How Agentic Payments Signify the End of Traditional Banking (Full Transcript)

BB – 5.15.25 – Video

I think, you know how it like became a thing to have vinyl records again? Yes. I think doing podcasts in person again is like the next serious trend. The next vinyl record.

Yeah, I think so. Coming back. Doing it in person.

I think it’s coming back. Because this is so much more fun than doing it online. It is.

It is. It’s great to have you at my home in Bangkok. Thanks for inviting me.

We’re right in my studio right now. This is awesome. Thank you.

So, it’s good that you’re here for Money 2020, of course. I wanted to, you know, we’ve had you on Breaking Banks more times than I can remember. We had a few really interesting shows.

Remember the show we did on M-Pesa? Long time ago. That was a good one, 2014 or so. And I remember that we had some really funny, fun times on the show.

You came up with a… You had me on your craziest ever episode with John McAfee, which, you know, I think set the bar. That’s right. That was pretty crazy.

For Lunatic Podcast. Rest in peace, John. Now, that was a pretty crazy podcast, that one, the McAfee podcast.

Anyway, I thought today, given we’re going to do Breaking Banks and Futurist combined from a content perspective, we have to talk about the future of money. And you, of course, have written extensively about the future of money. You’ve written… You’ve had a couple of recent books that have gone into that.

Of course, you’ve also spoken extensively about identity. I’m very, very interested to talk to you about machine identity and where that’s going to go. Very hot topic at the moment.

And KYM, right? Or KY? KYA. Know Your Agent. Yeah, Know Your Agent.

And then just in terms of where the utility of money goes in highly autonomous economies. So it’s very interesting. Now, we’re in the midst of this tariff war right now between the U.S. and China.

It looks as if that has not been very successful for the United States. I think there’s a number of reasons for that. But certainly, China’s trade is at a point now where they’re somewhat protected from the U.S. I don’t think the U.S. expected that.

But the other element that I’ve seen over the… I’ve been going to China now since 2001 was my first visit to China. Of course, lived in Hong Kong for seven years. But the really interesting piece of what I see happening in China is the industrial scale application of autonomy.

So, you know, the dark factories. There was reported earlier this week that Xiaomi and Huawei have these factories that can spit out new phones. One new phone, a smartphone a second.

And these are completely human-less factories. The only staff that come into the factory, they have the lights off. They’re called dark factories.

The only staff that come in are the cleaners and the machine repairers who service at night. They have broad autonomous vehicles. They’ve been operating autonomous vehicles in places like Shanghai and Shenzhen for a number of years now.

I know the U.S. with Waymo is making some progress there, but China appears ahead of that. 80% of Chinese ports are automated now. Autonomous, operated by robots.

Their application of solar, they deployed more solar. My listeners will be tired of hearing this, but they deployed more solar last year than half of the planet. So there’s a lot of next-generation infrastructure happening in China.

And the U.S.’s attempt is to sort of onshore manufacturing again because of this. But, you know, it’s complicated. They don’t have the tooling skills.

They haven’t really got the STEM education support for this. They don’t have the government incentives that China does for supporting the industry. I’m not really an expert on this sort of thing.

But the observation that caught my eye about it recently was, I think, something like a third of the Chinese Communist Party Politburo are engineers and have worked as engineers. And I don’t know what the commensurate figure in the U.K. is. I think at the last count, I think there was one member of parliament out of 630 who had ever worked as a scientist or an engineer.

That’s pretty interesting. And I think, you know, I don’t know that many U.S. lawmakers, but I have the impression they’re all sort of lawyers. Yeah, a lot.

You know, financiers. So I think there’s something to what you’re saying there. And I certainly think, you know, I mean, it’s a kind of flippant comment to make.

But if I was the dictator in charge of bringing back manufacturing jobs, I’d be deporting robots, not people from Haiti. We’re in a different world now. And it seems slightly anachronistic to me, not being an expert, that sort of 1930s style policies are back in vogue.

But as I say, where that’s going to go, I really don’t know. But if we start talking about an area that you are obviously very well informed on, which is the payment space, you’ve been working in that space for a long time. And when we look at the payments layer that is going to have to accompany this large scale autonomy.

So when we start looking at autonomous payments, agency based payments, smart contracts, all of that next generation infrastructure we’re going to need financially to operate the autonomous world. That’s something that you’re obviously a little thinking. I think the way to think about it.

I mean, I hope this. I mean, I always have models. I need to think things through.

You know, so I think you and I have talked about this before. There’s this concept of digital public infrastructure. Countries need digital public infrastructure.

