working on this. And I'm just glad that I can, like, see their thoughts on Twitter and their releases on Twitter to learn what's next. [00:37:18] S1: I see. Any maybe lesser known company that you might be allowed to talk about that you think more people should know about?
00:37:18 · Leo, Varun Nair, Oskar
[00:37:18] S1: I see. Any maybe lesser known company that you might be allowed to talk about that you think more people should know about? [00:37:26] S2: I mean, usually, wouldn't come here and promote any company particularly. That being said, as I said earlier, Robocurve is definitely one of the ones... A project that I really like is that people are trying to do the first, like, humanoid that climbs Everest. So, and I heard about that very recently from a friend. So I'm gonna follow that one. If there's a humanoid climbing Everest, that's going to be amazing. So I'm excited for that. Let's see what happens. It's for next April, whether it happens or not, I don't know, but I'm keeping an eye out for that. [00:37:55] S1: I think it's a guy who has a unit three g one, and he took it already through mountains in Chile, in Patagonia. Right? [00:38:03] S2: Wow. [00:38:03] S1: It is probably the same guy. [00:38:05] S0: Yeah. And and I saw on your website, you also mentioned a couple of investments that you did. One that stuck out to me was, Avea Robotics. I mean, as far as you can talk about it, why did you choose them and what are they doing exactly?