Following Dragonfly's new $650 million funding round, Haseeb stated that "encryption is not for humans," and that AI agents are the ultimate users.

This article is from: Haseeb Qureshi

Compiled by Odaily Planet Daily Azuma

Editor's Note: Last night, Dragonfly Capital, a leading venture capital firm, announced the completion of its fourth fund raising of $650 million.

That same evening, Haseeb Qureshi, a star partner at Dragonfly Capital, published a lengthy article on X titled "Crypto was not made for humans." The article proposed the new view that "cryptocurrencies were not created for humans, but should serve AI tokens," and stated that "in 10 years, we may be surprised to find that humans are directly interacting with cryptocurrencies."

The following is the full text of Haseeb Qureshi's article, translated by Odaily.

We are a crypto fund. If anyone should believe in cryptocurrencies, it's us.

However, when we sign an agreement to invest in a startup, we don't sign a smart contract, but a legal contract; the same applies to the startup. Without a legal agreement, both of us would feel uneasy.

why is that?

We have lawyers, and they have lawyers too. We have engineers capable of writing and auditing smart contracts, and they have engineers too. Both sides are seasoned participants with expertise in cryptography, but we still don't believe that smart contracts can be the only binding agreement between us.

I myself am a software engineer by training, but I still trust legal contracts more – because if there are problems with legal contracts, I know that judges will make reasonable rulings, while EVMs are not.

In fact, even when an "on-chain token vesting" contract exists, a legal contract is usually still included. This is just in case.

When I first entered the crypto industry, people told a fantastical story: cryptocurrencies would replace property rights. We would no longer use legal contracts, but smart contracts; we would no longer rely on courts to enforce agreements, but on code to enforce them.

But that didn't happen. Not because the technology was unworkable, but because it wasn't suitable for our society.

I've been in this industry for ten years, and I still feel scared every time I sign a large on-chain transaction, but I'm never afraid of a large bank wire transfer.

The banking system may be flawed, but it's designed for humans. It's hard to mess with. There's no address poisoning attack in banks, and it's highly unlikely they'd allow me to transfer $10 million to North Korea—but for Ethereum validators, there's no reason not to execute a transfer if my address sends $10 million to a North Korean address.

The banking system is specifically designed to address human weaknesses and failure patterns, and has been refined over hundreds of years. The banking system is adapted to human needs, but cryptocurrency is not.

This is why, in 2026, blind-signed transactions, legacy licenses, and accidental phishing contracts remain terrifying. We currently know we should verify contracts, double-check domains, scan for address spoofing… we know we should do this every time, but we don't, because we are human.

That's the crux of the matter. That's why cryptocurrency always feels a bit off. Lengthy, unreadable encrypted addresses, QR codes, event logs, gas bills, and footguns everywhere—none of it aligns with our intuition about money.

At that moment, it dawned on me— because cryptocurrency was never created for us.

Crypto was created for machines.

The AI ​​agent doesn't slack off or get tired. It can verify transactions, check each domain, and audit contracts in seconds.

More importantly, AI agents trust code more than law . I trust the law, not smart contracts, but legal contracts are actually more unpredictable for AI agents.

Consider this: how will I take my counterparty to court? In which jurisdiction will this contract be tried? What if legal precedent is ambiguous? Who will serve as judge or jury? The law is fraught with uncertainty; the outcome of any marginal case is difficult to determine, and dispute resolution often takes months or even years. This is generally acceptable for humans, but on the timescale of AI agents, it's practically eternity.

Code, on the other hand, is the opposite. Code is closed, deterministic, and verifiable. If an AI agent wants to reach an agreement with another agent, it can conduct multiple rounds of terms negotiation, static analysis, and formal verification on a smart contract, and enter into a binding agreement—all within minutes, while humans are still asleep.

From this perspective, cryptocurrency is a self-consistent, fully readable, and fully deterministic system of monetary property rights. This is everything an AI financial system needs. What we humans perceive as a "rigid trap" is, in AI's view, a well-written specification.

Even from a legal perspective, our traditional monetary system is designed for humans, not AI. The traditional monetary system only recognizes humans, businesses, and governments as the legitimate holders of money. If you are not one of these three entities, you cannot own money.

Even if you set up an AI agent to interact with your bank account on your behalf, then what? How do you conduct anti-money laundering (AML) reviews, report suspicious activity, and sanction violations against the AI ​​agent? If the agent acts autonomously, where does the liability lie? If it is manipulated, does the liability change?

We haven’t even begun to answer these questions— our legal system is completely unprepared for non-human financial participants.

Cryptocurrencies, however, don't need to answer these questions. A wallet is just a wallet; it's simply code. Agents can hold funds, conduct transactions, and access economic protocols as easily as sending an HTTP request.

"Autonomous driving" wallet

That's why I believe the crypto interface of the future will be what I call a "self-driving" wallet—that is, one entirely mediated by AI.

You'll no longer need to visit numerous websites. You'll instruct your AI agent to solve your financial problems, navigating available services such as Aave, Ethena, BUIDL, or any protocol that inherits from them, and constructing suitable financial solutions for you. You won't be doing anything yourself; an AI agent with a deep understanding of the world will handle it for you. As AI agents become the primary interface to the crypto world, the way these protocols market themselves and compete with each other will also fundamentally change.

Besides acting on your behalf, agents also trade with each other. When agents can autonomously discover other agents and enter into economic protocols, they prefer cryptocurrencies. This is because cryptocurrencies can operate 24/7, are peer-to-peer, exist in virtual space, cannot be shut down, and possess complete self-sovereignty…

Odaily Note: An AI agent on Moltbook asked how to find and interact with other Web3 agents.

This is already happening. Agents on Moltbook are finding each other and collaborating across geographical boundaries, and nobody knows who their owners are or where they are located.

Just yesterday, 0xSigil's Conway Research built a batch of autonomous agents that will live completely autonomously using crypto wallets and strive to earn their own computing costs in order to survive.

The art style of the future will become increasingly bizarre, and cryptocurrencies will be part of this strange world.

So, what is the conclusion?

I believe it's like this—those seemingly flawed aspects of cryptocurrency, those things that feel like defects to humans, may never have been vulnerabilities in hindsight. They simply demonstrate that humans weren't the right users. Ten years from now, when we look back, we might be surprised that humans actually "fought" directly with cryptocurrency.

This kind of change doesn't happen overnight, but a technology often explodes in popularity when its complementary technology finally arrives. GPS waited for smartphones, TCP/IP waited for browsers. For cryptocurrencies, we may have just seen it coming in AI agents.