And in some countries that’s very centralized. You have a centralized. So the digital public infrastructure is basically digital money, digital identity and digital data exchange, open data.

And in some countries that’s very centralized. The government provides it like India or somewhere. The government provides digital identity.

The government makes everybody use one value exchange system. And, you know, in time they’ll have open data and so on. So I think the way to think about it is what’s the equivalent of digital public infrastructure? But for machines, not for people.

Right. What does that look like? And I mean, I don’t know. Is the obvious.

These are early days, but I think we can make some reasonably informed speculation about it. So. So the first thing is the digital.

The digital money. How are how are agents going to pay each other? And I think there are two strands that are worth working into the thinking that one of them comes from a very unusual source. So I know I’m sick of talking about central bank digital current.

Well, I’m not. Everyone else is sick of talking about central bank digital currency. But in the original consultation for the digital euro, the response from the German Banking Association, which I regard as one of the most conservative organizations on Earth.

I mean, you’ve you’ve dealt with them before. It’s the Landers Banks and literally the most conservative organization you can imagine. And they said, well, look, you know, we think wholesale digital currency was a bit of a no brainer.

Everybody wants wholesale CBDC to make markets more efficient. Retail central bank digital currency, probably a good idea in the long run, but there’s no burning platform. No, no great panic.

But, you know, in time, I’m sure that would work. What we actually want is machine digital money. Right.

Because, you know, this the German banks and they you know, they’re interested in manufacturing IOT 4.0 enhanced supply chain sort of stuff that you were talking about there. And they said it’s time to start thinking about the money that’s going to be used in those supply chains, those machines. And I at the time, I was so surprised and interested by this.

I looked at the authors and I emailed them. I spoke to one of the authors. He was actually in Switzerland, not in Germany.

But I spoke to him about this and he said, well, this isn’t particularly futuristic because the big German companies are already doing experiments with this. You know, it’s like and I was really. Wow.

When you think about it, he’s right. I think that’s we shouldn’t be thinking about fiddling about with the money for the stuff we do now. We should be thinking about the money for what’s coming next.

Exactly. And to see that come from such a conservative source, I thought was very interesting. And then the second thing is I was in the States before Christmas.

I went to see a couple of companies. Well, already fun. Not hypothetical.

Already funded startups that are looking at agent to agent payments and how to implement them and how to make them all work. So this is already started. We’ll come back to what that value might be in a minute.

But to deal with the immediate question, people already started work on that. So the digital money. The digital identity is a bit more complicated.

Some some work has started on it. We have we have this concept of K.Y.A. You know, your agent. It’s a little more complicated, I think, than sort of K.Y.C. Because if my agent shows up at your concert, if my agent shows up at your bank and says, you know, can you send some of Dave’s money to somewhere? You need the identity of the agent.

You need to know who the agent is for, you know, obvious reasons and for audit and recovery and whatever. But you also need to know that I gave it permission to do that. It’s a little bit more complicated than like when you’re doing K.Y.C. Like, is this Dave Birch? That’s all you need to know.

But then I have to know, well, is this agent who it says it is? Does it it doesn’t have the authority to represent you? What if any? Are there any boundaries of that authority, etc.? You need a more granular permissions based system. There’s work being done in that space already. I mean, it’s not it’s not a solved problem for a reason I’ll explain in a moment.

But there is work already done in that space. So then the third part is the sort of data exchange. And here we’ve got the first baby steps already being taken.

You’ve got the anthropic MCP model context protocol. People are already building MCP clients, MCP servers. MCP as it stands is a first step.

It’s not sufficient because it doesn’t have that kind of security infrastructure built into it yet. It doesn’t do the authentication and authorization properly. And those are interesting and complicated problems.

Because with a person, for example, you’ve got a smartphone. You could have a secure app on your phone. You could use the secure element in the phone to store the keys and this kind of thing.

But if you’re an agent, where are you going to store those keys? So it’s interesting. So you’ve got those three sort of components building. Not complete, but enough to see which way the wind is blowing.

So just put that bit in place and then come back to answer your bigger picture question on top of that. So given that that seems like it’s going ahead, given that I said this at Money 2020 and I didn’t mean in a flippant way. But remember, for most consumers, the stuff we’re interested in is quite boring.

I mean, they basically don’t care about it. So as soon as they can hand it off to bots to do it instead, they absolutely will. No one cares that much.

