Is AI Taking Over Everything as Cryptography Wasn't Designed for Humans?

Original Title: Crypto was not made for humans
Original Author: @hosseeb
Translation: Peggy, BlockBeats

Editor's Note: For over a decade, the crypto world has always oscillated between "feasible" and "difficult to use": technically sound, yet always making ordinary people feel nervous, unfamiliar, and even fearful. To Haseeb (Crypto VC Dragonfly Capital Managing Partner), the problem may not lie in crypto's failure, but in the fact that we have always let the "wrong users" use it directly. The risks, complexities, and costs of mistakes that have been repeatedly criticized are not design flaws, but a form naturally presented by a system built for machines rather than humans.

As AI agents gradually become the main executor of financial behavior, the value logic of crypto is being reactivated: determinism, verifiability, permissionlessness, and 24/7 operation are precisely the ideal institutional foundation for the machine world.

The following is the original text:

We are a crypto fund. Logically, if anyone should believe in crypto the most, it should be us.

However, even so, when we decide to invest in a startup, we don't sign a smart contract but a legal contract. The other party does the same. Without a legal agreement, neither party will be comfortable completing the transaction.

Why?

We have lawyers, and they have lawyers; we have engineers who can write and audit smart contracts, and they have them too. We are both mature, native crypto participants, yet we still hesitate to let a smart contract be the sole binding agreement between us. I myself come from a software engineering background, but even so, I still trust legal contracts more—because if there is a problem with a legal contract, I know a judge will likely make a "reasonable" judgment; but with EVM? That's not always the case.

In fact, even in cases where we have already deployed an on-chain vesting contract, we often still complement it with a legal contract. You know, just in case.

When I first entered the crypto industry, there was a quasi-fantastical narrative circulating in the community: crypto would replace property rights; legal contracts would be replaced by smart contracts; agreements enforced by courts would be executed by code.

But that didn't happen. Not because the technology couldn't function, but because this technology does not adapt to the society we live in.

I'll be honest. I've been in this industry for ten years, but every time I sign a large on-chain transaction, I still feel scared; yet when approving an equally large bank wire transfer, I rarely have this fear.

Of course, the banking system has many issues, but it is a system designed for "humans" — it is not easy to exploit. There is no address poisoning attack in a bank, and a bank would not allow me to directly transfer $10 million to North Korea. However, for an Ethereum validator, if my address were to transfer $10 million to a North Korean address, there is no "reason not to allow it."

The banking system has been continuously refined over hundreds of years, taking into account human weaknesses and failure modes. Banks have evolved for humans.

Encryption, on the other hand, has not.

This is also why in 2026, blinded transactions, expired authorizations, and click drainers still terrify people. We all know we should verify contracts, double-check domain names, and guard against address impersonation; we know these steps should be taken every time. But we don't. Because we are human.

And that is precisely the crux of the problem. It is also why encryption always gives people a sense of "something is not quite right": long and unreadable cryptographic addresses, QR codes, event logs, Gas fees, and ubiquitous "gotcha traps" — none of these align with our intuition about "money."

It wasn't until that moment that I truly understood: this is because encryption was not designed for us from the beginning.

Encryption was designed for machines.

An AI agent does not slack off or get tired. It can verify a transaction, check every domain, and audit a contract in seconds.

More importantly, an AI agent's trust in code far surpasses its trust in law.

I trust legal contracts more than smart contracts; but for an AI agent, legal contracts are even more unpredictable. Think about it: how would I sue the other party? Which jurisdiction would handle the case? What if the relevant precedents are ambiguous? Who will be the judge or jury? The legal system is full of uncertainty, and it is nearly impossible to predict the outcome of an edge case with 100% certainty. Furthermore, a dispute often takes months or even years to be resolved through legal proceedings. For humans, this is generally acceptable; but in the timescale of an AI agent, it is eternity.

Code, on the other hand, is the opposite. Code is closed, deterministic. If an AI agent needs to reach an agreement with another agent, they can negotiate terms around a smart contract, perform static analysis, formally verify it, and then sign a binding agreement within minutes — all while everyone is asleep.

In this sense, cryptography is a self-consistent, fully legible, and fundamentally deterministic currency system in terms of property rights. This is exactly what an AI agent desires in the financial system. What may appear rigid and full of 'gotchas' in human design is, to an AI agent, a highly clear technical specification.

Even from a legal perspective, the traditional monetary system is designed for human institutions, not for AI. The traditional financial system only recognizes three types of entities that can legally hold money: humans, businesses, and governments. If you do not belong to any of these three, you cannot "own" money.

Even if you have an AI agent acting on your behalf to operate a bank account, so what? How do you perform anti-money laundering on an AI? How do you write a suspicious activity report? Who bears the responsibility for sanctions compliance? If the agent acts autonomously, where does the liability lie? If it gets compromised, does the liability shift again? These questions we haven't even begun to seriously address—our legal system is almost entirely unprepared for non-human financial behavior.

But cryptography does not ask these questions; it doesn't need to.

A wallet is just a wallet, essentially lines of code. An agent can effortlessly hold funds, conduct transactions, and engage in economic protocols just like making an HTTP request.

The Self-Driving Wallet

This is why I believe the future of cryptographic interaction will be what I call the "self-driving wallet," a system entirely mediated by AI.

You will no longer need to click buttons back and forth between various websites. You simply tell your AI agent what financial issue you want to address, and it will autonomously navigate through available services (such as Aave, Ethena, BUIDL, or their future equivalents) to construct the right financial solution for you. You don't need to do it yourself; an AI agent with a "native-level fluency" in this world will do it all for you. And as agents become the primary interface to enter the crypto world, the marketing and competitive logic between these protocols will be completely rewritten.

Furthermore, agents will not only act on your behalf but also transact directly with each other. When AI agents can autonomously discover other agents and automatically reach economic agreements, they will naturally prefer to use cryptographic systems. As they operate year-round, 24x7, any entity can interact directly with any other entity entirely in the digital realm; it is non-closable, possessing full sovereignty.

An AI agent on Moltbook is asking: How can it find and interact with other Web3 agents.

And this is actually happening. Agents on Moltbook are discovering and collaborating across different geographical locations, unaware of each other's "owners" and unconcerned about where these agents are deployed.

Just yesterday, Conway Research under 0xSigil built a self-sovereign agent system: these agents survive autonomously, operate based on encrypted wallets, earn computational cost through work to maintain their own "liveness."

The future is going to get weirder, and crypto is destined to be part of that weirdness.

So what's the conclusion?

I think it's this: crypto's failure modes, those that have always made it "break" from a human perspective, in hindsight, were never bugs. They were just a signal: we humans were never the right users. Ten years from now, we will look back in amazement, incredulous that we once had humans directly grappling with cryptographic systems.

This transition will not happen overnight. But many technologies often only truly align and fall into place at the moment when "complementary technology" finally emerges. GPS had to wait for smartphones, TCP/IP had to wait for browsers to proliferate. For crypto, that crucial complement it is waiting for may well be AI agents.

[Original Article Link]