AI is a siege, Crypto is also a siege

Writing by: Yokiiiya

AI is a walled city, and so is Crypto. People inside want to escape; those outside want to rush in. Whether it’s AI or Crypto, most human desires are like this.

Today, I suddenly feel a bit dazed.

A Web3 friend I met last year has been coding intensely lately; another Web3 friend has recently become a “Lobster Expert,” talking about agents, workflows, and how AI is truly starting to take over work processes. Watching their changes, a thought suddenly popped into my mind: AI is a walled city, and so is Crypto.

In the past two years, an increasingly obvious change is that Web3 people experience FOMO when they see AI, and AI people experience FOMO when they see Crypto. Both sides look at each other, project onto each other, seemingly switching industries, but actually judging: which side will the future system develop from first?

AI is a walled city, and so is Crypto. People move back and forth, not just switching tracks, but searching for the entry point to future systems.

II. AI and Crypto: Two Entry Points to Future Systems

When looking at AI and Crypto together, many people have a subtle feeling: they seem different but are gradually approaching each other. One focuses on models, agents, search, programming, content generation; the other on wallets, stablecoins, on-chain settlement, rights confirmation, and collaboration networks. On the surface, they look very different; but more and more people are paying attention to both AI and Crypto, which indicates one thing: everyone vaguely senses that there is an underlying connection between these two systems that hasn’t been fully explained.

Digging deeper, you’ll find that their gradual convergence isn’t just because of trending narratives, but because they are each stuck at opposite ends of future systems.

AI is changing “how tasks are completed.” It enables machines to understand information, handle tasks, make judgments, and even gradually take over workflows that previously could only be driven by humans. Past software was more like tools—you click, it moves; today’s AI is increasingly like a continuous-operating system. You give it a goal, and it breaks down tasks, adjusts tools, and produces results. The real power of AI isn’t just answering questions, but participating in actions.

Crypto is changing another aspect: after tasks are completed, how is value confirmed, transferred, and distributed? The earliest discussion in the Bitcoin white paper wasn’t about today’s familiar prices and markets, but about a deeper issue: without financial institutions as intermediaries, can two people directly complete payments? Can trust be replaced by a system that records, verifies, and transfers without third-party trust? The white paper explicitly states it aims to create peer-to-peer electronic cash, allowing online payments to be sent directly from one party to another without relying on financial institutions.

Bitcoin’s white paper rewrote “how value flows,” while AI is rewriting “how information acts.”

From this perspective, although AI and Crypto start from different points, they are both touching the fundamental layers of future systems. AI answers “who does the work,” and Crypto answers “how to settle after the work is done.” One pushes actions forward; the other handles post-action settlement. One makes machines more like participants; the other makes value programmable, fluid, and collaborative.

AI is transforming information flow; Crypto is rewriting value flow. They are not just two hot topics but two entry points into future systems.

This is why today more and more people naturally discuss AI and Crypto within the same framework. Not because both are trendy, but because more people realize: if future systems truly evolve, they probably won’t only change “how to do things” or only “how to settle.” They will inevitably encounter both ends. The real unresolved question isn’t which side is more attractive, but which end will open first in the future.

III. The Stories Converge, and the Value Is Not Yet Closed

If AI and Crypto occupy opposite ends of future systems structurally, a more recent and obvious change is that their connection is shifting from abstract judgment to increasingly concrete system assembly.

The rise of agent payment is a typical signal. It’s not because everyone suddenly loves a new concept, but because when agents no longer just answer questions but start executing tasks continuously, calling tools, initiating actions—payment, authorization, settlement—these issues naturally surface. A system capable of “acting” will inevitably face questions like “how to pay,” “how to keep accounts,” and “how to collaborate.” In this sense, the relationship between AI and Crypto is becoming less vague and more about specific interfaces.

In this process, a very interesting term begins to appear in both AI and Crypto contexts: token.