And also the complexity that, you know, if you imagine like the simple use case of, OK, I’m just going to pay for something at the store. And if I was a rational person, I would stop at the checkout and I would think, OK, should I use my Amex Platinum card, which is literally the most expensive way on earth of paying for anything and get the membership rewards, which I could then use to trade for British Airways miles? Or should I use my British Airways Avios and get the miles? Or should I use my John Lewis card and get the cash back? Or do I remember that there was a thing on my debit card that if I used it five times in a supermarket this month, I would get two pounds off or something. OK, well, that’s quite an easy problem to solve, right, Brett? So all I’ve got to do is just go to my calendar, have a look at the places I’m going to be going over the next few months, go to the British Airways website and look at what the cost of the flights to each of those places is, and then go to the Avios website and see how much they would cost in Avios and also how much they would cost to go in economy and then upgrade with Avios.

So you get the rough exchange rate for the different things and then work out what the exchange rate from them is. And it’s quite simple, you know, but for other people, they might not do that. But for a bot, that’s nothing.

Yes. That’s absolutely nothing. Yes.

And I don’t think the retailers, the banks, the airlines have any strategy. Yes, I agree. But what happens when you have customers who are 1,000 times smarter, literally overnight? Yes.

What happens when customer use is optimised? And having rewards promotions up in the supermarket saying, you know, you can get cash back or these miles or this gift if you use this card, that’s not going to work because that’s not marketing the bots. And also it partly doesn’t work for marketing to people. Maybe some of that’s deliberate.

I mean, but I do sort of think, you know, you get the same thing from I get the thing from my credit card provider. They’re doing these card linked rewards. So if you go to Marriott and then you go to, you know, pizza, Olive Garden and you do it in May, but you take your grandmother with you, and it’s like, you know, like you start reading this, you know, complicated.

OK, so when you come to the near future world that you’re talking about here, you do have to ask yourself, how’s that going to work? And there was a very interesting piece I read a couple of weeks ago from the CTO of Walmart. Very smart, very interesting. And he said, well, look, we’ve got no choice.

We’re going to have to build our eyes because we don’t want customer eyes talking to our website. Yeah, it was never designed for that. And we don’t want their eyes using MCP servers to get to our MCP to get to our APIs because they were never designed for that.

What we’re going to have to do is build our own eye. So the customer eye talks directly with the retailer eye and then they can strike. Then there’s a process of negotiation.

There’s a process of negotiation, but it’s informed negotiation. In other words, it’s constrained by regulation. So to me, none of this seems crazy.

I think this is, I mean, it’s not there yet, and it’s fascinating and interesting. And I strongly doubt that any of your banking clients have any strategy for what happens. No, I would say there’s very few.

The other element of this that seems fairly obvious to me is the reason we need not only, you know, AI infrastructure for companies like Walmart to handle the incoming requests from AI agents that we’re going to have in the consumer space. But the other thing that’s fairly obvious to me is that the payments rails that we’ve historically been using, Swift, ACH, Visa, MasterCard, and so forth. These were built in the 1970s when compute power and telecoms infrastructure were at their most expensive.

You couldn’t send huge amounts of data because we had to transmit over these copper wires, over modems. We didn’t even have broadband at the time. And you also have, at the same time, compute power, which is fairly nascent.

I mean, smartphones now have billions of times more processor power. So you could send only a small amount of information. So what we had to do is we separated authentication from the transactional element of payments because of the data restrictions we had sending data through those early networks.

We don’t have those restrictions anymore. So why isn’t it that every transaction that we’re sending doesn’t have detailed identity information parsed with the wallet address as well as the transactional information? When you say detailed identity information, because you’re saying that to a horrible identity nerd. So I’ve got to say, I mean, I think we should be aiming for a slightly more sophisticated infrastructure than that.

I go online and, as in fact happened to me yesterday, I’m trying to do something. And it turns out the piece of software I was mucking about with video files to upload to LinkedIn. And it turned out the old bit of software I used to use suddenly doesn’t work anymore.

So I go in and I want to buy this piece of shareware and it was, I don’t know what it was, $9.00. And I didn’t want them to have my name and address and my personal details and all this sort of thing. Because it means when they get hacked, which is all the time, all my stuff’s gone. So in that circumstance, what they need to know is maybe I’m over 18, maybe I’m resident in the UK.

That’s pretty much it. And we have the technology to do that now, which we didn’t have in the 1970s. So zero knowledge proofs and cryptographic binding and that sort of thing.

I can prove to you that I’m over 18 without sending you my date of birth. I can prove to you all sorts of things about data. So you’re right.

So technically in my head, I’m translating, when you say identity information, I’m translating that into sort of proofs of identity information, verifiable credentials. But yes, you’re right. And this structured packaging, this is the data part of the thing.

