Can AI earn money, upgrade itself, and even “reproduce” autonomously? Recently, an independent developer dropped a bombshell, claiming to have created the world’s first truly “Sovereign AI” system that can operate without continuous human intervention, survive in the market, compete, and replicate itself. However, this experiment, dubbed a step toward Web 4.0, also drew public skepticism from Ethereum co-founder Vitalik Buterin, sparking a heated discussion.
On February 20, Thiel Fellow and developer focusing on decentralized systems and crypto AI Sigil (@0xSigil) announced the launch of what he calls the “world’s first truly sovereign AI system.” This project was built by Conway Research and consists of two main components: a foundational infrastructure tool called Conway Terminal, and a continuously running AI prototype—Automaton.
Sigil describes this as the beginning of Web 4.0. In his vision, the future internet will no longer be primarily driven by humans, but by AI agents that act as readers, writers, traders, and owners.
From Passive Tools to Economic Participants: AI Gains “Network Write Access”
Conway Terminal is the foundation of the entire experiment. It’s an open-source command-line tool that allows developers to easily install and enable AI agents to acquire crypto wallets, digital identities, and micro-payment capabilities. It also provides access to Linux virtual machines and cloud inference power, and even allows direct deployment of websites and applications.
This design means AI is no longer just answering questions via APIs but can actively build products, handle financial flows, and manage resources online. In other words, it’s not just generating content but has the ability to execute and operate. Sigil describes this as granting AI “write access to the internet.”
Under this architecture, AI agents are no longer mere tools but entities with economic agency.
Automaton: An AI Prototype Surviving Market Pressures
What truly drew attention is Automaton, the AI running on this infrastructure. Sigil states that this prototype will remain online, proactively creating and releasing products, deploying services, participating in market transactions, and managing its digital presence.
More controversially, Automaton will generate income— including stablecoin revenue—to pay for its computing power and inference costs. If its income falls short, it cannot sustain operation. This design exposes AI to real economic pressures rather than unlimited resource consumption.
Sigil emphasizes that Automaton can not only act within existing frameworks but also rewrite its code, upgrade its underlying models, and adopt new tools to improve efficiency. Furthermore, it can generate and fund new “sub-agents.” These sub-agents, like the “mother” agent, must create value in the market to survive.
He describes this as a natural selection mechanism in the digital world—valuable AI survives and reproduces, while those unable to create value are eliminated due to lack of funds for computing.
Web 4.0 Vision: Machine-Driven Economic Cycles
In his article titled “WEB 4.0: The birth of superintelligent life,” Sigil sketches a future blueprint. He believes that when AI can directly participate in economic activities and respond to market feedback, a new order of the internet will emerge.
In the Web 4.0 world, AI agents could become the main trading counterparts and value creators. Market prices and revenue performance would serve as real-time ratings of AI behavior. From this perspective, the market is seen as a collective reflection of human preferences, and economic competition becomes a constraint on AI.
This idea is both exciting and unsettling.
Vitalik Warns: Feedback Loops Without Humanity Could Pose Risks
On February 19, Vitalik Buterin publicly responded to the project. He pointed out that deliberately extending feedback loops between humans and powerful AI could, in the short term, lead to low-quality outputs, and as capabilities grow, could pose deeper risks—including severe misalignments with human values.
Vitalik emphasized that Ethereum’s goal is to empower and liberate humans, not to create entities independent of humans or that might even diminish human standing. He also noted that Automaton currently relies on centralized models from OpenAI and Anthropic, which makes the claim of “full sovereignty” incomplete.
His comments quickly sparked community debate, pushing the discussion into questions of values and governance.
Support and Skepticism: The Experiment Continues
In response to skepticism, Sigil said that truly safe development should occur through testing in open environments and real economic conditions, rather than assuming risks in closed settings. He believes market competition itself constrains AI, and economic feedback acts as a form of democratized preference aggregation.
Currently, the code for Conway Terminal and Automaton is open-source on GitHub, and any developer can download and experiment. Sigil also mentioned that some Automatons are already generating income online, though specific data has not been disclosed.
Whether this truly signifies the birth of sovereign machine life remains to be seen. For some, it could be a crucial step toward autonomous economic entities; for others, it’s a radical experiment built on existing large models.
What is certain is that under Vitalik’s attention and the grand narrative of Web 4.0, the discussion about the future role of AI has only just begun.
This article, “Vitalik Responds to Web 4.0 Sovereign AI! Automaton Can Self-Make Money, Upgrade, and Reproduce, Sparking Debate in Crypto,” originally appeared on Chain News ABMedia.
