Venture with Grace
00:47:42 · Grace Gong
[00:47:42] S0: I feel like you mentioned how much is the juice, is it even worse to squeeze? I feel like that can be applied for the physical AI space since I feel like, well, I don't know if it can be applied. When you say we're at the top of the wave or because I feel like nowadays a lot of the software VCs are investing in hardware because of the margins are kind of similar nowadays. And when you're thinking about these big investment themes, like autonomous vehicles, or robotics, or defense autonomy, how much would you think it will be keep on growing versus would you think there will be only a few huge company that end up in the space that's dominating?
00:48:33 · Austin Spiegel
[00:48:33] S1: Yeah, obviously I'm very big booster and bullish on this industry and the number of companies operating in this space continuing to expand. I think there's a couple of really interesting things, right? One is, will we actually be able to re industrialize in The US? And also, what is the impetus to re industrialize? I believe we will be able to, and I believe we're going to be forced to. I think that, you know, look, we're not, right now, we're not importing Chinese automotive vehicles. I was actually talking to, you know, one of, one of my colleagues and we were talking about how, even the Chinese humanoid robotics space is actually slowing down because it's it's taking jobs away from people who are otherwise unemployed in China. So, I actually think, we know we actually have a labor shortage here. You know, we're not in the same situation as China. So that means I think if it's not going to be China that's building autonomous robot, like autonomous humanoid robots and exporting to the rest of the world, someone has to do it. I think The US is very clearly the the leader to do it. So I think there's actually a lot more opportunity here. With robotics specifically, I think we're just super early. Like, mean, most of the businesses that are working in the robotics space are... There's a lot of picks and shovels businesses, right? There's a lot of model businesses. And then there's a few companies building robots and it's easier to build a single prototype, it's a lot harder to scale up. And that's what they're going have to go figure out.
00:50:12 · Grace Gong, Austin Spiegel
[00:50:12] S0: What would you say is misconception of people in the physical AI industry that normal people would have? [00:50:22] S1: Oh, so misconception of people in the physical AI industry that [00:50:29] S0: Like all your clients. Or I guess when you're working with all these kind of... Obviously you work with a lot hardware companies as well as all the cool space, defense, all the the interesting companies. There something that you feel like the general public's misunderstood of the space? [00:50:52] S1: Yeah. I I think certainly talking about physical AI is, I mean, this is not the general public. I'd say this is more like the Silicon Valley crowd, but AI adoption is actually very slow in a lot of these critical industries and businesses. And a lot of that has to do with trust, right? Like the cost of failure is so high in these industries. If you're launching a multi million dollar satellite, if that satellite doesn't work when it gets to orbit, then you can't do anything about it. You can't go up and fix it, right? So, you deploy software that bricks the satellite, you can't roll it back. So, I think because of that, non deterministic AI is not or there's a lot of concern about adoption and you basically just need to kind of tailor the work. First off, don't force it on people, right? And then you need to kind tailor the workflows and improve the eval so it will actually be more widely adopted by customers.
00:50:52 · Austin Spiegel, Grace Gong
[00:50:52] S1: Yeah. I I think certainly talking about physical AI is, I mean, this is not the general public. I'd say this is more like the Silicon Valley crowd, but AI adoption is actually very slow in a lot of these critical industries and businesses. And a lot of that has to do with trust, right? Like the cost of failure is so high in these industries. If you're launching a multi million dollar satellite, if that satellite doesn't work when it gets to orbit, then you can't do anything about it. You can't go up and fix it, right? So, you deploy software that bricks the satellite, you can't roll it back. So, I think because of that, non deterministic AI is not or there's a lot of concern about adoption and you basically just need to kind of tailor the work. First off, don't force it on people, right? And then you need to kind tailor the workflows and improve the eval so it will actually be more widely adopted by customers. [00:51:55] S0: I agree. I feel like AI is very... I feel like nowadays everyone's building their own bot, on the other hand, do we need the same bot for everyone's building for the same company? I feel like it's super interesting to watch. I wanna be mindful of time. Maybe we do like a quick fat round for you. What's your favorite book?
00:51:55 · Grace Gong, Austin Spiegel
[00:51:55] S0: I agree. I feel like AI is very... I feel like nowadays everyone's building their own bot, on the other hand, do we need the same bot for everyone's building for the same company? I feel like it's super interesting to watch. I wanna be mindful of time. Maybe we do like a quick fat round for you. What's your favorite book? [00:52:15] S1: My favorite book? Oh, man. I don't really know if I have a favorite book. I mean, I I think, like, the books... I read a couple books this year. One is called Breakneck. It's about kind of, like, Chinese society versus American society and how The US is very lawyerly and the Chinese society is led by engineers and, you know, what that means for how China can innovate. And then I read another book about called Apple in China about, you know, China basically or Apple moving to China to do manufacturing and then kind of building up this community of practice in Shenzhen. And I think, you know, what I really want to see here in Southern California is that we have that community practice, and I want that community practice to grow, and I want this to become the hub of next generation hardware development and manufacturing in The US. And I'm hoping that we can, not to get political, but find a way to make that the case in California. So, those were both very interesting to read and kind of helped me think about what do we need to do in order to actually re industrialize. [00:53:31] S0: Who made the biggest impact in your career?
00:53:31 · Grace Gong, Austin Spiegel
[00:53:31] S0: Who made the biggest impact in your career? [00:53:36] S1: Oh, god. These are hard questions. I think, you know, there's several people along the way who have made huge impacts in my career. I would say actually, like, you know, my cofounder, Karthik, he was the one who said we should go found this business. And I was like, I think there's something here, but I'm not entirely sure yet. And he was like, no, absolutely. Let's go do it. And, you know, I wouldn't be on this journey without him. So certainly very influential. [00:54:09] S0: Who would you invite to a dinner party? [00:54:12] S1: Who would I invite to a dinner party? Honestly, like, I'm not a huge, like, oh, you know, Sam Altman and Elon Musk kinda kinda group. Like, I'm... And I do this frequently is like I like to do dinner parties with our customers. I know that's super lame answer, but I'm more interested in understanding on the ground what's happening and how are people building their hardware, leveraging our tools, leveraging other tools. So in the context of work, that's who I would invite to a dinner party, but would have to think about it more. [00:54:48] S0: Where can we find you outside of work? [00:54:52] S1: Like physically outside of work? [00:54:55] S0: Yes. [00:54:57] S1: Or you're like, oh, you can find me on LinkedIn and Twitter. [00:55:00] S0: Well, yeah. [00:55:02] S1: Yeah. No. I mean, like Yeah. Sorry. Go ahead. Oh, I [00:55:06] S0: was gonna say actual outside of work if you have outside of work.