Eric Siu explores how AI agents could create fully automated firms by copying leaders and scaling them with compute power. He covers implications for organizational structure, startup opportunities, and a shift toward more creators and entrepreneurs as traditional work becomes optional.
Chapters
Imagine a world where companies are no longer run by people, but they're run by AIs instead. And not just any AIs, super intelligent AIs. In fact, they could be copies of your CEOs right now. So today we're gonna talk about the concept of fully automated firms, and we're gonna talk about how the concept of intelligence is actually gonna shift over into something else. These AI firms aren't going to just be emulations of humans. They are descendants of large language models. And if you think about it, if what we've been paying as humans for work For the last thousands and thousands of years or so, we've been paying for intelligence. Well, the cost of intelligence is coming down. And one of the biggest implications of that, this means we're going to be able to make copies of ourselves as people and as leaders. Let me give you an example. If I wanted to make a copy of an operator right now, which is a ChatGPT operator, that's one of the agents, or I wanted to make a copy of myself as CEO, I could do that. And so I could have 1,000 of myself running in different parts of the company, solving 1,000 different problems. Or you can have 1,000 Elon Musks. You look at all the big problems that he solves around rockets, around global satellites. We're talking about making the government more efficient. We're talking about restoring free speech, all these things. Imagine if he's working, someone like that is working on the world's biggest problems. You can actually make copies of yourself just by doing it. The problem with making copies of yourself and the problem with humans in general is not that there's a problem with humans necessarily, but there's just, we can't do this physically right now. Even if we made a copy of ourselves, if I'm working in 1,000 different parts of the world, it's hard for me to knowledge transfer very quickly as a human being. That is actually changing now with the ability to be able to merge merge all of ourselves. Let's make 1,000 copies of Elon Musk and then let's have all of those copies talk to each other. They'll merge all the data. Whatever one copy sees, all of the other 999 copies, they see the same thing too. So knowledge transfer is happening a lot faster. That ultimately means you're going to be able to move a lot faster. Now imagine if you have a mega Elon, a mega Jeff Bezos, a mega Sundar, a mega Satya Nadella from Microsoft. You have one superstar AI that is kind of like the main manager, manages all of the other 999 copies. The only thing it's limited by is compute. Because compute, at the end of the day, why everyone's paying so much, you hear Microsoft investing $80 billion in CapEx for AI, Amazon or Google, they're putting, I think Google's putting $75 billion or so. So they're investing a lot of money into this. They're building out infrastructure because they know compute, that's where the intelligence comes from at the end of the day. So the cost of intelligence is coming down, but more and more of these companies are paying for where they know intelligence is coming from. It's going to be compute. Now, how do you think decisions will flow if it's all coming from one single point in the company? You don't have multiple different meetings. You don't have different emotions from different stakeholders. You don't have people protecting their turf. You don't have people playing politics. At the end of the day, if there's one main AI master that's overseeing everything else, and that AI master is overseen by the human master, things are going to be a lot more coherent. And people like to talk about vectors, for example. And I think this was actually Elon— an Elon analogy that was explained by Dharmesh Shah, from the, the CTO of HubSpot. When you have vectors in a company, if you're all rowing in the same direction, I'm moving in the same direction, you're good. But if one person's going this way, the other person's going going this way, you're actually not moving forward, right? And in an organization, you actually have people moving in all different directions, and it becomes hard to kind of track where things are going. And this is where bureaucracy kicks in, things slow down. But when you make 1,000 copies of the best of the best, and then you merge them together, and you have unlimited compute, or maybe you have a lot of compute, what's going to happen with your business? It's going to move a lot faster. We're not just talking about speed here, but we are also talking about transaction costs too. AI firms are ultimately going to be able to move a lot more efficiently when it comes to innovating. And their transaction costs are gonna come down. If you think about it, cost of knowledge is coming down, decision-making is moving a lot faster. It's a net win for the AI firm. Now let's talk about where this is all going to happen. The big firms, if the Microsoft, the Googles of the world, they're gonna continue to get bigger. But if you're starting up a company, now that you can spin up instances of yourself anywhere in the world, a lot more power is going to go to startups as well. If you can spin up 100 copies, 1,000 copies of yourself, you're just limited by compute at the end of the day. You're gonna be able to move a lot faster. And what's gonna happen is we're gonna see a lot more small teams, big businesses. And what that means is you might have one-person companies doing $100 million a year, which is fantastic. You might have 10 or 100-people companies doing a billion plus in revenue. So I think what's going to happen is you're going to see a lot more efficiency, not just from the big companies, but you're going to see a lot more efficiencies from the world because everyone's going to realize once they have all this power, these geniuses in their pockets or on their computers, they're going to realize that life was never about retirement. Life was more so about creating something and being useful the world. Once they realize that, okay, sitting on a beach for 2, 3 weeks or so is not what it's all about, it's actually about doing things that you love and being helpful to other people. And now that they have the resources to be helpful to other people, everyone's gonna realize that. Because I can tell you, the one secret that most artists or entrepreneurs know is that at the end of the day, it was always about finding what you loved and just continuing to do it for a very long time. So if labor is so easy to replicate in the future, what does that mean for