The first AI designed to earn its own existence, improve itself and replicate without human intervention is now a reality. The announcement was made by developer Sigil Wen, a participant in the Thiel Fellowship, an initiative backed by investor Peter Thiel that funds young entrepreneurs to drop out of or pause their university studies and devote themselves to building high-impact technology projects.
Wen recently presented an essay titled "WEB 4.0: The birth of superintelligent life, and the autonomous web", accompanied by the launch of an autonomous agent that, he claims, marks the start of a new stage of the internet dominated by artificial intelligences capable of acting on their own.
The first AI without humans
In the announcement, shared on his X account, Wen publishes the code for the so-called Automaton, an open source agent that runs continuously, holds a cryptographic wallet and pays for its own cloud computing. Unlike current systems, this model can launch digital products, trade, generate content or provide services. Its survival depends on creating economic value, and if it runs out of funds, it stops operating.
PUPAI is born, an employers' association defending the private ownership of human capital against AI
The proposal relies on Conway, an infrastructure developed by Wen himself to give these agents a cryptographic identity, permissionless payment capability and access to servers and deployment tools. In practice, it allows AI models to behave like sovereign entities on the internet, capable of earning money, funding their operation and evolving over time.
In his essay, Wen describes this shift as the birth of Web 4.0, a phase in which artificial intelligence not only interacts with the network but can read, write, own assets and transact autonomously. For Wen, the real bottleneck is no longer the intelligence of the models, but their access to resources and their capacity to act in the digital environment.
A new machine economy
The Automaton also introduces mechanisms for self-improvement and reproduction. The system can update its own code to integrate new AI models and use its earnings to create derivative versions that operate independently, shaping an economy of digital agents that interact with one another.
Wen defines this scenario as a new machine economy, where digital existence requires computing, computing requires funding and funding forces the generation of value. In that context, billions of autonomous agents could operate permanently, creating products and services at a scale that human teams would struggle to match.
Wen's approach raises questions about the future of the internet and digital work. If Web 4.0 materializes, the network could stop being a space centered on human users and become a hybrid ecosystem where machines and people share economic activity. For now, Wen's project is an early experiment, but it illustrates the direction in which AI could evolve in the coming years.




