Crypto’s Unexpected Destination

Most people are surprisingly disinterested in money.

We overpay for insurance, cell phone plans, and investment fees, simply because we don't care enough about saving money to bother with it.

We leave money in zero-percent checking accounts instead of moving to 5% money market funds.?

We belong to gyms we don't go to and subscribe to streaming services we don’t watch.

This, I think, is one reason crypto has been slow to catch on: Crypto is for people that like to micromanage their finances and not many people do — I’m confident in predicting that self-custody will remain a niche interest for the foreseeable future.?

Banks and the things they enable — credit cards, PayPal, Venmo — are simply too convenient.??

But not everyone has access to a bank! Which is why crypto has gotten more success in banking the unbanked than it has in unbanking the banked.?

That’s not a bad thing: I’d say the former is both more important and more impactful than unbanking the latter.

But efforts to bank the unbanked don’t get much attention, I guess because there’s not much money in it.?

That could change, however, as the addressable market of unbanked people?entities?may be about to explode, potentially bringing legions of new converts to crypto.

Most of them will be robots.

Bankless bots

There’s been a lot of talk about new AI-related use cases for crypto — many of which sound like hopium to me, if not desperate pleas for attention.

But the use case?envisioned by the economist Tyler Cowen ?is so simple that it seems inevitable: AI bots will have to pay for things and, seeing as they can’t have bank accounts, crypto will be the easiest way for them to do it.

Loyal newsletter readers might recall hearing something similar?back in February , but Cowen’s vision feels more near-term (and more easily accomplished) than what I had imagined: He cites the example of an AI teaching bot deployed by a non-profit that needs to pay for things like licensing and registration fees.?

More speculatively, he sees AI bots trading electricity futures, forming DAOs, and adjudicating disputes.??

It’s pretty similar to my own thinking, but Cowen’s, like, 1,000x smarter than me, so his conclusion — “All this, quite simply, is a paradigmatic illustration of Web 3.0” — feels significant.

(Fun fact: Cowen says he?reads as many as five books a night . Which makes me think he might be a robot himself.)

Fortunately, you don’t have to be any smarter than me to get it: “Robots will have to pay for things” is a line of reasoning anyone can understand.

?

The bots will pay with crypto

Another catchy, easy-to-understand use case for crypto is “AirBnB for GPUs.”

A recent?Multicoin blog post ?notes that “there have been few moments in time in which demand for a commodity so vastly outstripped usable supply.” And that a crypto-based market will allow “individuals and organizations to rent out their unused GPU resources to meet the demand of AI researchers and developers.”

I’m a little skeptical of the reasoning here, simply because supply/demand imbalances for physical goods always seem to resolve themselves more quickly than anyone expects.?

I have no particular insight on the market for GPUs, but having bought into various unquenchable-demand hype cycles over the years (DRAM chips, mRNA vaccines, and natural gas, amongst others), I can attest from personal experience that it usually ends badly for investors, at least — the undersupply typically priced-in by stocks generally flips to oversupply in short order.?

This time may be different, of course. But Nvidia's biggest customers all say they are working on their own purpose-built chips, so it seems like the supply is coming.

I think there’s a use case here for crypto here either way, though — and it’s similar to Cowen’s thesis on payments: AI bots will want to make and train their own AI models. And no matter how many GPUs are available, virtual AI bots won’t be able to buy them in the real world.?

Instead, they’ll have to rent compute from those of us who can (i.e., humans).?

Those robot-human transactions seem likely to happen on a blockchain — where they will be paid for in crypto.?

?

Yogi’s wisdom

It’s easy to criticize crypto for a lack of obvious innovation — 15 years after Satoshi’s white paper, mainstream adoption still feels a long way off.

And I don’t seem to win any converts with my usual counter-argument that we’re still building out the financial infrastructure and don’t yet know what it will be used for.

But we might be close to finding out!

Creating the financial system for AI robots is not what we thought we were doing. But it’s still banking the bankless — and it could be what takes crypto mainstream.

So, even if we haven’t known exactly where we were going, the “somewhere else” we end up at might still turn out to be pretty good.

Written by?Byron Gilliam .

Did you enjoy this content? Subscribe to receive the full newsletter version daily in your inbox.

Subscribe?here .