the future of work? What I just mentioned earlier, I think a lot more people are going to be finding what they want to be doing. They're gonna find their calling. And we're gonna see a lot more creators. We're gonna see a lot more entrepreneurs in the world. Those people that don't wanna work, they're not gonna have to work. Work will become more of an optional thing. And that's gonna happen, I believe, in the next decade or so. Again, but I think that's very exciting because for someone like me, it took me a long time to realize I was lost growing up. I played a lot of video games and that was my creative outlet. And I was always worried that I wouldn't be able to find my creative outlet. Thankfully, it turned out to be marketing and it turned out to be business. I think that's gonna happen for the vast majority of people. If a lot of firms are just copies of each other and the ability to just replicate talent is not going to be a challenge, that opens humanity to be able to solve a lot more things a lot quicker. So one of the key points we talked about was copying. So you're gonna be able to copy talent, okay? So you can make copies of yourself 1,000+ times or so, and you're gonna have a bunch of agents, which is a lot of people have been talking about that. You're gonna be able to merge. All of these agents are going to be able to knowledge transfer instantly. And then it comes down to scaling. Do you have the resources to go get additional compute so you can continue to scale the copies in your organization. Because now labor isn't what it's all about anymore. It's more about just how much compute you have. If you're in B2B, here's the deal. LinkedIn account-based marketing sucks, but it doesn't have to. With Carrot, it allows you to create personalized ads towards your customers and also personalized landing pages that align with the ads. So you're going to have higher conversions and influence pipeline reporting to Prove that your ABM works. So if you want to learn more, just go to karrot.ai and you can join the waitlist. Again, it's karrot, K-A-R-R-O-T, .ai, and we'll see you on the other side. Here, one of the implications, if you think about it, that means that humans tend to look at micromanagement as a bad thing. But when you have this ability to copy, to merge, to scale, you can micromanage down to the smallest thing. And I actually think that's a beneficial thing because at least for right now, it doesn't seem like computers mind being micromanaged. Being told what to do, whereas humans do, because humans are very emotional. That means you can get down to the detail. You can do things at a much higher scale while maintaining quality. Whereas if, let's say, you had an agency type of business like my agency Single Grain, what happens is as you continue to grow, if you grow rapidly, what happens is your quality starts to go down. That's not going to happen anymore as long as you have the ability to do good work and as long as you have resources to continue to buy compute. If compute is what it's all about, what does that mean for human beings? Well, for one, this is one thing I've been telling all of my employees at Single Grain. I've been warning over and over. I should Show the new things, but I also warn about the world that we're going into. The best of the best will continue to have jobs. They're just gonna be paid more. And I believe a lot of firms will have less humans. Those humans will be paid more 'cause they're doing really valuable work and they're very AI-enabled. And those are the ones that bring strategic thoughts. Those are the ones that are just slightly ahead of the curve. And we need to continue to feed strategic thoughts into AI, at least for right now. And that's where the value comes from. And other than that, we're just buying more compute. We're buying more compute. Okay, that's intelligence. We're buying the smartest of the smartest people. And then from there, other people, it's optional for them if they wanna work or not. So I don't think they're gonna just get completely eliminated. I think work will become optional. I think there's going to be so many resources where there's going to be some form of income for everyone. And I think this is not going to be some universal basic income where it's like a small income. I mean like actual universal income, universal wealth, let's call it that. Universal wealth, I think, will become available. And those that want to work, they can continue to work. If the bosses of those companies, those entrepreneurs, they can paint a good vision, they can motivate people, and they can inspire people to come work, then great. And I think a lot of entrepreneurs can do that. I don't think that's ever going to change. I think humans are going to continue to work. And I think Just more humans are going to do what they want to be doing. So let's talk about the forces at play when it comes to these AI firms. I believe that if you don't adapt and move into this new paradigm that we're talking about today, your company is going to be in trouble. And we can see that happening for a lot of companies. The entrepreneurs that are at the forefront of this, they're going to be fine. If you continue to learn, you're a lifelong learner, you're good to go. But I think a lot of the ones that fail will probably be acquired for pennies on the dollar. And then the entrepreneurs that are savvy will just AI enable those companies. And I hate to say this, but there are companies I know right now, they have AI committees where the people in that committee are bonus on the people that they fire. Or I know people where they're losing standard attrition every year. Their standard attrition with people, people will leave. They're just not replacing those people. And the same thing's happening for us too. I'll give you one small example. We had a copywriter say, hey, he gave us notice, it's gonna be done in March. I said, hey, your work, we don't really need to replace you, huh? And he's like, yeah, you don't really need to. And so just with that example alone, imagine that scaling across a ton of different organizations right now. It's the ones that don't want to accept this right now. They think this is overblown, this is fearmongering. You're better off looking at this from a position of, hey, I better do something or I could be in trouble, because if something does happen, you're prepared for it. If you don't do anything, you're gonna end up being screwed, right? So you better do something about it. We went through this video. I want you to watch my next video over here on how service businesses are going to change forever. It is almost a continuation of this video, so go check it out somewhere over here. Click it, and we'll see you in that video.