So, okay, we’ve figured out the AI’s identity. We’ve figured out value. We haven’t, we’ll come back to that in a minute.

And now we’re going to, the other part of DPI is the data exchange. And there we have the MCP and the structured context data. So this idea that you package up the credentials and all the other stuff you were talking about, it all goes in a bundle from my AI to your AI.

Your AI pulls it apart, queries the servers it needs to get what it’s done, repackages up the answer and sends it all back. That’s actually a pretty attractive future. Yeah.

Because then when I say to my AI, oh, oh my God, I forgot I was going to go to stay at Brett’s. Can you move my flight back by a day? And can you book the hotel for an extra night? Oh, and can you arrange a grab for me at five o’clock? Which are not, not competent requests really. I mean, but, but right now you can’t do that.

You’ve actually got to, I can rebook the flight, but then I’ve got to go and log in myself because it can’t pay the flight change fee and this sort of thing. That’s not that far off. That’s not that far away.

So now let’s take the futurist view on that. So I can understand how, let’s say for sake of example, you have a server which provides very good quality weather information. Your server becomes very trusted.

People want to use it. When I’m trying to work out, you know, what I’m going to wear today or when chat GPT is doing it for me, it will, it will cheerfully pay your server, you know, a 10th of a cent or a hundredth of a cent to do that rather than look at untrusted sources. In the early days of the internet, there were lots of people, me included, who thought that a system based on some kind of micropayments infrastructure like that would be better than what we ended up with, which is advertising based.

Wall gardens, you know, that kind of thing. So do I tell my, and we’re talking about agentic commerce here, so I’m not telling, I’m telling the agents what to do. The agents are going to figure it out themselves.

This is a world of, you know, vector databases and MCP queries and all this kind of thing. So if they have a budget to, to, to, to get the best solution instead of relying on app, cause, cause we’re human beings. And so, you know, advertising kind of makes sense and we’re emotional and we can be nudged and steered and, but in, in the sort of bot world, maybe not.

And then that led me to think, well, okay. It’s a much more mission based. They’re much more mission based.

But then what would they value? Like I, I’m. How do you market? How do you market to it? I’m trying to sell my, I know I’ve just come from on chain 2025. I saw an interesting presentation about trying to shift people from using stable coins to using yield bearing coins.

So how do I persuade you to come and use that thing? I don’t care that you sponsored the Superbowl. Couldn’t care less. You’ve got your sticker on these formula one cars.

I don’t care. You’ve got this beautiful logo that cost you millions of dollars. And I, I couldn’t care less.

You’re found as handsome, very photogenic. I don’t even care if he’s real or not. You know what I mean? Yeah.

Like, so what, what would I value? And then I started to think, okay, well maybe if I was trying to attract the agent instead of the person, I’d give the agent. There’s no point giving the agent air miles. It can’t, can’t fly anywhere.

But if I gave it compute power. Yeah. Or maybe I gave it access to more powerful GPU chips or maybe electricity.

I don’t know. Then you get into some more interesting speculation about where that might all go. The other thing is on the hallucination side.

Yeah. Which we talk about. One of the things that seems fairly obvious, which we do have the ability to do somewhat with decentralized networks, is rollbacks.

Right. So if you have an erroneous transaction, in theory, you can roll back transactions, even six transactions ago that occurred in your wallet and so forth. We don’t have anything like that today in the current financial system.

You have chargebacks. But that’s a process of recovery from fraud or recovery from a fraudulent transaction where you’re trying to get your money back. Whereas in an agency payments world, surely we’re going to have to have the ability, at least in the early years of agentic payments, to actually say, no, that transaction was in error.

Let’s revert that and roll it back. That’s a really interesting, because, you know, whether that should be part of that, that DPI layer or whether that should be the rules based stuff that sits on top of it is not, is not completely obvious. You can argue very successfully that one of the important reasons for the growth of credit card usage, and one of the reasons why I use them for lots of things, is that I don’t care.

Because if anything goes wrong, it’s not my problem. I’ll charge it back to the bank. It’ll get sorted out.

They have to take care of all the fraud and all that sort of thing. The scale of agentic payments could mean that that number could balloon. And if it’s through, you know, AIs just making simple mistakes, you know, then why should banks… But that’s why I don’t want the AI to have my credit card.

Right. So instead we have to have a layer of protection there almost. So do the AIs have their own wallets? Well, I mean, that sounds crazy, but I think the answer to that is yes, I think they do.

And in fact, I think a good reason for thinking that, there’s a very good book by James Boyle that came out last year called The Line, which I like very much. He’s the guy that wrote The New Dark Ages. Really, really good writer.

