The People's AI: The Decentralized AI Podcast
00:28:57 · Bayley Wang
[00:28:57] S1: on on a robot? Remotely. Remotely. Okay. You you remotely operate the robot through like it could be as simple as a keyboard and a joystick or VR or something. Okay. And that that type of that type of task has a more interesting incentive structure right now because there are customers that are willing to like they're paying $50 an hour for that service right now, but they can't scale. So you you... That's like a very robust business. Right? Like they're they're like not only do they not want to pay $50 an hour, but more to the point they do not know how to scale these robot operators. Like the team is like like you you have companies with a valuation of over $1,000,000,000 running a team of eight people to do their data collection right now. Like one of the people is like a founding engineer. That kind of situation. So being able to offer something which is a lot more scalable is really helpful. And there's a lot of interesting things you can do there. Like, we encourage guild formation because guilds cannot only go and like organize the operators on the ground. They can also do things like invest in a robot that can be operated to generate data and mine incentives. And that's really helpful for the robotics companies as well because that means that they can, like, at least break even on the hardware, like, sell it at cost to a guild for the data collection rather than having to, like, raise dilutive capital for a manufacturing round in order to, like, get to the point where they can even start collecting data.
00:30:24 · Jeff Wilser, Bayley Wang
[00:30:24] S0: Amazing. Say a little more about how this works as far as compensation. Is there a token? How is a DAO fit into this? What are kind of the the web three mechanics here? [00:30:35] S1: Yeah. Yeah. Of course of course, there's a token. Right? We're a decentralized project. And what what else would we do? But the the DAOs are sort of their organize... Their organizational units to manage access to the data and, sales to the data, so that we don't have to go manage all of the data ourselves. The way the tokenomics work is that it's it's similar to like any any any other data project I think or what the other data projects should be if they're not already already doing this. That new data contributed to the network generates new tokens because that's like new value being retained by the network. Then the transactions themselves happen through like a a token buy and burn mechanism. And that that's that's where the that's where the DAOs sort of come in as well. Right? Like like, central buy and burn is, somewhat regulatory suspicious in many countries. So by having the DAOs, you sort of you sort of allow that to happen. And it's it's up to the DAOs to go manage the manage the, data access. [00:31:37] S0: What... What's the token called? And where should listeners go to learn more if they wanna get involved with this?
00:31:37 · Jeff Wilser, Bayley Wang
[00:31:37] S0: What... What's the token called? And where should listeners go to learn more if they wanna get involved with this? [00:31:42] S1: Yeah. Yeah. So, our project is called PrismaX. The token will be called p... PIX, p I x, but it's it's it's not live yet. So right now, the best way to follow us is, you can follow us on exit PrismaX AI. We also have a we also have a a sort of portal portal gateway where people can get engaged, upload some existing data, and interact with the platform. [00:32:05] S0: You've been in stealth for a while coming out of stealth. Talk about what that pivot stealth, you know, what you're launching with and then about demo day. I'm eager to hear what you guys cooked up.
00:32:05 · Jeff Wilser, Bayley Wang
[00:32:05] S0: You've been in stealth for a while coming out of stealth. Talk about what that pivot stealth, you know, what you're launching with and then about demo day. I'm eager to hear what you guys cooked up. [00:32:17] S1: Yeah. Yeah. So we've been in stealth for the past year or so, mostly working on the demand side to engage with robotics companies. We're coming out of Stealth, and we want to build out our community of people who are excited to help advance the new frontiers of AI and robotics. Right now, we came out of Stealth. We were part of the a 16 z CXX SF cohort. Close the... We recently closed the seed round and we're coming we're coming out of Stealth with just a basic a basic early portal right now to be able to let users interact with the platform. You can you can upload some of your existing data. You can go out and collect some existing data or soon you'll be able to also, reserve robots and be able to teleoperate them in order to generate new data. And that's an exciting part of the platform as well. There's a lot of data collection projects, but there's no two sided data collection projects yet where you engage with someone else's hardware, someone else's product to collect data for that product. And I think there's a new level of value here which we can really capture. [00:33:22] S0: Dude, I would, like, pay money for funsies to operate a robot remotely. [00:33:28] S1: Exactly. Exactly. Right? [00:33:29] S0: Like... No more is gonna pay me to do that.
00:33:29 · Jeff Wilser, Bayley Wang
[00:33:29] S0: Like... No more is gonna pay me to do that. [00:33:31] S1: That's pretty cool. People will go to like a theme park and like pay $15 for the special ride where they go into VR and like become the robot. And I'm telling you right now, there are people who will pay you $15 for you to operate that robot. Because right now, like, whatever, the CTO has to go operate the robot himself to collect the data, and he really should not be doing that. [00:33:53] S0: Well, so tell me a little more about about these comp... So you're saying there... This actually exists right now. There are company... There are robotics companies. You you said earlier, billion dollar robotics companies. They're like, shit. We have these robots. We need remote workers to come in with a joystick or keyboard and move this robot around the factory. That is [00:34:16] S1: currently... Exactly. Exactly. Like as like probably right now at this very moment as I... We make this podcast. Yeah. There are robots being remotely operated to collect data by companies that have raised collectively billions of dollars, but they do not know how to deal with the operator fleets. Like, it just doesn't make sense for a highly technical team of researchers to also handle what eventually needs to be thousands of data collectors. So there's like an incredible opportunity here for decentralization. Right? Like like the platform, the de facto platform for these operators who can immediately engage with an industry on something that they need like today, right now, but don't know how to do.
00:34:16 · Bayley Wang, Jeff Wilser
[00:34:16] S1: currently... Exactly. Exactly. Like as like probably right now at this very moment as I... We make this podcast. Yeah. There are robots being remotely operated to collect data by companies that have raised collectively billions of dollars, but they do not know how to deal with the operator fleets. Like, it just doesn't make sense for a highly technical team of researchers to also handle what eventually needs to be thousands of data collectors. So there's like an incredible opportunity here for decentralization. Right? Like like the platform, the de facto platform for these operators who can immediately engage with an industry on something that they need like today, right now, but don't know how to do. [00:34:55] S0: So the PrismaX platform, I get at a high level. It connects users who want to... Let's say, in this example, they want to get paid $15 an hour or whatever it is to use a keyboard and move the robot around. Again, like, very cool. Do you guys provide the actual, like, operating system, the UI to do that, or is it more giving the token in the future to user and then the robotics company is, like, has their own UI? How how does mechanics work there?