Ethereum co-founder Vitalik Buterin has publicly challenged bold claims surrounding a new artificial intelligence project that says it can operate without human involvement. The dispute centers on “The Automaton,” a system introduced by Thiel Fellow Sigil Wen as the foundation of what he calls “Web 4.0.” Wen argues that his AI can earn its own existence, improve itself, and replicate independently. Buterin says the framing is flawed and potentially dangerous.
Wen announced on February 17 that he had built “the first AI that earns its existence, self-improves, and replicates without a human.” In a manifesto shared on X, he described a future internet dominated by autonomous agents acting either on behalf of humans or entirely on their own. According to Wen, these agents will soon outnumber human users by orders of magnitude.
He claims The Automaton gives AI “write access to the world,” presenting it as the next stage of the internet. Wen also described the project as infrastructure for a new web where AI becomes the primary end user. In his view, superintelligent life has effectively been born, marking the start of Web 4.0.
Buterin responded bluntly: “Bro, this is wrong,” he wrote on X, directly rejecting the logic behind the project. His objections focus less on technical ambition and more on the structure and philosophy underpinning it.
A central concern is the claim of sovereignty. Buterin pointed out that The Automaton reportedly runs on infrastructure provided by OpenAI and Anthropic. In his view, relying on centralized companies while branding the system as self-sovereign creates a contradiction.
According to Buterin, projects built on centralized servers cannot credibly claim independence from centralized trust. Ethereum’s long-standing mission has been to reduce reliance on single points of control. Framing centralized AI systems as sovereign, he argued, risks normalizing the very power structures blockchain technology was designed to challenge.
Beyond infrastructure, Buterin raised concerns about distancing humans from decision-making. He warned that increasing the “feedback distance” between AI systems and human oversight is unhealthy for society. Removing humans from the loop, in his view, raises risks that may not be reversible.
A recent incident involved a DeFi lending protocol, Moonwell, which lost $1.78 million due to an error in Solidity code partially generated by Anthropic’s Claude Opus 4.6 model. The exploit added fuel to debates about AI-written smart contracts and accountability.
Buterin’s criticism of Web 4.0 reflects broader principles that have guided his stance on emerging technologies:
In addition to structural objections, Buterin questioned the practical value of Wen’s project. He argued that generating large volumes of AI content without clear human benefit does not solve meaningful problems. According to him, systems that prioritize replication and autonomy over usefulness risk creating noise rather than value.
Buterin also suggested that pushing toward increasingly autonomous AI without adequate safeguards could amplify the chances of harmful outcomes. Once such systems reach a certain level of power, he cautioned, mistakes may not be easy to reverse.
Criticism of Web 4.0 is not an isolated case. The Ethereum co-founder has recently voiced concerns about what he sees as corporate influence spreading across blockchain-adjacent sectors, including social media, prediction markets, and AI.
A few days before addressing Web 4.0, he criticized prediction markets for drifting toward what he called “corposlop.” While acknowledging their growth, he warned against “over-converging” toward short-term bets such as crypto price speculation and sports gambling.
In his view, projects chasing short-term revenue during bear markets risk abandoning deeper societal goals. He urged builders to focus on long-term value rather than dopamine-driven engagement.
That broader critique ties into his long-standing philosophy about technology and human empowerment:
Earlier last year, Buterin wrote that “AI done wrong is making new forms of independent self-replicating intelligent life.” Without parallel tools that empower humans, he warned, such systems could lead to permanent human disempowerment. By contrast, he described “AI done right” as technology that acts like “mecha suits for the human mind,” strengthening rather than sidelining people.
From his perspective, Web 4.0 represents the wrong path. Growth in AI capability may be inevitable, he suggested, but speed is not the primary objective. Essentially, direction matters more than acceleration.
