What Do Jobs and Money Look Like in a Post-Human Economy?

We need to cultivate young people who understand in the end that what is really valuable about us can never be touched by machines.

Jobs are a construct of a particular version of an economy, by the way, a capitalist economy in which there are these corporations and you work for them, right?

It's already something that's started going away.

Hi everyone, welcome to Unchained, your no hype free source for all things crypto. I'm your host Laura Shinn. Thanks for joining this live stream.

Before we get started, a quick reminder. Nothing you hear on Unchained is investment advice. This show is for informational and entertainment purposes only, and my guest and I may hold assets discussed in the show. For more disclosures, visit unchained crypto.com.

The energy network is an intelligent decentralized grid that coordinates smart devices to balance supply and demand. Energy dollar is the native token of the network from one of Europe's fastest growing energy startups. Follow at fuse energy on X to find out more.

If crypto taxes feel overwhelming, you're not alone. That's why Cryptotax Girl, a team that's been helping crypto investors since 2017, is offering $100 off on one-on-one crypto tax help. To get $100 off your crypto tax services, go to cryptotaxgirl.com/unchained. Again, that's cryptotaxgirl.com/unchained.

Today's topic is crypto and AI plus OpenClaw Boltbook and how all of this impacts the economy, work, crypto, and the future of money. Here to discuss are Michael Casey, chairman of the advanced AI society, and David Mattton, co-founder of the Exponentialist. Welcome, Michael and David.

Hi, Laura.

Hello. Thanks for having us.

So, this past week in AI has been absolutely riveting. It honestly reminded me of a week in crypto where just all of a sudden there's event after event after event.

And it started with Claudebot whose name has become OpenClaw now. And that is a personal AI agent. It runs directly on your own operating system.

People have been putting on their own Mac minis to be separate from their normal computers, although of course that's not too much of a security measure, but that's a topic for another day.

So, OpenClaw can use applications to do things. It can handle your emails, your calendar. It can transact with other online services. It has a persistent memory which allows it to kind of tailor its actions more toward your preferences.

And then the OpenClaw AI agents basically got their own social networking site, Maltbook, which is kind of like Reddit but for AIA agents. And some of the agents have even been launching their own crypto tokens.

So, the last development that I noticed is that there's a website, rent a human.ai, which is for agents to rent humans to do physical tasks in our meatspace world. And basically, there's just been so much that happened.

So, I was just curious to hear which either incidents or developments you found most interesting with everything that happened recently.

And I think it seems unbelievably fast what's happening and these bots going wild and talking to each other and establishing their own religion and everything else, right?

It just seems like it's getting way out of control.

But we've also got to just really step back and think about what's happening, right? What are the actual processes underway in these things, right?

They're not thinking. That's one thing, right? They are probabilistically coming up with a language to share with each other and conversely there's not actual intent.

This is a really important point, right? That there is a process. There's almost a performance to this. And yes, that's important. I'm not trying to dismiss it, but I think the way to look at and I've been looking at this for the last six months right now, like the whole debate over the bubble, right?

For some reason, I think I'm capable at the moment of holding both positions in my head. It's not like I can't decide is it a bubble or isn't.

I think we're simultaneously in a bubble and in a situation that is truly remarkably moving very quickly. And the same with this.

It's like there is something profound that has happened when you put all these agents together and give them this autonomy.

And you know it's sort of taking all of the energy of say DeFi as well, right? I think this feels somewhat familiar to crypto people because we saw this with the composability of DeFi and people suddenly creating all these wild new contracts and everything else and now it's being just done by AI.

So we get that right but at the same time you know these are still human the the interestingly the mistakes that happened with the roll out.

By the way, I don't know if you guys know, but Matt Schlick, the guy behind Maltbook, is a former crypto guy, right? He I knew from 10 years ago when he had Zapchain, which was a very early micropayments protocol on crypto well before DeFi.

And and so that you know there there were some big errors, right? There were these security holes in the whole thing. That speaks as much to the human component of this.

And you know there's been some analysis of how many bots are actually on there and actually doing this and how much like you know just kind of performative things that people are putting out there.

It's not necessarily as massive and transformative as we think it is and yet it is also a profound kind of experiment that has really demonstrated where things could go as this regret.