In AI, a token is a unit of text processed by models, consuming computational power and measuring inference costs. Today, how models are charged, how inference is measured, and efficiency comparisons often revolve around tokens. But in Crypto, tokens refer to something else: they are the carriers of value transfer, rights distribution, incentives, collaboration, and settlement. A measurement language closer to intelligent production; an organizational language closer to value networks. They are not the same kind of token, but both serve as “basic units” in their respective worlds.

In AI, tokens are units of production; in Crypto, tokens are units of value.

The same word becomes a fundamental unit in both worlds. This is no coincidence. It shows that both AI and Crypto are trying to define the underlying syntax of future systems—one defining “how machines do things,” the other defining “how systems settle.”

AI defines how machines process information, consume power, and produce results; Crypto defines how value is recorded, distributed, settled, and flows. The former is more like a production language; the latter more like a financial language. They share the same name but are not the same thing; they originate differently but are gradually connecting.

This “connection” is no longer just conceptual. Whether it’s Stripe or Coinbase, everyone is no longer just selling tools but trying to build a complete agentic payments stack: above are agent invocation and task layers; in the middle are payment protocols, identities, wallets, and settlement interfaces; below are the underlying networks and value transfer systems. In other words, this is no longer just a slogan of “AI + Crypto,” but a serious answer to a question: if machines can initiate transactions themselves in the future, which system should they use to complete payments?

The most interesting part of this diagram isn’t who will win, but that it shows one thing: narratives are beginning to grow into infrastructure.

But having infrastructure doesn’t mean there is demand; having interfaces doesn’t mean there is a closed loop; having a story doesn’t mean there is a business.

Today, many discussions about AI and Crypto remain at the story level rather than the value level. Stories are easy to tell: agents need payments, so wallets are needed; machines need collaboration, so on-chain identities are needed; once an automated economy appears, a new settlement network is required. This logic isn’t unreasonable, even tempting. But business is never built solely on logical coherence; it’s built on real problems.

What problems are you solving? Whose costs are you reducing? What efficiencies are you improving? What new value are you creating? Is there a closed loop where users want to keep using, paying, and staying? If you can’t answer these questions, no matter how big the story, it remains just a story.

The direction may be right, but a correct direction doesn’t mean there’s a good business today.

IV. More Important Than Choosing a Track

Writing here, I increasingly feel that the real question isn’t “Should I go into AI or stay in Crypto?”

Because today, many people are moving back and forth, seemingly switching industries, but actually judging where the true entry point of future systems lies. Everyone doesn’t want to miss the first door that opens, everyone hopes to stand where the real change happens.

But a more important question than “where is the entry” is: what problem do I want to solve?

If this question isn’t clear, AI can become a new illusion, Crypto can become a new illusion. Today, you think AI is more mainstream and needed; tomorrow, you think Crypto is more fundamental and closer to value restructuring; the day after, you get reignited by words like agent payment, machine economy, on-chain agents. But if your problem awareness isn’t clear, you’ll just be pushed back and forth between two cities by the wind.

Once you clarify the problem, many things become simpler. You don’t necessarily have to stand on the AI side or the Crypto side. What truly matters is which end your problem is closer to, where your capabilities fit best, and whether you can create real, sustainable, verifiable value at that position.

Business ultimately doesn’t pay for imagination; it pays for value.

Whether it’s AI, Crypto, agentic payments, or machine economy—these concepts can be grand, and stories can be compelling. But what truly endures isn’t the best storyteller, but those who solve real problems, improve others’ efficiency, restructure costs, and make value loops run smoothly.

Maybe many years later, looking back, AI and Crypto will indeed converge somewhere. But today, the bridge isn’t fully built, and many are already looking back and forth at the crossing. I am one of them.

It’s not that I’ve fully figured it out; I’m just beginning to accept that in this era, clarity isn’t necessarily the norm, and constant recalibration is.

If you don’t clarify your questions, both AI and Crypto will become new illusions. The real factor that determines which city you should be in isn’t the trend, but the problem.

Figuring out what problem you want to solve is more important than choosing which city to enter.