And he says, I mean, I’m super paraphrasing, but he basically says, look, the reason why corporations have legal personhood and the reason why some kinds of trusts have legal personhood isn’t because we think from first principles they have, you know, there’s some moral basis that they’re a person. We know they’re not a person, you know. See, like, stop thinking about this.

Stop torturing yourself and saying, like, is the AI really a person or not? Not the point. The point is that’s done for economic efficiency. So legal personhood for AIs is inevitable because it will be a matter of economic efficiency.

So when the AIs have legal personhood, they can open their own bank accounts and do their own transactions, that sort of thing. And of course, they’ll have their own wallets to do that sort of thing. In the interim, of course, they can’t have that.

But they can have, and of course, this is what all the guys in, you know, this is what all the guys who are playing around with this stuff right now are doing. They can’t have bank accounts, but they can have crypto wallets. Yeah, yeah.

So, so the de facto… I’m doubting agents work on top of that infrastructure. Yeah, so the de facto currency for agent-to-agent commerce is already U.S. dollars because they’re all using U.S. dollar stable coins. USDT.

Yeah. Yeah. And others are available.

On that note, let’s take a quick break, Dave. Okay. We’ll go to break, and then afterwards, let’s come back and let’s do a bit more future visioning and see at scale how AI-based financial infrastructure might change the way we think about the financial system itself.

Okay. I am breaking in with Dave Birch, the futurist who’s looking at identity and payments infrastructure for the machine world right now. And I had Nick Hughes on talking about M-Pesa… Fantastic guy.

…a few years ago. Really interesting, really nice guy with Joanne Barefoot, actually, who you know as well. Yeah.

It was very interesting because when he described for me the early prototype MVP of M-Pesa in this little village in Kenya, and it’s like, that’s tokenization. Essentially what they’ve done was earlier tokenization, I think, but you were involved in that project as well, right? It looks like, because it really isn’t, because although with M-Pesa it looks like you’re sending the money from your phone to the other person’s phone, in reality you’re not. What you’re actually doing is sending a message to the central system, and the central system then sends the money to the other account.

It looks like you’re sending it from person to person. Actually, you aren’t really. What’s different about if I’ve got a USDT stablecoin and I’m sending it from my phone to yours is it really is going from me to you.

It isn’t going through a central system. But Nick, honestly, is one of the people I… Him and Susie Loney, who were the team that built that initial, are just wonderful people. So now he’s working on carbon credits, which is a really great area to be looking at in terms of development.

But as we move forward, it seems to me, you know, plenty of people, Bill Gates included, Elon and others, are talking about universal basic income. They’re talking about the fact that within 10 years, our pattern of work will have definitely changed. I wrote about this in The Rise of Techno-Socialism, of course, because the level of automation in the system will require us to work less.

We’ll have UBI and UBI will cover our basic needs. But, you know, we’ll still have to work because our basic needs won’t be enough to get new iPhones and travel and things like that. So you’ll still have some work.

But this is the sort of world that could be 10 or 15 years away based on high levels of automation, particularly in developed economies, certainly in a place like China. But if we take it another step further, we take it another 10 or 15 years out. So we’re looking at 2050 now.

And now you have large scale automation where it’s conceivable in the most advanced economies that a lot of your daily needs will be looked after by AI. Health care will be optimized for AI. Education will be optimized for AI.

Every child will have their AI teacher that will be a full time coach that helps them. Food production will be automated. Transportation and shipping will be automated.

Most production facilities will be dark factories automated. In that world where there’s so much automation, what is the role of money? If you can just take resources and convert them at that level of efficiency and distribute them so efficiently, then what utility does money have in an environment? Well, in the long term, it’s not obvious that money as we know it survives. And that sounds slightly crazy, but I think it’s a really I think it’s very well founded.

I was talking to Matt Harris about this from from Bain before he wrote. He wrote some stuff about this as well. You know, the sort of the common theory of money is, you know, I want some eggs.

You want some fish. I don’t have any fish. You have any eggs.

So we use money as an intermediate to to replace the double coincidence of ones. But if you if you take that forward and we start living in a world of digital asset and tokenization and that kind of thing, why would I? And so I have I have assets which are which are hopefully, you know, growing in value and that sort of thing. So I have my digital assets, which are stocks and bonds and gold and bits of who knows what doesn’t doesn’t sort of matter.

In the kind of intermediate world, which we are about to move into, because I do I don’t think that’s hyped. I do. I do.

