The People's AI: The Decentralized AI Podcast
00:36:50 · Jeff Wilser, Bayley Wang
[00:36:50] S0: Got it. And congrats again on your recent demo day with a 16 z. Talk a bit about what you you unveiled and showcased there. [00:36:59] S1: Yeah. Yeah. The a 16 z's demo day was a pretty short like three minute presentation. But against that backdrop, we came out of Stealth on X. We unveiled the we unveiled the portal so that people can start engaging with the platform if they're excited about it. And we're, like, we're we're sort of, like, unveiling this plan for people, like, ordinary people to be able to engage with what is an emerging multibillion dollar industry. Right now, it's only a multibillion dollar industry, but it's going to be hundreds of billions of dollars down the line, right, potentially, as they say, like, replacing all of the, like, replacing all of the, like, labor in the world. Let... Let's go there now. [00:37:36] S0: Well, let's pivot to the future. What what excites you most about robotics use cases in the future?
00:37:43 · Bayley Wang
[00:37:43] S1: Yeah. I think robotic... There there will be more and more robots and the world... The way the world is going to shape up, I don't think people are going to get replaced by robots. I think people will work sort of like work as the back end of robots to do things like make sure robots complete tasks reliably, do teleoperation. So it's not like people will no longer have jobs. It's that one person will be able to do the work of seven or 10 people and that means a heck of heck of a lot more time to either do more or like make money doing a do... Make the same money doing less and then like having more time to go like do the things that people actually find enjoyable. As far as concrete use cases go, I I really would... I I really excited to see robots step in into tasks like, I don't know, kitchens and restaurants are I think are are I think a big one. Food is important to everyone. I'm sure, like everyone needs to eat. And I think that with technology, there's a lot more there's a lot more opportunities to, like, branch out and do something different. If you have more automation, that means your restaurants can, like... It's it's not just about being cheaper. Right? It means, your restaurants can be more diverse, serve more interesting dishes. The people have... The humans have more time to be creative, and like leave the actual repetitive work to the robots. I I would say among the near term use cases that's probably the most interesting one. Like you'll see it, you'll feel it, and your life will become better and more exciting because of this technology shift. As far as the long term goes, the end game vision for everyone, I think, is a robot in every home where you can wake up and, like, all of your chores have been done for you by the robot. But this this this is probably the most challenging of robotics use cases, and I think it'll take a... It'll take us a little bit of time to get there.
00:39:30 · Jeff Wilser
[00:39:30] S0: What... And and I fully agree with you on so much of that, especially the idea that there are some really, obvious wildly useful applications like cleaning. Like, as I've had other guests in my podcast say, hey, you're telling me a robot can come and, like, clean my toilet. Like, yes, please. Who's not going to want that? Who's like, no. It's important to me that I clean my toilet the old fashioned human way. Right? So if if if keep enough, of course, we'll have robots scrubbing, mopping floors, cleaning toilets, and so on. What do you think is the biggest obstacle to overcome to get to a world where, let's call it, the toilet cleaner robot is mainstream? What what would it take for that to be ubiquitous in daily life?
00:40:19 · Bayley Wang, Jeff Wilser
[00:40:19] S1: The answer is very simple. In order for the toilet cleaning robot to be mainstream, we need a heck of a lot more data, because that robot needs to be able to get to your door. That's data. It needs to be able to navigate through your house, find the toilet, and like know how to clean the toilet. And there's a lot of diversity right there. Every house is different, doors are different, layouts are different, and that robot needs to be able to handle all of this. And I could probably right now build you a machine which using conventional machine vision, like finds your house, opens the door, and then like navigates to the toilet and cleans just the toilet, but it would be pretty useless. It wouldn't do anything but clean toilets. So in order to have, like, enough value to ship, that machine needs to be able to do a lot of things. Right. And for that, we need to scale up. And I think there will be a bridge. I think there will be a bridge where we go from videos of people cleaning toilets to teleoperated robots cleaning toilets all the way to autonomous robots cleaning toilets. [00:41:20] S0: What's your prediction of how long that will take?
00:41:20 · Jeff Wilser, Bayley Wang
[00:41:20] S0: What's your prediction of how long that will take? [00:41:25] S1: I I used to say I used to say ten years, but one of my friends who's good at robotics as well, told me that that was too early of an estimate. So these days, I avoid make... I I I avoid giving a hard estimate. But I think it's gonna be about a decade before the technology really gets to the... The technology and the models get to the point where we can we can handle general domestic tasks. It may be less for some of the high value ones like cleaning, but we'll leave that for the we'll leave that for the founders to go find those tasks. [00:42:01] S0: Okay. Last question for you. I I won't pin you on a hard timeline, but give us one other prediction of out there in the future of how our... You mentioned at a high level earlier that we'll have more time to do things we actually enjoy. Give us a little more concrete prediction of how our lives and how the world could be different in the, let's call it, five plus years in the future. What might surprise people of how basic life looks?
00:42:28 · Bayley Wang
[00:42:28] S1: I I think what's interesting is that five years in the future, the robots will silently, like, be part of your life. And you won't you won't really notice it, but you'll find that your life has just gotten better. Right? Like, you'll you'll be able to like go to a restaurant and things will be much more timely because there are robots working in the back to make sure that like things are getting done on time. You'll be able to go to a hotel and things will be cleaner and people like people will be working to engage with you and give you a good experience rather than going around and like folding the sheets and doing the laundry because there will be robots handling that. But they won't be they won't be highly prominent in five years. I think that this this idea where like robots look like people and walk around and just see them on the street, that's a little ways off. It's sort of like the early days of of AI, right? Like we got cool features like better Google Translate and and like Google Reverse Image Search. And that was that was all AI. Those were all based on like embeddings and transformers. But they weren't like in your face. Like, they they were just silently making the world better. And the the the big moment like was ChatGPT where you like explicitly interacted with the AI. And I I think maybe someday you'll explicitly interact with the robot, but I'm also not sure that this is the world where you want. Right? Like where the robots are like squarely in your face all over the place. Maybe what you want is a tool that like silently works in the background to make your make your life better.