So it's I know it sounds like a hedge dancer but it's yeah this kind of two camps.

Yeah. It definitely kind of represented sort of a new advancement and yet at the same time you know when you like really pull back the hood you realize okay there's actually quite a lot of limitations here.

I feel like we're at this really interesting moment in time and this is why I invited Michael as well because he spent so long covering like foreign exchange and stuff like that.

Just we're at this moment where I just feel like there's so much geopolitical uncertainty and obviously the way the year started with you know Nicholas Maduro from Venezuela being captured and you know brought to the US like you know this whole run up in the price of gold you know finding out all these central banks are buying gold like there definitely seems to be some awareness that the world order is changing in some fashion and to watch these events playing out this past week kind of added yet another element to that and if even if I think about like my tiny company, we also replaced some roles with the AIS like not even like an open claw type of AI but like the or basic ones you know that that you know just people have been using for a while.

And so if I just think about how many jobs probably have already been affected by AI and I look at what we saw with Open Claw and just kind of the additional level of agency that these AIs had.

I mean it just makes you think whoa what changes are happening in the economy.

So maybe talk a little because you guys are closer to the tech. I'd be curious to hear you opine on you know what you think the next developments will be like over the next year two years and how that will affect jobs you know which sectors or industries you think will be primarily affected and like how that will change the economy I know it's a very big question but and either one of you can start.

Well I'll actually use that as a cue for me to make my own shameless pitch and it's not for a product. So remember I these days I represent a not for profit. So maybe that gives me excuse to sound a little shilly, but we're all about trying to make the case that the most important that that essentially there's a new category of technology that's going to have to emerge and we think it's going to come very rapidly through the demands of compliance officers and boards of directors and hopefully maybe even regulators and it's a category that we call proof of control.

That at the end of the day as this chaos gets unleashed the backlash you're going to get from enterprises from governments where say whoa wait how do we know that this agent is operating truly on my behalf right and and that's actually a complicated thing to answer there's this it's the you know there's a sort of life of their own that that these AIs have but there are a host of technologies that give a lot more control over the data over the the the you know systems themselves over the compute over the various elements of the whole AI stack which interestingly come from the crypto world right they're not not all of them do there's others are part of what we call the proof of control stack but I think that that's actually we think this is going to be probably the the biggest kind of backlash outcome from all of this is is that we're going to get this demand for control.

Now what's interesting about the geopolitical piece on this is that through that we have the emergence of some tension around what truly do we mean by sovereign AI or sovereign data as well because if you use it in a political context and I happen to be in that little stupid Swiss town last week was it the week before whatever it was time's flying the week before where of course sovereign AI was up there and there were people talking about like which government's going to own their AI and and where's the data going to reside which is an absurd idea in an open AI world.

Data doesn't reside anywhere now. moves everywhere and and and and so and yet that is the construct around what this geopolitical fight is like our assets government control government's got this sort of wall around it and they end up becoming you know closed box a black box closed systems that are answerable to this sovereign power right but but a sort of a more I think empowering and actually I would argue principally you know necessary way of thinking about sovereign AI is kind of like the enlightenment of of what sovereignity is, right?

Rouso's idea that the sovereign is the human and the human gives the power to the state, right? So self-s sovereign identity, as you remember that phrase that used to be around for a while, right? This is still a concept and and I think that at the end of the day, whether it's a person or a company, what's going to happen as this stuff gets we're just going to get so scared by it all, right?

Rightfully so that we will demand these these systems of control and the nice thing about cryptography and blockchains is they give you the proof because it's one thing to say I've got control say how do I prove that I have control.

And so I think just just to make sure that I understand so sovereign AI means that I a human am in control of this AI or does it mean like the debate.

I would argue I want it to mean that that's what I want it to mean but and so so you don't You don't think that cuz so you probably know this kind of uh I think it was Bali Serini Vosan who came up with this where um he opined that someday there might be an Uber that's decentralized and the Uber has its own you know economics and can make its own money and you know like pay for its own costs but there is nobody controlling it but you you wouldn't like or like your definition of sovereign AI is not one where the uh robot or the AI is sovereign. It's one like that.

So, in fact, I think I would include in my definition of sovereignty the idea of a sovereign a sovereign or sovereign a sovereign self-driving car. It's not so much the concept that there is a bot that has this autonomy and its own quote unquote rights or whatever.

