The story of the internet is often told as a series of expanding capabilities. Web1, the read-only era of the 1990s, let people access static web pages but not interact with or change them. Web2, starting in the mid-2000s, introduced the ability to write and share — think social media posts, videos, and comments — but in doing so, platforms took ownership of users' data, content, and social connections in exchange for providing the service.
Web3, emerging from around 2017 onward, reintroduced a form of ownership through blockchain technology: people could hold digital assets like tokens directly, without a platform's permission. However, this ownership was largely limited to that one layer — the ledger. The surrounding infrastructure, such as identity, hosting, and naming services, remained controlled by the same centralized providers as before.
Web4 extends this ownership principle across every layer of the internet at once — identity, connectivity, computing power, naming, software, applications, and hardware — rather than just one narrow slice. It also adds a new capability on top of reading, writing, and owning: the ability for a person's own automated agents to act on their behalf, under clearly defined and revocable permissions.