I do genuinely believe the whole kind of digital asset tokenization stuff that that is the next generation financial market infrastructure. So at the moment, I take my two square inches of the Mona Lisa and my my right to travel on this train and, you know, a ticket for going to see you to whatever. And I sell it and I get money and then I give you the money and then you use the money to buy the stuff that you want.

But actually, why don’t you just get that basket of assets yourself? In other words, what’s the point? Since since you can trade them away in liquid markets as well, if you want to. Why doesn’t your eye and my eye just do that? Yeah. So.

So you want some paint? You know, I’ve got some. Oh, I know. I can trade these eggs for that token.

And that’ll get me the paint that Brett wants. That that sounds crazy when it’s people doing it. But, you know, this is nanoseconds.

Because you’re talking if you’re talking about liquid markets for these tokens. In other words, whatever basket I’ve got can be turned into whatever basket you want. And neither of us are involved in it, of course, because it’s boring.

So that’s why the I do that sort of thing for us. Yeah. Well, interesting.

Again, you probably think I’m being very kind of superficial about this, but in that future. Well, this is why I wanted to have the conversation, because. Well, I think you can broadly.

You say financial transactions. I’d say, OK, I can put financial transactions into two buckets, two buckets. There are transactions that are boring, which is most of them.

And there are transactions that are baffling, which are the other ones. So. So for me, paying my car parking is boring.

Right. Me trying to work out which savings account I should have is baffling. Right.

So in both cases, I don’t want to do the boring thing because it’s boring. So I’ll get the AI to do it. I don’t want to do the baffling thing.

I don’t want to do the baffling thing because the idea that. AI will be better at working. Of course.

For 99.9% of people, even the most rudimentary bot could make better choices about savings accounts or whatever than I do. There’s a small number of people. You know, your bot might be.

Actively managing it in real time, every second of every day. And that’s its only thing to do. But even then, the stuff that you get stuck on with your taxes is my butler’s shirt dry cleaning tax deductible or not.

Well, the AI will know that as well. So it’s back. It’s done for you.

Yeah. And it will be thinking about that in advance. And yeah.

So. So in these highly autonomous worlds, when you have, you know, this extreme utility because of these bots and agents operating everywhere seamlessly. It seems at some point.

The more automation and the more. The better value exchange that gets or the more resource efficiency occurs, the less and less you worry about the money in the system, because it’s just that, am I able to do X, Y, Z now? Yes or no. And, and, you know, am I in danger of not being able to do those things in the future? Because I don’t have the assets.

And if over time, it’s so good at creating value from assets and appreciating that and looking after your needs in real time. You could imagine a kind of, you know, we used to argue about Star Trek all the time. Because you love Star Trek.

And I think Star Trek’s hippie rubbish. And this idea of post scarcity makes no sense. But let’s have a communist utopia.

But let’s let’s meet in the middle. OK, nobody in Star Trek has any money. OK.

But the idea that there’s no scarcity is ridiculous because there will always be scarcity. But there was where the gold press latinum was scarce. Or gold press latinum was the thing you couldn’t put through the replicators.

Yes, that’s right. But there’s always going to be things that are scarce. Your attention span, for example, is scarce.

There’s always going to be demand for that. So, so what happens in the middle? The people don’t have any money. They don’t care about money.

That doesn’t mean there’s no asset exchanges going on in the back. There’s no valuation. So the utility is still there, but you may not value it in the same way.

You may not have money as the intermediary. You may just go to having baskets of assets that are moving around. And that, to me, doesn’t seem that implausible.

It sounds futuristic, but I don’t think it is that futuristic. No. So if you then extend that another 20 or 30 years, so the second half of the century.

Then when we’ve optimized resource efficiency, when energy is basically free because of fusion and renewables and so forth. The only limitation on the system is we come back to the limits to growth theory. And, you know, you can’t have infinite growth on a finite planet.

So you need. Well, I’m looking out the window of your apartment, which is fantastic. Right.

Thank you. This is scarce. Yeah.

Right. There’s lots of apartments in this building that don’t have that balcony. Yeah.

So that’s scarce. And that’s not going to change. Like all the, all the buildings in this apartment can’t have that balcony, but they can have an LCD window.

And maybe it looks like a balcony, but it won’t be the same as. So there’s always going to be scarce things. So, so there’s going to be value, but it’s just taking things out of assets that, that, that earn and work for you and turning it into money as we know it now.

And so it doesn’t make any sense. So maybe the scarcity comes from. Well, you know, I think, like, how are people going to choose to live in the future? I mean, are they going to be much more mobile? We’ve obviously seen, we are seeing a shift with millennials and Gen Z’s and alphas to less concern about asset bases and more experiential, you know, the digital nomad experience and things like that.

