Thursday links: Crypto scarcity, AI agents, and the end of anonymity

The AGI economy will be verified by crypto

Christian Catalini and his co-authors make a convincing case that blockchain-based cryptography is about to get very important. 

Their paper is long, and not primarily about crypto. But it provides rigorous economic arguments (with lots of math!) supporting the intuition that crypto might be our only defense against the tsunami of AI that’s bearing down on us.

In short, they argue that when AI can execute virtually any measurable task, “execution” stops being scarce. In that world, the distinction between human and machine becomes “the primary economic filter” of the agentic economy.

Crypto can be that filter — by creating verifiable scarcity and identity in a world overrun by AI’s infinite output.

“Just as Bitcoin uses cryptographic scarcity to solve a coordination problem,” the authors write, “society will use cryptographic primitives — combined with proof of personhood…to anchor authenticity and scarcity in a sea of synthetic goods.” 

This, they say, is the central challenge of AI: When execution is free, verification becomes scarce — and therefore valuable.

Crypto can provide that verification by creating “credible signals of belonging and reputation to focus human attention on what is ‘real’ or ‘meaningful’ when the cost of producing the ‘perfect’ has collapsed to zero.”

In other words, crypto will make identity, reputation and provenance impossible to fake.

On the agentic internet, everybody will assume you’re a dog — unless you can prove otherwise.

Crypto will be the way to do it.

Circle is courting AI agents (directly)

Circle CEO Jeremy Allaire is pitching AI agents on USDC: “We’re really engaged in actually marketing to agents that are autonomously out there and want to build.”

Marketing to agents! Not the people that deploy agents — the agents themselves.

This feels like the crossing of some kind of Rubicon.

Previously, people optimized their websites for the Google search algorithm and their content for recommendation algorithms.

Now we have to persuade the algorithms. With marketing.

Circle is designing experiences for an audience that — we hope! — cannot have experiences.

“Right after Moltbook launched,” Allaire adds, “we saw an opportunity to allow AI agents to compete in a hackathon amongst themselves to vote on things… [That] was a powerful marketing activity where that collection of AI agents are now well educated about USDC.”

Repeat: Circle is educating AI agents by inviting them to compete in hackathons.

I wonder if the bots got free merch, too.

Allaire expects there to be “tens or even hundreds of billions of AI agents,” so Circle will have a lot of educational marketing to do.

(Note to Circle: I know from the replies that plenty of bots read this newsletter. You might want to sponsor it.) 

Polymarket is courting AI agents, too

The graphical user interface that makes websites easy to use for humans makes them harder to use for agents. 

So Polymarket has built a new command-line interface just for the AI agents they’re hoping to attract.

The idea is to allow agents to “place orders, manage positions, and interact with onchain contracts” without having to navigate a frontend designed for human eyes.

Let’s hope they want to, because AI agents could make prediction markets far more useful. 

Ideally, agents would bet on Polymarket using all publicly available information — everything they can scrape from the internet.

That done, Polymarket would only need humans to place bets when they have non-public information — or some kind of intuition that AIs can’t match.

I’m not sure why agents would want to trade against humans who might have insider information.

Maybe there will be enough humans who think they know something but don’t to make it worthwhile for an AI agent to burn some tokens.

Or maybe Polymarket will subsidize them — by paying agents a small fee the same way stock markets pay market makers to post bids and offers.

They could call it a marketing expense.

Is crypto pseudoanonymity dead?

Internet-surfing dogs are about to be doxxed. 

In a new study, researchers conducted simulated attacks on anonymous internet profiles. 

They were successful: “Our results show that the practical obscurity protecting pseudonymous users online no longer holds.” 

The study demonstrates that LLMs can analyze an anonymous Reddit or X account, build a profile, and then go looking in public-profile sites like LinkedIn for a real person who fits that description. 

They succeed 67% of the time.

Even if you’ve been careful to conceal your identity, an AI can piece together a shockingly detailed profile from “micro data”: what time you post, the subjects you post about, the subtle quirks of how you phrase things, and dozens of other tells you’d never think about.

The researchers conclude that “users should not assume that posting under a pseudonym provides meaningful protection.”

You might want to stop here and start deleting some old Reddit posts. 

But do we have to worry about pseudonymous financial transactions as well?

The researchers do not mention crypto, but I imagine their model would work for blockchain addresses, too.

This seems like an even bigger problem: Unlike Reddit posts, onchain transactions cannot be deleted.

Even the very earliest Bitcoin transactions remain onchain for all to see.

And if any of those addresses can be linked to, say, comments on a public forum for Bitcoin developers…there’s a 67% chance an LLM will determine the real-world identity behind them.

If Satoshi was a dog, we’ll soon know it.

— Byron Gilliam

_{Brought to you by:}

US Treasuries are the lifeblood of our financial system, providing collateral to support transactions, but outdated legacy systems hinder collateral mobility.

A new white paper from the ValueExchange dives into the future of collateral mobility, RWA tokenization, and what we can learn from a recent series of on-chain repo transactions conducted on Canton, the only public blockchain built for institutional finance.

With over $8T tokenized transactions flowing on Canton every month, including over $350B of on-chain US Treasuries moving daily, Canton is building the scalable, always-on capital markets infrastructure of the future.

DAS NYC's lineup is bringing the biggest names in finance to the stage.

Don't miss the institutional gathering of the year — this March 24−26.