01:32:36 · Adam Miller, Michelle Gilmore
[01:32:36] S5: That that weighs in at a time. Yes. [01:32:38] S1: I just got a good recommendation. Oh, they just so there were attributes that it had to meet. Yeah. And an agent is probably going to be way better at that than the subjective emotional human that I paid a lot of money to do that for me. And so, again, context is going to change that. Whereas if it's my family home that I'm living in with my child and I'm... Then I'm going to want to probably see that. But if it's enterprise software, do I wanna do a handshake? Has or no? Get out? I don't I don't wanna think about that. Whatever. Like, what's... Is it is it a value? Does it meet these attributes? Is it going to give my business a return? Great. Move on. [01:33:21] S5: I I think I'm thinking on, like, the, I don't know, hundreds of thousand dollar plus contracts. Yeah. But I I quite still about it. Okay. So you wouldn't, like, wanna sign off? Like, you'd let an agent make that full decision end to end?
01:33:21 · Adam Miller, Michelle Gilmore, Daniel Palmer
[01:33:21] S5: I I think I'm thinking on, like, the, I don't know, hundreds of thousand dollar plus contracts. Yeah. But I I quite still about it. Okay. So you wouldn't, like, wanna sign off? Like, you'd let an agent make that full decision end to end? [01:33:33] S1: I may want to, over a certain amount, approve it. But once I'm comfortable, if all of the parameters have been met and their evaluation frameworks and I trust that whole system... When I say an agent, I mean all of the frameworks in place and the parameters. Yeah. Why would I... I mean, companies spend six months evaluating the right CRM and spending so much time and money and then use the same crappy one that they always use. And that is such a waste of the people that they're paying a lot of money to do that. And, actually, there's a very few high priority attributes that the piece of software has to achieve. It's not complex software. It's a simple workflow. There's three providers. Let's be honest, it comes down to which one's gonna give the biggest discount. Like, I I just don't think that's where the high leverage thinking needs to be. [01:34:28] S3: And I think, like, part of my point is that it's actually really hard, if not impossible, for you to predict what you actually care about in the future. Like if I take it back to the music analogy, in 1920, could we have predicted that in one hundred years, Palmer would be listening to some experimental, electronic beat boot music that... [01:34:48] S5: I would [01:34:49] S3: have been able [01:34:49] S1: to guess. [01:34:51] S5: This guy. He
01:34:51 · Adam Miller, Xharmagne Carandang, Daniel Palmer, Michelle Gilmore
[01:34:51] S5: This guy. He [01:34:52] S0: doesn't say it's [01:34:52] S3: born in 1928. [01:34:55] S1: Need a more random example. I think you're listening. [01:35:00] S3: That's supposed... [01:35:01] S5: But like, I think it would have been [01:35:02] S3: hard to project what music would have resonated. You can't project, I can't project what music would resonate. What news would be? And so for the same reason, I don't think I can project what will matter transactionally in business in this future either, because the norms will shift, that could just be the thing that we are so used to agents taking care of, that we place that human to human value on some other unpredictable part of business. Maybe it's the morning messages you send to the owners of the other company, that was a terrible example but what I mean is there will be something I'm sure you will care about in that future and you will find that human connection in. I just couldn't tell you what it is and I don't think we can say right now, definitely not that, or definitely not this. [01:35:45] S0: And not just the human connection part, but you described it should meet these criteria. We are assuming that everything you can easily define criteria for. Absolutely. Maybe not. [01:35:55] S1: You won't be able to. Yeah. Because if there's criteria, it means you have to be able to codify, define [01:36:00] S0: Rationally think through it. [01:36:02] S1: Right, actually, most people aren't very good at. [01:36:04] S0: Yeah.
01:36:04 · Xharmagne Carandang, Michelle Gilmore, Adam Miller
[01:36:04] S0: Yeah. [01:36:05] S1: And so I... I'm not sure it's going to be that simple or general. Mhmm. And I think that also where we put our cognitive load will change. So what I'm okay making a decision about today comes down to, you know, what fires are there? What do I have to prioritize? What... If something comes into my world that isn't a priority, I may... I'm just gonna outsource that today, but tomorrow, I may not. Yep. So I I I just think that it's adaptive, and I'm not sure there's going to be anytime soon general rules. Yep. Mhmm. Because one one business owner will want to be across every single decision always and micromanage everything, and another one will be like, I don't need my cognitive load on that. I'm just gonna outsource that. [01:36:53] S5: Mhmm.
01:36:53 · Adam Miller, Michelle Gilmore, Sachin, Daniel Palmer
[01:36:53] S5: Mhmm. [01:36:54] S1: Whether it's to a human or AI workforce or hybrid, I personally think of my workforce as as agents and humans, like, when I talk about it. And some are humans working together, and some are agents of workforces being governed by a human. I think that's something that we're probably seeing changing faster and that we are talking about these hybrid notions. Tasks getting done by groups of agents working together being governed by a human. Mhmm. But to do that well, you need to codify it. You need to understand what you're automating. You need to be able to do quality assurance. You need to be able to think about... You know, there's a whole there's a whole lot of geek fail, obviously, right, that we do every day that is incredibly challenging, which is why I think we're comfortable that that's not happening at scale anytime soon. [01:37:48] S2: I think as we kind of round out this dinner, what could be nice is if everyone around the table says that something that they would want to tell something that themselves in ten years would want to tell themselves now. Because we've been talking about all these timelines and where we're going to go and all these kinds of things. What would your plus ten years self tell to [01:38:10] S1: you now? About the businesses or our work or [01:38:13] S2: just something related to AI? Doctor [01:38:17] S3: What a question. [01:38:23] S2: I'll kick us off and I need [01:38:25] S5: to think of something on the spot now.
01:38:25 · Adam Miller, Sachin
[01:38:25] S5: to think of something on the spot now. [01:38:29] S2: I think I would say to myself that right now, the way AI works, way next token prediction works is very much like, how can we rationally consider the next kind of sequence, right? And I'll tell myself that if that is the way the world is going, pay attention to things outside of that, pay attention to things like emotion, pay attention to things that cannot be defined in that framework, because it's very clear that that's only gonna accelerate and all the money that's going on there is only gonna accelerate. So those are the things that kind of really matter. And I'd also encourage myself to not be too hyper capitalistic about this. I say this as someone that's VC backed, etc. But I do think it's easy to see this as a big financial opportunity, which it is for all of us. As I think someone says, we're probably all going to be fine in that world, right? But there's billions of people out there that don't even have access to food, water and mobile phones, right? Like I would also encourage myself to think about those people because yeah, it really scares me to think about what's going to happen.