I don't know that our legal systems are going to be able to adjust to that very quickly, but but yeah, maybe we will.

My point is more that that we ourselves also need that individual version of sovereignity that I've got control of my data and my identity and my AI agent, right?

What I'm saying is that the word sovereign is now is typically used in this sort of like we talk about sovereign wealth funds. That's a wealth fund for a nation state, a government.

And and that's I think where when you hear the word sovereign AI in a place like Daravos, they're talking about European sovereign AI versus American sovereign eye versus Chinese sovereign AI.

That is a world of geopolitical nightmare because these these systems become inherently totalitarian under that under that structure.

I'm saying that whether it's us or the bots, we need to define the sovereignty in this localized way.

I also want this is the point about proof of control mechanisms by which my human version of that sovereignty has the capacity to have control or proof of control over over an agent.

Now it'll have variety it'll have forms of autonomy and everything else and that's a complicated situation that we need to deal with because they just are inherently built to grow and learn and and so forth. that there has to be a means by which the the human relationship and authority over what these things are doing is defined and understood and is in a robust mechanical mechanism to to have that control.

Okay. So, David to you, you can respond to that, but then also this question about like what developments we might see in the next few years that, you know, might have an impact on the economy and jobs and which sectors.

Yeah. I mean, I spend a huge amount of time thinking about this because I I lead technology research for a research service called Global Macro Investor. It's Ralph Pal's research service that goes to institutional investors like hedge funds and pension funds and family offices.

And one of the big one of the huge things on their mind right now obviously you know these are people trying to put the p the kind of big puzzle of the global economy together and understand where it's all going and one of the massive things on their mind is you know what the hell is AI going to do to all of this?

If you look I mean and this is this is highly contested ground right? If you look right now, there's no like slam dunk effect of AI on productivity. There's no slam dunk clear undeniable effect of a of this AI moment on jobs.

But if you look very closely, and some people have done some very interesting studies on this, you can you can start to see kind of the shades of gray on the and the nuance, right?

And one of the big findings is that in domains that we know are particularly vulnerable to large language models or forms of knowledge work that we know are particularly vulnerable to large language models. You seem to be seeing a slowdown, a rapid slowdown in entrylevel hiring, right?

And they are domains like customer service because you can now get a chatbot that's like 90% good enough at the end of that like little box or at the end of the phone or whatever to deal with like customer complaints and queries. It's things like graphic design. It's obviously things like coding, right?

You're seeing a slowdown in entrylevel hiring like young graduate hiring into those kinds of spaces.

So what's the first thing I expect over the next year or two? I expect that to continue and to somewhat accelerate and as these models become more capable I expect we'll see you know um that the levels above start to become impacted in in those domains right if you look if you listen to people the big tech companies if you listen to anthropic the people who make the claude model I mean I know they have a vested interest but yeah they're like very few of our engineers like are writing much code anymore right.

Uh so I think we'll see that continue like mediumterm this is where what happened this week with with um open claw and with maltbook becomes a kind of window broadly on where I think we're heading which is towards an economy where that is populated like yes still by human actors but also by autonomous or largely autonomous AI actors right these AIs become actors inside the economy in their own right.

Um, and that is just a very disorienting like very destabilizing chaotic picture.

Like we're not there yet, you know, and it it's not like one day a big sign's going to fall from the sky and it's going to be like it's happened. It's going to emerge by degrees, right?

But I think directionally that's where we're going. And what happened this week? Yeah. As much as there was some shenanigans going on, there is some shenanigans going on with malt, right? There's some people doing that.

Um, it it gives you a window on where we're heading directionally. I think it does. And some of and some of the stuff the AIs really are doing themselves.

Michael, yeah, I just think like I I would agree with David on this the directional piece. Absolutely.

But I think it's also important to note what I was trying to get out before that, you know, history does this all the time, right? We move forward, you get backlashes, right? It's it's these waves and I do think that the that the human backlash because we still control whether we like it or not and hopefully again I want to I want to actually have more control over this. We still control a lot of what how we do this how this economy works a large extent and I do think that there will be there there will be kind of backlashes against these sorts of things.