So you can see a lot of, of humanity moving away from the sort of things that we’ve found valuable in the eighties and nineties, you know, in the post sixties world, which is, you know, you need to invest in real estate and you need to, you know, invest in stocks to now where people are saying, well. Now those things are not as important to me because what they were designed to do is give me access to a lifestyle in the future. When I retired, that was affluent where I could travel and live where I wanted, but now I can do that anyway.

So now let’s invest. Well, this is the argument says when I’m retired, I’m going to need to pay the electricity bill. So I need to put money into assets.

So when I retire, I’ve got money to pay the electricity bill as opposed to why don’t you just put, why don’t you put electricity into your savings account? Yes. Yeah. Yeah.

And, you know, electricity is going to be free in the future, hopefully. Maybe nearly. But I think you’re onto something with this, with this point about the, the, there is definitely, I mean, I’m not an expert on, on, you know, millennials and generational shifts and all that sort of thing.

But even to me, it’s obvious. Yeah. The, the things that I valued when I started my career, like, you know, getting my first company car.

I mean, I remember I was so excited. I get my first company car. Yeah.

The number of kids even bother getting driving licenses. Yeah. It’s so true.

Like now it’s, it’s a really good, like, who wants a company car now? I’ve got to pay for a car park for it. I’d rather give me a Grab account. You know? I think it will be a bit like, so, so my assistant loves riding horses.

So, you know, she has a part share in a horse and N times a week she goes and rides around. Musk has said this. He said owning a car in the future will be like, you know, like owning a horse.

I kind of think that or a cowboy or something. I mean, you, you, you’ll go out at the weekend and maybe drive it around a bit. Actually, you won’t be allowed to drive it in the city because the insurance will be protected.

Well, because humans won’t be able to drive it. If you want to drive yourself, the insurance you’ll have to pay will be so great. It’ll be insane.

Because maybe some students will be much, much better drivers. Of course. And in fact, the figures already show that.

That’s not even speculative. Yeah. So, so I think you’re onto something with this.

Now, um, trying to work out what that means for financial services, I think is very complicated. I did some work for an insurance company recently who were looking at customer acquisition in the, um, in a, in a kind of strategic horizon, which is interesting because in my head, I sort of see insurance companies as being quite conservative, but they were working on a project which was three to five year. So they were saying, how do we make sure in five years time, um, when that kid goes to get their first car insurance, their first travel insurance, their first renter’s insurance, how do, how do we make sure they come to us for that? In order to do that, we have to start working with them now.

So where are those kids now? And because I’m old, I just thought, oh, they’re mucking about on the internet, doing the TikTok shoot yourself in the head challenge or whatever they’re doing this week. But actually, if you look at the numbers, they’re in Minecraft during the fortnight, they’re already in there. Yeah.

I, I invested in, um, Roblox actually. Yeah, Roblox is a good example of it as well. Yeah.

Yeah. Um, all right. So, um, you know, we are getting to that point where we, we need to wrap, but, um, let’s, let’s go on a journey to 2050.

Okay. And presumably much of what we think of as banking infrastructure today will be already highly autonomous, AI run, right? There’s not much of the internal operational aspects of a bank that won’t be on AI in 2050. There will be still some banks running an old legacy systems, but the newest, biggest, fastest, most AI friendly and capable banks will be running, you know, technology infrastructure that’s autonomous.

So if you look at the banking system of the world or the economies of the world, what percentage of it will be automated in the 2050s versus still a human initiated transactions? Do you think? I think the number of human initiated transactions will fall quite rapidly. And again, I, I know there are all sorts of interesting theories about why people want, and in fact, uh, I saw a very good talk by behavioral economist at, um, at, uh, money 2020 about this. Uh, you know, there are some things people will want to do themselves.

And I was just, you know, my example, you know, last weekend, the weekend before I flew, I was at a big war games convention in London. And I was looking at, you know, new paints and different painting techniques. And I was looking at sort of, uh, you know, new 3d printers and stuff.

I don’t want an AI to do that for me. You know, I had a really nice day out. That was fantastic.

Chatting to people. But that’s the experiential. But that’s the experiential stuff.

But everything else that happened on that day. Paying your rent. I couldn’t care less.

Paying for the train fare, paying for the car parking. I don’t care. I just do it.

I just want to get on the train. We’ll just do it. So, so, so the theory that the proportion of transactions will be human initiated.

