00:15:32 · Michelle Gilmore, Sachin, Adam Miller, Xharmagne Carandang
[00:15:32] S1: I I think from an analogy point of view, this one makes me twitch a little bit because, if we transfer it into the application layer, what we're saying is that supplier A and B will both be able to just turn on a line bulb. Right? So AI is way more complex. But if we just do use that analogy for a moment, I think that it being available to anyone and becoming cheaper and a commodity, you know, if that... If we play with those notions, I think that's really interesting. And we're trying to understand what pricing looks like. You know, we don't have playbooks like traditional SaaS companies did. Right? I mean, I sometimes drool over the simplicity of a traditional SaaS company. Like, wow. That... I'm sure it was hard at the time. But as you said, this intensity, I think that there is no playbook, which is incredibly inspiring and exciting, but also we're trying to play with notions of will quality be the thing that determines a price continuum? You know, do I pay for higher quality, which means I am going to choose a different provider or providers? And in what context does quality matter and in what context does it not? [00:16:54] S2: I was thinking in what timeline? It's like sometimes I get myself in loops where I'm like, Oh, but we'll get to AGI of five years, and then I'm trying to figure out pricing in Asia, then I step back and I'm like, Oh, wait, this is the least of my worries. [00:17:08] S5: Yes. There needs to be some sort [00:17:09] S1: of discipline around how... When we go there and when we don't. But they are connected, [00:17:14] S0: I think.
00:17:14 · Xharmagne Carandang, Michelle Gilmore, Adam Miller, Sachin
[00:17:14] S0: I think. [00:17:14] S1: Right? And you do have to think in those concentric circles. At Juno, we are making bets on quality. So having done the job that Juno does for fifteen years, I'm fairly certain that I can make better subjective decisions on what good looks like in this context of interviewing humans. So doing user research, doing market research, having done that all over the world, I'm confident that I will make a better decision on that than someone that's got much more money than us, to your point, Xharmagne. I think that was a really interesting point. But how do we educate a market on what quality looks like in this context, and how do we apply values to that, and how do we do that really quickly? Right, so that we can stay alive and generate revenue and increase our valuation and keep growing. [00:18:09] S0: Yeah. Because what I like about this is that, say, it becomes a commodity, then we can focus on what actually matters, which is what is the customer trying to get out of it rather than we charge by tokens, whatever. It's like we charge, say at Lorikeet, if we answer the customer's question in a quality matter [00:18:27] S5: Yeah. [00:18:27] S0: Kind of thing. And that means our incentives start to align with our customers. I think that's great. [00:18:32] S2: Yeah. The SaaS incentives are just so you don't rip them out. [00:18:35] S0: Basically. Yeah.
00:18:35 · Xharmagne Carandang, Adam Miller, Georgie Healy, Michelle Gilmore
[00:18:35] S0: Basically. Yeah. [00:18:36] S5: And I'd love to hear, like, what are the businesses that won't thrive if the cost of intelligence does trend towards zero? Mhmm. Like, maybe, Georgie, you might have some opinions because you see a lot of startups in Google and your former VC, this is gonna have a lot of implications that might look like great businesses today, but in a few years it's like anybody can build that with your swarm of agents like you have from Melbourne. [00:18:57] S4: I Yeah. Often ask myself, is this winner takes all? Like, will I just be a few key players? Is it like a commodity like oil where it's a very capitalist market and then, you know, the the biggest will win? I hope not from an innovation perspective. I hope that's not the case. I know I work at Google. But to be honest, like just zooming back a little bit and talking about the electricity question, I do wonder how much you guys as founders think about electricity versus the cost of chips. Cost of chips and GPUs maybe, what, 50% of the total cost? But everyone keeps talking about electricity 15% of the cost? So I'm curious why people keep talking about electricity. Maybe I'm... [00:19:49] S1: Sam Altman said it. [00:19:50] S4: Okay. All right. Now I know. Yeah. Yeah, he's got a notebook.
00:19:50 · Georgie Healy, Michelle Gilmore, Sachin
[00:19:50] S4: Okay. All right. Now I know. Yeah. Yeah, he's got a notebook. [00:19:55] S1: I think we, you know, we grab hold of what very few people say because the media sends that out everywhere. Right? So I think I think there are probably more useful analogies Yeah. If we're talking about something becoming available to everyone to fuel or power various use cases that humans and societies deem worth paying for. Like, that's kind of the most general... I've been deliberately abstract invade rights to [00:20:31] S2: try to [00:20:32] S1: bring us there. Yep. And I think there's probably useful analogies in there, but the reason why we're probably grasping for one is because we are in an unprecedented time. Like, it's it's kind of closer to the industrial revolution than it is to a .com boom in that we're talking about complete workflows, crafts Productivity. Being duplicated, augmented, in some cases, replaced, and also, in some cases, amplified, which is what is unique about AI versus the industrial revolution. Mhmm. Right? There was a very reductionist outcome to the industrial revolution, whereas in some cases... I mean, we've got some customers using Juno that are selling way more interviews than they ever were before. Right? Because they're able to do more, so they're charging more units. Yep. And therefore, they're going through more units because Juno sells the unit of an interview. So it's it's... There... That's a really interesting comparison, I think, if we think about we just need to go back a bit further.
00:20:32 · Michelle Gilmore, Daniel Palmer
[00:20:32] S1: bring us there. Yep. And I think there's probably useful analogies in there, but the reason why we're probably grasping for one is because we are in an unprecedented time. Like, it's it's kind of closer to the industrial revolution than it is to a .com boom in that we're talking about complete workflows, crafts Productivity. Being duplicated, augmented, in some cases, replaced, and also, in some cases, amplified, which is what is unique about AI versus the industrial revolution. Mhmm. Right? There was a very reductionist outcome to the industrial revolution, whereas in some cases... I mean, we've got some customers using Juno that are selling way more interviews than they ever were before. Right? Because they're able to do more, so they're charging more units. Yep. And therefore, they're going through more units because Juno sells the unit of an interview. So it's it's... There... That's a really interesting comparison, I think, if we think about we just need to go back a bit further. [00:21:35] S3: I think it's a potentially fraught question, but I guess why wouldn't it have the same outcome as the Industrial Revolution? Is it because AI is more accessible?
00:21:35 · Daniel Palmer, Michelle Gilmore
[00:21:35] S3: I think it's a potentially fraught question, but I guess why wouldn't it have the same outcome as the Industrial Revolution? Is it because AI is more accessible? [00:21:47] S1: Yeah. Really good question. So the industrial revolution put in physical machines to take out human labor. Yep. Right? And the whole idea was to create everything... To to scale the creation of physical things really, and to have humans in that loop but for humans not to be actually making the things. It was the birth of craft though Right? So there was this idea of after that, people then did go and continue to make the thing they were making by hand, but they were standing up their own markets or they were creating what then became things like Etsy. Right? So, it's not that nothing good came from that. But at the time, those workers that were replaced, they didn't necessarily have an ability to upskill or to amplify. It was just about replacing a human production line with a machine led production line. [00:22:47] S3: Yeah. Mean, like, I think I agree very strongly with that comparison, but I guess, like, I'm curious to dig into more of like, is the difference... Because as you say, we're seeing a similar thing here. Like, I mean, it's happening fast and slow, but there are jobs that are now being taken by agent and sync automation, AI enhanced workflows. Why will the people displaced in that process be more capable of being enhanced by the technology?