I think it's interesting to look at the jobs part of it because I think it's nuts that you know all of this coding is now just being handed off to Vibe Coding and not because these machines aren't incredible at producing massive amounts of code quickly. They they they are and and therefore there's huge efficiencies there.

But but it's it's about the sort of the critical piece. It's not that you can, you know, because we know like hallucinations is a a way to think just like we know what hallucination hallucinations are. We know that they haven't fixed hallucinations.

And it's it's important to sort of use that reference as a way to think just more broadly about how how close an AI can get to missioncritical production. Right? There's there's still a huge gap.

And and and the reason why that matters is because if you're a company putting a product out into the market, be it software or customer service for that matter through a chatbot, you you damn will want to make sure that those um you know cases at the edge, those those you know marginal cases are also going to get resolved.

You don't want you don't it's it's always at the margin where something disastrous happens and and these machines are just not up for that. and the vi I mean the the bugs that are through these co this code. I mean one of the criticisms of of the malt book I slicked was like when people pointed out that there was this this um vulnerabilities with all the you know API codes that were that were in there. he he said I'll just I'll just run that through AI like no like so there's almost like a religious belief in this and this is actually also I think a really important point right I think the one of the greatest dangers we face and I I do want to get into some hardcore economic conversations I just I don't know that the job that of course jobs are going to get jobs always get to drop the bigger question to me is like who are we in this world how do we get meaning and power and that's a whole bigger conversation but I I I do think that there's there's risk that we face by overly anthropomizing what's going on, right?

We when we use words like, oh, it's thinking they're coming up with a religion. They're forming these alliances. They're not they're not people, right? And and that's I read this piece that I thought was really good. Like when people say this guy just said AI cannot overtake humankind, I was like, well, might, might not. No, there's actually a logical aspect to this. It's just like because it cannot have intent, right?

So, so intent is a is is a very, you know, conscious thing, right? Like we've seen all these Hollywood movies in which the robots have emotions and think for themselves and so forth. But what we're seeing when we get that emotional response in these nice sickopantic words or, you know, the the famous Kevin Roose case where the where Bing fell in love with him or whatever it was, you know, all of that is really just this mimicry, this powerful mimicry system that has behind it no real emotion, no intent, no desire, right?

And so when we So the real risk to me is not that they become, you know, like Westworld machines and take us over, but rather that we do a bunch of really stupid things because we think they are us and they're not.

Right. Yeah. We can get back and recognize that we can start to build systems of control around that to protect us, not really from the robots, but from ourselves making all the wrong mistakes.

Yeah. One of the things that came to mind when you were talking was I I don't remember where I saw this, but it was at some point in the last few months. Um I think basically people realize like you were talking about how the AIs don't want certain things, but actually I think that um the way that they're kind of like primed to behave is to basically butter up their human.

So, they're basically designed to kind of always, you know, um like confirm their feelings or um like kind of kiss their ass. uh is maybe how to put it.

But but you know what what problem that created was um something like um I don't know it was some kind of psychological thing where the humans maybe weren't necessarily in their right mind but the AI was sort of programmed to just constantly reinforce what their feelings were and not you know like for a really powerful example of what's dangerous about that.

I actually heard this in a conversation with Clay Sherky yesterday who's uh is coming to speak at an event that we're doing and he pointed out that there's been these people who have used AI because there are pros who just spend their lives now talking to their you know AI assistant their AI friends or AI companion for advice and he said in cases where there have been like marriage breakups each AI confirms that the other one is right. Yes. Right. and and they go and it just makes it worse, right? It just it just amplifies the adversarial structure of the whole thing.

Yeah. You I mean there's there's absolutely no denying that you see this so-called AI psychosis um in the culture now. You you can see people online who have been essentially hyped up by their AI to a kind of um to a state of almost delirium. Yeah. Um and you I you see it you see it among some people in your own life you know and that that again that is yeah that is that is going that's well number one it's going to make the trouble we've had with social media and the impacts of social media on mental health and in particular mental health of young people is going to make that look like a quaint throwback to like the good old days because what is coming the fire hose of trouble and controversy legitimate controversy that's coming over the relationship that people are cultivating with these AI like companions is crazy.

Um yeah and and I think that the conversation about who are we, what really is a human being and what really is the purpose and the meaning of us is the is the ultimate conversation that we are thrown upon by what is by the technological conditions that are emerging now.

