Dream Machines: Physical AI, Robotics, Being Human
00:12:36 · Evan Helda, Aravind Kandiah
[00:12:36] S1: That's interesting. Question on that 30%. So how exactly do you guys determine that number and and quantify that? And and how do they... How does the customer or you understand what a 100% should be? [00:12:50] S0: Yeah. So I think the interesting part is we'll never be able to hit a 100% because of the thing, like, you can you can never, like, get all the long tail scenarios. [00:13:02] S1: Yep. [00:13:02] S0: Right? So the way we think about it is there's two types of, like, coverage we call... The first one is... It's quite straightforward. It's just, like, operational conditions. So the way you can think about this is let's take, like, the the boat example. Right? Let's say you're deploying an autonomous boat and you're putting it in a port. There's actually a few key conditions that you care about. So maybe it's sea state, types of boats, different types of weather conditions, different types of, collision approach... Like, approach angles. All of these parameters, you can actually ask any sea captain or any, like, domain expert, and they'll able to list them all out for you. We list them all out, and then this forms like a permutation grid. Right? So you can think like you're doing parameter sweeps, but you're doing parameter sweeps across scenarios. [00:13:54] S1: Mhmm.
00:13:54 · Evan Helda, Aravind Kandiah
[00:13:54] S1: Mhmm. [00:13:55] S0: Right? And this forms, like, this elaborate grid where it covers every single permutation. Right? So that's, like, bucket one. Bucket two is what I call these anomalous long tail scenarios, where it's like a giant whale comes into the scene. A plastic bag is covering a a sea lion, and it's flailing about. All these very weird things. [00:14:20] S1: Mhmm. [00:14:20] S0: These things actually, like, AI agents are really, really good at coming up with these long tail scenarios. And humans, as they see these things come up, they add them to the the long tail bucket of stuff as well. Mhmm. And this is the same approach that Tesla has taken as well. And it's just the thing that you constantly grow this, like, a long tail scenario library. Right? And the interesting thing that we have seen... Sorry. Go ahead. [00:14:44] S1: I was saying that's a really helpful way to break it down. [00:14:47] S0: Yeah. Because this one, the first bucket is really easy to quantify. But the second bucket, which is like all these anomalous long tail things, you can't really quantify because it's an infinite number of scenarios that can happen. Mhmm. Right. [00:15:01] S1: So you used a fancy word earlier that I doubt most of my audience knows, which is parameter sweeps. And since I'm on the mission to help educate people on the technical elements of the space, can you quickly define what parameter sweeps are and why they matter?
00:15:14 · Aravind Kandiah
[00:15:14] S0: Yeah. So the way you can think about parameter sweeps are... Let's just take a super simple scene. Right? Let's say you are outside and you want to be able to just detect a chair. What a parameter sweep would look like is, let's say I want a parameter sweep lighting conditions. It means I want to move the sun angle to all the different parts of the... Like, from sunset to sunrise through the... All the different angles so you see the chair at all the different lighting angles. So that's what it means, like, parameter sweep through the lighting angles. So you could also do a parameter sweep through different types of chairs. Right? So modern mid century chair, an old antique chair, chrome metal chair, and there's, like, parameters sweeping through the object types. And there's also parameters sweeping through the angle in which you're viewing the chair. So you can see how these, like, compound really, really quickly. And the goal really is just, like, can you provide these teams with these high coverage datasets so that they can test and see where their model is failing even before they deploy it? So a lot of the times, they would generate these tests, these parameter sweeps, and then they're like, oh, actually, it fails on a Chrome chair at this specific angle and this specific lighting condition. We didn't know that this would happen, so they'll go in and they'll try to fix it.
00:16:37 · Evan Helda
[00:16:37] S1: Got it. Super, super helpful. So I... We went deep in the technical rabbit hole, which I like doing in the beginning because it gets people hooked on what this is and why it matters, but let's let's pop up a level. Put... Go from CTO perspective to CEO perspective. Charles, take us take us back to the beginning and maybe give us that headline thesis and mission statement you guys had when you started the company, and tell us a of the origin story about how you guys met, how you guys decided to jump into this problem, and kinda what was that moment that this problem is worth solving and that you guys have the right team to go solve it.
00:17:14 · Charles Wong, Evan Helda
[00:17:14] S2: Yeah. For sure. And I'll I'll also maybe answer a little bit about your question your question earlier about, you know, what does it take to convince people that we know what we're doing and, you know, that, you know, seem to real is surmountable, that, you know, fake data... Our synthetic data is as good as real data. Right? So taking us back, you know, I... So me and me and Aravind met in in freshman year of college, actually, we've basically been partners in crime ever since. Mhmm. The... I think we... And, you know, we noticed in each other a bit of an ambition a desire to do something a little more than what was just, you know, in the textbooks. Right? And that I think sparked... That was enough of a delta for us to go and do our own thing. Right? We took slightly different journeys together. Know, for me, I went down the autonomous vehicle route. So I was working at Newtonomy for a while. Actually, I remember my first ever internship at Newtonomy was barely even an internship. I basically got a warm intro to the director of autonomy at the time and just texted him, hey, I don't have any skills. Can I come do something? Can I go work for you? And my jobs there included making Ethernet cables. I can like like, who makes Ethernet cables? Right? I was like, yeah. And also driving their their cars dangerously to... For scenario testing. The... You know, I also worked a little bit at a VC firm in the Bay Area. That was an internship as well. Just trying to get, you know, just trying to see what I like on the other side of the table. [00:18:49] S1: Mhmm.
00:18:49 · Evan Helda, Charles Wong, Aravind Kandiah
[00:18:49] S1: Mhmm. [00:18:50] S2: You know, Aravind, you wanna give a quick intro? [00:18:52] S0: Yeah. Yeah. Charles didn't mention this, but he was also building race cars. [00:18:58] S1: Oh, yeah. That's cool. [00:19:00] S0: He built a race car from scratch. He's he's underselling himself for sure, but he did a lot of cool stuff. He inhaled a lot of carbon fiber as well during that time. So his lifespan is significantly short creativity. [00:19:12] S2: Indeed. Indeed. I will I will outlive Charles. So