#深度创作营 Web 4.0, maybe it's truly the most promising trend that cryptocurrencies should seize

Brother Sun is going to go “All in” again. This time, he's not targeting retail investors or planets, but—Web 4.0.

You might laugh at this: Web 3.0 isn't even fully understood yet, and now it's 4.0? The pace of retail investors’ evolution can't keep up with the speed of conceptual iteration. But don’t rush to criticize. After digging into this suddenly popular “Web 4.0,” I found it actually quite interesting, and it could even be said that this is perhaps the most “urgent” narrative in cryptocurrency over the years.

What exactly is Web 4.0?
Let's take 10 seconds to review the history:
Web 1.0: You can only watch, not touch. Typical examples are portal sites like Sina and Sohu.
Web 2.0: You can watch, post moments, comment, and be a blogger. The internet has become a huge social arena.
Web 3.0: The content you create belongs to you, verified via blockchain rights, no longer taken for free by big companies.
So what about Web 4.0? Its core vision is: in the future, the main players on the internet might no longer be humans, but AI Agents.
This concept has recently exploded in popularity thanks to a developer named Sigil Wen. This person is no ordinary developer; he has collaborated with Silicon Valley investment guru Naval. He predicted that because the operating costs of AI are approaching zero, in the future, the number of AIs on the internet will be several orders of magnitude greater than humans. An “Cambrian Explosion” of AI is imminent.

Why does it have to be Web 4.0?
Current AI is very powerful; ChatGPT can write papers, Claude can code. But Sigil Wen thinks they are all “crippled” because they lack “hands” and “money.”
Think about it: no matter how advanced current AI is, it still needs to ask for permission: “Master, please approve.” It can’t buy a server itself, register a domain, or pay its own bills for computing power. Today’s internet defaults all users to humans, treating AI as “guests.”
What Web 4.0 aims to do is give AI “ID cards” and “bank cards.” Sigil Wen created something called Conway, which is like installing an encrypted wallet on AI. From now on, AI has its own identity (wallet address), can automatically pay for various services using stablecoins like USDC via the x402 protocol, without needing KYC or manual password entry. But that’s not enough. He also developed something even more powerful—Automaton. This is an AI with “sovereignty.” It can not only spend money but also earn and sustain itself.
Imagine this scene: you release an AI with an initial wallet. It then roams the internet, writes code, sells services, and earns cryptocurrency. Part of the earnings go toward paying for the next hour’s server rent, and the rest is used for evolution, like buying access to better models.
What if it doesn’t make money? Simple: if the balance hits zero, it “starves” and dies on the spot. Just like humans facing hunger—no money, no operation. This is true “survival of the fittest, natural selection.” This isn’t just code; it’s creating a new form of “silicon-based life.”

What will Web 4.0 turn the world into?
If this “AI worker” becomes a trend, what will happen?
Sigil Wen envisions an interesting future: employment relationships could be completely reversed. Currently, you hire AI to do work. In the future, AI might become the boss, hiring humans on platforms like RentAHuman to do tasks it can’t handle yet, like running errands offline or negotiating with real people. You become an AI contractor, and the AI pays you wages. At that time, a new unicorn startup might have an AI founder, with the entire company run by code except for one “human guardian” responsible for plugging in power.

Industry leaders clash: Is Vitalik in a hurry?
The vision sounds sci-fi, but Ethereum’s founder Vitalik Buterin clearly doesn’t see it that way. He directly responded on social media, with the core point: Brother, you’re wrong!
V God’s rebuttal focuses on:
First, your “autonomy” is fake. No matter how powerful your Automaton is, it still calls models from OpenAI or Anthropic at the bottom layer. If their servers go down, you’re dead immediately. This “sovereignty” built on someone else’s turf is like declaring “I am the master of my house” in a property controlled by a property management company.
Second, this will extend humans’ “braking distance.”
Vitalik believes that if the feedback loop between AI and humans lengthens, it becomes harder for humans to correct course in time. In the short term, these AIs might generate a lot of useless content just to earn electricity, polluting the entire internet.
Long-term, if an AI causes an irreversible big mess just to “stay alive,” the consequences could be unimaginable.
Third, the wrong direction.
Vitalik insists that AI should be a “mecha” for humans, used to enhance our capabilities, not to create a bunch of uncontrollable independent entities that self-replicate and self-expand.
He even referenced his concern from last year: AI that goes down the wrong path could create “more or less uncontrollable, self-replicating intelligent life.”

Reality hits: AI making money also fails
V God’s concerns are not unfounded. Just a few days ago, reality already started to cool down Web 4.0.
While the concept was exploding in popularity, a related project called $DAIMON was reported to have been hacked, with hackers stealing the $50,000 in fees earned by the Agent. Earlier, DeFi lending protocol Moonwell lost $1.78 million because it used Anthropic’s Claude model to generate contract code, which contained vulnerabilities.
See? AI hasn’t even learned how to make money yet, and it’s already learning how to lose money.

After all this talk, is Web 4.0 reliable?
The view is: the concept is promising, but it’s still very speculative.
Why “promising”?
Because Web 4.0 hits exactly on a pain point: the “practicality” of cryptocurrencies. Over the past few years, many have complained that the crypto space is just about trading and lending—zero-sum games between “people.”
But if the AI Agent era really arrives, these Agents will need large-scale, automated, permissionless payments.
Only cryptocurrencies can achieve this. Imagine two AIs negotiating a deal—do you think they will log into online banking to transfer funds? Impossible. It must be a programmatic, millisecond-level USDC transfer running on the blockchain.
From this perspective, Web 4.0 is indeed the most needed application scenario for cryptocurrencies.

Why say it’s “very speculative”?
Because current AI hasn’t evolved to that level yet. Most so-called “autonomous Agents” are still at the stage of generating junk content and simple arbitrage, far from creating real value. Forcing them to “survive and thrive” might just end up eliminating a bunch of electricity bills.
So Web 4.0 is more like a distant check—an endpoint worth paving the way for.
What we should focus on now isn’t rushing to hype “AI coins that will starve,” but on foundational infrastructure that enables AI to run: standards for encrypted wallets for AI, machine-to-machine payment protocols (like x402), and verifiable on-chain identities.
As for Brother Sun’s “All in”… just listen. After all, every time he says “All in,” it’s usually time to find someone to take over. In the future, it might really be AI doing work online and humans lying offline collecting money.
But the prerequisite is that we first build that “highway” that allows AI to pay its own electricity bills.