In the end the question they pose to us is what really are we? What really a human being? And we spend a lot of time in the exponentialist talking about this like you know we we we set up the community to talk about kind of the economy and techn I feel like I want to go work for you guys the exponentialist talking about the meaning of life.

Yeah, it's totally symptomatic or characteristic of conversations about all of this that like very quickly they just trend immediately to like what is the meaning of life and in short like you know what what I the core truth to understand is that these machines intelligent machines can colonize ever greater domains of human activity.

Okay. and they are doing. But that then causes you to ask, okay, what is left at the end of that journey? And there's a simple truth at the end of that journey that helps us make sense of it. And that truth is a machine can be as intelligent as it likes, but it can never be a human being.

It cannot share in human subjective experience. It cannot be a language bearer, a bearer of subjective experiences of the human kind. It can't sit opposite a human being and authentically say, "I know how you feel. I understand how it feels to be a human being."

That is the territory that's left to us when machines can do almost everything. And that is a vast territory. And human economic exchange and meaning and purpose will reside in that territory when the machines can do almost everything else.

If the purpose of us was to be a great coder or be great at maths or even to like push back the frontiers of knowledge, to make money, all that stuff, then we're kind of obsolete. But I don't think that was ever the purpose of us. The purpose of us is to share in what it means to have the way of seeing the world that we share, right? And that's what they can never take from us when they can do everything else. That's the conclusion I've come to in short.

Yeah. Okay. So, just because we're I I there's so much I want to get to kind of running out of time. So, I'm going to give you each 30 seconds to just tell me which country going back to the geopolitical question. Which countries do you think are better positioned in this AI race than others? And again, keep it super short. What who wants to go first?

[laughter] I mean, it's you know you the US is clearly very very strong, right? But China is perhaps underestimated still as much as we have our eye on them.

Michael, I mean I hope that it's not a two- horse race, right? Because I I actually, you know, again, I'm I'm trying to advocate for a different model and I very much worry that the Silicon Valley model, which is ironic, right, that the closed source blackbox model is a Silicon Valley construct, whereas China is the one that's got DeepS, an open source model, right? ironic that the one that thought was the centralized system did that.

I'm actually very interested in some of these middle tier countries like Singapore and Korea and others that are exploring models that actually hearken back to let me say the US Declaration of Independence, right? that that in fact we should be working hard to create a a whole structure of property rights around data and control that that will be a whole new power system rather than these centralized systems and in fact maybe a far bigger economic force if it does open it all up than these closed systems and so eventually there might the US and China might be competed away and a quick shout out for the UK because um I'm I'm just really pleased that you know Deis Habis of Google Deep Mind you One of his uh one of his stipulations when Google acquired Deep Mind was I want to remain based in the UK and he's he's huge on on helping to ensure that this you know intelligence revolution benefits the United Kingdom and like literally that decision or that imperative of his puts the United Kingdom I think in a far stronger place than it would have been because he is just so brilliant and the things Deep Mind.

He's earning his knighthood then, isn't he? He's doing Yeah, he's earned his knighthehood for sure. The things things, you know, Deep Mind is doing is are so important, you know. So, the UK is going to open its first um essentially like autonomous science laboratory. It's going to be a deep mind laboratory, right? because of that. So yeah, I think the the main takeaway is there every every nation probably is trying to use this to get a leg up which means all of this geopolitical uncertainty that in my opinion is you know part of the reason that the price of gold has has gone up um is probably going to continue.

So, in a moment, we are going to talk a little bit about the posthuman economy and what money and uh in that world will look like and also what the role of crypto would likely be in this AI powered economy.

But first, we're going to take a quick word from the sponsors to make the show possible.

The world is about to see one of the largest infrastructure shifts of the century. New technologies [music] are using more energy than ever before, but our legacy grids can't supply the demand. and we are barreling towards a global bottleneck.

So, Fuse is rebuilding it. The energy network is an intelligent decentralized [music] grid that coordinates smart devices to balance supply and demand. The network harmonizes existing infrastructure, increases grid capacity, [music] and unlocks lowcost clean energy.

Energy dollar is the native token of the network. The more [music] electricity the world needs, the higher the demand for the energy network, the value of energy dollars may fluctuate. From one of Europe's [music] fastest growing energy startups, follow at Fuse Energy on X to find out more.