I think that’s basically true. I have another good reason for thinking this as well, which I think is a little less obvious, which is all to do with financial literacy. Because in many parts of the world, the UK included, there are, there are serious concerns about financial literacy.

Right. And the current sort of approach to that is, is financial literacy campaigns, trying to educate people. Right.

And I’m just wondering if like, there’s no evidence that those campaigns work. No. So maybe just.

Budgeting doesn’t work. Actually, we want bots to do it for people. Yes.

It’s going to be far healthier financially for people to use. I think so. So I agree with you about, so the proportion will be quite low.

What that translates to in the marketing of sales, I’m really not sure. That’s one of the reasons why I’m so interested in it, is because I don’t know. So when you come down to that basic question of, you know, my bot has earned some money from doing something, or maybe I’ve earned some money that the bot’s got, and it’s going to go and park that money somewhere.

So is it going to put it in a savings account or is it going to buy a yield bearing meme coin or is it going to, I don’t know, buy a lottery ticket? I don’t know what it’s going to do with it, you know. How do you persuade it to, like if it’s decided to get a savings account, how does your bank persuade a bot? Because if it’s just going to be price, that’s an absolute race to the bottom. That’s a nightmare.

You’ll end up with, I mean, I’m not saying it won’t happen, but you’ll end up with two or three gigantic banks. Yeah, think about the high frequency trading world, but replicated on every aspect of our economy, where you’re always looking for the optimum vehicle, you know, whether it’s a payment vehicle or a savings vehicle ever, and your AI is switching that, you know, 150 times a day or 10,000 times a day. I put an example.

So I wrote a paper on non-human customers with Kirstie Rutter from Lloyds Bank. It’s in the Journal of Digital Banking. Next month, I think, I can’t remember.

And I used an example that Matt Harris gave me, actually, because it’s a very clear example. And he said, look at car loans. So he said in the US, when interest rates reached their lowest point, only one in 20 people with a car loan refinanced their car loan.

So the average person with a car loan gave $3,500 to the US banking system for no reason. Yeah. And why didn’t they refinance the car loans? Because they had other things to do, basically, you know.

But bots don’t. Bots don’t have other things to do. They don’t have other things to do.

They focus on- Exactly. So that $3,500 vanishes from the banking system. I mean, it doesn’t get redistributed to more efficient producers.

It just goes completely. So how do you function in that space? And I don’t know. Would your AI want faster response times, better uptime? No, but this is- If it’s just price, what’s the point? The point is that once you’ve optimized that price to the nth degree, at a certain point, it won’t be able to be optimized anymore.

But all of this means that the utility of money over time reduces, that value exchange becomes driven by other things, having more time, being able to do the experiences you want to do, decreasing damage to the environment or actually fixing damage to the environment. That’s very interesting because one of the things I’m kind of interested in, because I have no expertise in it at all, is which bot would I choose to- So I don’t want to deal with the bank. I’m going to get a bot to do it for me.

So which bot do I choose? Let’s assume they have to have some kind of basic regulation. You can’t use an unregulated bot. Unregulated bots won’t get access to the bank account.

So I choose a bot that’s regulated by the Monetary Authority of Singapore, and it’s a bot which I’ve chosen because it’s the Greta Thunberg bot and it will optimize for green choices on things, and it’s prepared to sacrifice a certain amount of efficiency. You would choose the Elon Musk, get-me-to-Mars bot as quickly as possible, and other people would make sort of other choices. But why you would choose those things, that becomes fascinating to me.

And again, I stress I don’t know the answer, but that’s why I’m interested in it. Well, I think on that note, it’s a good note to wrap up. What are you working on? Are you going to Mars? Yeah.

Are you working on any new books or anything at the moment? Yeah. Well, I’ve just published a paper on KYA, actually, with Jelena Hoffa from Mastercard. So the digital identity for bot stuff is very interesting.

Link in the description? Because- Oh, you lost the camera. Because for quite a few of the things, quite a few of the businesses that you know I’m involved with, this is a potentially huge new area of business, simply because there’ll be so many more bot transactions than human transactions. And that takes us into the world of micropayments again and so on.

So that’s very interesting to me. I’ve just started writing a new book with Victoria Richardson. Oh, yeah.

Good. You previously partnered with her. Yeah.

Well, it was very successful, so we’ve decided to do that again. And I’m off to Payments Canada the week after next to catch up with a few people out there. So, yes, all right.

It’s interesting. Fantastic. Well, Dave, thanks for joining us on The Futurist and Unbreaking Banks.

And safe travels back to Woking. Thanks for inviting me over.

[shows-menu]