etn.
00:18:43 · Sam, James Lowe
[00:18:43] S1: I think following on from that, the conversations we've been having over the past few weeks has been a lot about I think when Anthropic released the S1, there was a little bit about the potential for like the new UBI, the new universal basic income, what that looks like. They obviously released the public benefit fund as well. And then and then SpaceX kind of did some something similar. And that opened the question of like, what is the purpose of human beings? We asked Nick Clegg about this on Wednesday, and he had a really fascinating, fascinating answer. I'd love to ask you the same. Like, what what does purpose look like in, you know, a post AGI world or whenever these companies get so big that they're able to, I guess, share the profits with, you know, a vast amount of people and therefore might not work, you know? [00:19:29] S4: Yeah. So so we don't really subscribe to the belief that there should be some kind of UBI at some point where everybody's just like taking some meager, you know, share of the profits being generated by a few giant companies. In our view, what's very likely to happen and kind of a new business model we're pioneering on our end is to help enterprises post train their open weights models. But whereas they post train and deploy these models in their company Mhmm. We split our margin in half, and we pay that as royalties back to all the employees and all the external experts [00:20:01] S1: No way.
00:20:01 · Sam, James Lowe, Ed Leon Klinger
[00:20:01] S1: No way. [00:20:01] S4: Who provided, you know, data to to train those models. Alright. Now the reason why it's interesting is that, you know, as companies deploy AI more in their, you know, in their work, they're also gonna reward all the employees who had all that institutional expertise encoded in AI. And as a person, what it also means is you can translate your expertise into evaluations and then turn that into a scalable form of your mind. Right? Where everywhere around the world, as AI deploys and as it leverages your expertise, you're gonna earn more in passive income. You can spend that income on all the things you wanna do. And so that's what we believe the world's gonna go to, and it doesn't preclude the proprietary labs. Right? The proprietary labs are also gonna build fantastic products, and they're also gonna do what they... That they're gonna do, and we're gonna work with them to help train their models too. But what we see this sort of, like, duality in the world where we can grow the AI economy and the human economy at the same time. [00:20:48] S1: Interesting. So royalties will be a huge profit generator for the individual, you think? [00:20:52] S3: Yes. [00:20:53] S1: Is there a time in history where that's happened before? All I can think of is like the music industry. Mhmm. Is that a is that a parallel? Can you kind of talk to us a little bit about about that? Like, how did you come up with, hey, like royalties are how, you know, this is gonna work?
00:20:53 · Sam, James Lowe, James Foster
[00:20:53] S1: Is there a time in history where that's happened before? All I can think of is like the music industry. Mhmm. Is that a is that a parallel? Can you kind of talk to us a little bit about about that? Like, how did you come up with, hey, like royalties are how, you know, this is gonna work? [00:21:07] S4: Yeah. So there's a version of this where it's kinda like Spotify for the mind. Yes. So where, you know, essentially... Trademark that. Trademark that. We gotta do this. Brittany's gonna be so happy about this. But Spotify for the mind. And and you can kinda, you know, figure out how do you split royalties at scale. Now the challenge with music specifically is that once you've created a a piece of work, it basically stays static. Like, that's the piece of work. [00:21:31] S7: Of course. [00:21:31] S4: And so as a result, that's always been a challenge where, you know, people go, well, you know, there's a fixed amount of potential revenue and it's not splitting properly. Right? So in this particular case though, once you've encoded your mind as evals, that's not a fixed piece of work. There are many, many ways what you know can be incredibly valuable around the world. [00:21:49] S1: Mhmm.
00:21:49 · Sam, James Lowe, Kosta
[00:21:49] S1: Mhmm. [00:21:49] S4: And part of what AI can do is discover where your expertise could be useful. So some examples we see today, for example, is companies come to us and say, I wanna find customer feedback on my product. And a lot of our users every day talk to our voice agents about, you know, their experiences, how they use their product. Yeah. And so through that we found there are hundreds, sometimes thousands of users who have already used this product before, and we contact them and say, hey, we found that you you've actually used this product, so do you want to now do a $200 per like thirty minutes call to like talk about your product experience and go deeper? And suddenly that that has uncovered new earning experiences for them. [00:22:24] S1: Yeah. Right. [00:22:25] S4: And so really this is about how can AI discover all the ways your knowledge is economically valuable in ways you couldn't have even imagined, and then scale your earnings over time. So it can go way beyond what Spotify on its own can do for music. [00:22:38] S0: Yeah. And maybe maybe one last question before I have our next guest on. Tell us what was the moment internally where you saw this sort of hockey stick? Like, what what... How how are you feeling internally? Because it's a very rare thing to experience. There's probably a handful of companies that have ever experienced it in the whole world. What's going through your mind? What was the moment where you saw like, oh, wow. This thing's, like, scaling almost uncontrollably?
00:22:38 · Kosta, James Lowe, Sam
[00:22:38] S0: Yeah. And maybe maybe one last question before I have our next guest on. Tell us what was the moment internally where you saw this sort of hockey stick? Like, what what... How how are you feeling internally? Because it's a very rare thing to experience. There's probably a handful of companies that have ever experienced it in the whole world. What's going through your mind? What was the moment where you saw like, oh, wow. This thing's, like, scaling almost uncontrollably? [00:23:00] S4: Yeah. Yeah. Honestly, it feels like a blur. Right? So, like, whole team went from... You know, the ramp was literally something like 200, 400 k, 3,000,000, 22,000,000, 27,000,000, you know, 43,000,000 per month. Like, it actually sort of went like that. [00:23:14] S1: Same. [00:23:15] S4: And so, as the team is is dealing with grappling with this, you're literally going every day going, my god. Like, we're fighting another fire, fighting another fire, fighting another fire. This is probably the first one we've we've had to slightly breathe and go, wow, we've gone into a completely different level of stable state. We've built all these systems. For the first time, everybody in the company feels like we could delegate to AI to kind of solve a a lot of these problems. We're building really good systems now to to scale further. And so it felt like a complete blur, and I don't think we we were expecting any of this. And every month, we just have to unblock the most important problem step by step, step by step Yeah. Until we build the systems to to keep scaling from
00:23:15 · James Lowe, Kosta, Sam, Ed Leon Klinger
[00:23:15] S4: And so, as the team is is dealing with grappling with this, you're literally going every day going, my god. Like, we're fighting another fire, fighting another fire, fighting another fire. This is probably the first one we've we've had to slightly breathe and go, wow, we've gone into a completely different level of stable state. We've built all these systems. For the first time, everybody in the company feels like we could delegate to AI to kind of solve a a lot of these problems. We're building really good systems now to to scale further. And so it felt like a complete blur, and I don't think we we were expecting any of this. And every month, we just have to unblock the most important problem step by step, step by step Yeah. Until we build the systems to to keep scaling from [00:23:55] S0: And and that time frame is what? Sorry? [00:23:57] S4: That was all happening in six months. [00:23:58] S1: That's insane. [00:23:59] S0: I don't [00:23:59] S1: know how you I don't know how you're sleeving. We we we have a nice guest coming in. [00:24:04] S3: So thank [00:24:05] S0: you so much. Wonderful as always. [00:24:07] S1: So good to see you. [00:24:07] S4: See you. [00:24:08] S1: See you [00:24:08] S0: soon. Dude, always... We'll see [00:24:09] S4: you later as well. Take care. [00:24:10] S1: Cheers, dude. Get well. Sorry. Are you good? Dude? Okay. There you go. [00:24:13] S0: Think we are ready for our next guest. Sam, do you maybe wanna see or James check who's I think we have one of the [00:24:18] S1: a 16 z partners. Yeah.