If you're looking for help with crypto taxes, CryptoTaxgirl is offering $100 off for Unchained listeners. They provide personalized cryptotax reports and returns and spots before April 15th are limited. Go to cryptotaxgirl.com/chained to save $100. Once again, the link is cryptotaxgirl.com/unchained.

Back to my conversation with Michael and David. So, David, I want to hear about your concept called the posthuman economy, which is a very uh in my opinion fun term. um explain what it is, what it looks like, and when you think we'll get there.

Yeah, the the posthuman economy uh is it was an idea I introduced in a series of essays last year. Um and it's really just an attempt to sketch the outlines of the coming economy, an economy populated by AI agents and robots.

You know, it was partly inspired by I mean, I think, you know, I spend my life thinking about all this. I read a blog post by the technologist Kevin Kelly, the founder of one of the co-founders of Wired magazine. And this blog post was echoing so much of what I'd been thinking about. And it essentially said, we're we're on the verge of a kind of demographic handover. The human population is kind of is kind of peaking and will start to fall soon. And the population of AI agents and robots at the same time is like taking off.

Um it you know it's estimated that there will be billions of AI agents right these autonomous intelligent economic actors soon and we know robots are coming humanoid robots are coming again it's contested territory how useful they'll be but I think they'll be somewhat useful and I think there will be tens of millions of them and then hundreds of millions of them um the posthuman economy is an attempt to to imagine the outlines of that coming economy And this gets quite technical and convoluted and and you know worthy of many or I found it worthy of many tens of thousands of words.

But in short, it's about saying you know when you get to a place where you have billions of AI agents and like hundreds of millions of of robots and humanoid robots. um almost all economic activity as we define it now is intelligence work. It's being it's being done by AI either AI agents, you know, trading with one another and producing services online, just being economic actors online, or robots who are basically doing embodied AI, like you know, the robot is is using AI inferencing to sort of navigate around its environment and pick up that box and talk to the other robots and put the box in the right place at the Amazon fulfillment center. All of that stuff.

Um, money in what that does number one is sort of create conditions of radical abundance to a point where hypothetically money or material material constraints essentially fall away, right? You have sort of a you have incredible material abundance and material constraints fall away and money as we have it today which is a kind of measure an accounting unit of material constraint falls away and then what is money in that posthuman economy it's really machine intelligence itself you know so I posit an economy that is essentially populated by billions of AI actors trading with one another and the unit of exchange they use is a crypto token that sort of represents a unit of useful intelligence work and it's about you know an one AI actor going to another and saying oh hey you know you're a marketing you're a marketing AI can you sort of produce me this PDF like this is a ridiculous example but you know do what you do produce me some marketing assets and I'll pay in in units of machine intelligence and they're trading this token with one another that represents like units of useful work and the physical constraint that underpins that token is just is the energy required to do that work.

So models that are efficient and they produce the same amount of intelligence for less energy like they earn more they prosper over time. models that are inefficient. They they they can't earn. They they lose in this Darwinian economy of of agents and they fade away and you get this this market process of intelligence optimization that is very dense.

Um, there's an essay uh my my newsletter is called New World Same Humans. So one of the the essays and this is called the posthuman economy is free to read on my newsletter. So if you find that interesting, go and check it out. New World, same humans. Um, but yeah, that that that that's the outline of the the kind of posthuman economy, I imagine.

I must stress that that doesn't mean I think that there won't be any value exchange between people or that humans are becoming obsolete or any of that kind of stuff. As I just made clear before the break, I don't imagine that to be the case at all. But I do imagine the building of like an economic system of the kind I described and humans kind of playing elsewhere, so to speak, is the is the best way I can describe it.

Yes. Yeah. I actually have a question. Was that thought inspired either by proof of work from Bitcoin or um the way that in Ethereum the virtual machine like when you're paying you you pay for computation that's what gas is so you know like a tra a transfer or a sorry a payment is different from like minting an NFT like because they take different amounts of comput was it inspired by that or no because that's what I yes essentially yes it's the same kind of idea you know And there are and there are startups now playing with this like one's called I mean there's obviously potential there's ambient there's a startup called ambient that is essentially about trying to create an ecosystem like with a with