September 26-27, 2023 - Meet us at Current hosted by Confluent! Stop by booth #211 to hear how you can build streaming data applications in minutes, not months. / Register Now

A Web of Agents

The Internet vs. The Web

The internet is a global network The web is a global library of resources The web is an application of the internet The web lives alongside other applications of the internet, like email, voip, etc.

The Web has become the dominant application of the internet

Even adjacent internet applications, like email, are now commonly accessed via web interfaces.

Most mobile apps communicate over APIs built on Web technology.

Defining characteristics of the Web

Universal addressability. The idea that you can uniquely identify any resource on the planet with a short string of text (a URI)

Hyperlinks. People forget about the importance of hyperlinks these days. Hyperlinks are what make the Web a web. Without links, it’s not a web, it’s a sandbox.

Semantic content. The Web began with a semantic language for documents: hypertext. The Web had a notable failure in the early 2000’s to become even more semantically rich (the semantic web). The semantic web didn’t fail because people didn’t want semantic content; it failed because the existing semantics were good enough. There’s a cautionary tale here fir Web3.

A Web of Tokens?

The idea of a unique, non-fungible token (NFT), on the one hand, seems like a natural extension of universal addressability. An NFT is a unique identifier of ownership.

Two problems though. First, MFTs just prove that you own the token; they don’t prove that you own the thing that the token may claim to represent.

Second, we already have unique tokens that can convey both identity and ownership: URIs do both if these things already.

Centralization vs. Decentralization

The argument against URIs as ownership is that there are central authorities invoked in provisioning URIs, namely the Domain Name System (DNS) and various industry providers.

The other argument against URIs is that the documents they reference can be changed, and their history can’t be tracked.

DNS is highly decentralized

The Domain Name System is managed by s central authority, called ICANN. But ICANN delegates top-level domains to thousands of other DNS registrars.

The Pirate Bay has no problem obtaining domain names.

Web2 is more democratized—and more decentralized—than Web3

A common claim is that Web3 will further democratize the Web. That power. over Web2 is too consolidated in Big Tech.

This is nonsense. Anyone can still run a web server. You have to pay to play with Web3 too; you just have to pay a different cabal.

Blockchains can be censored

The idea that Web3 will be a libertarian paradise is also a crock.

Blockchain processors are free to reject transactions. They’re highly incentivized not to do so today because they’re on.the up-swing of a pyramid scheme.

Bitcoin is deflationary by design. As Bitcoins dry up, they will shift to processing fees to incentivize the network. Processing fees are a much less lucrative incentive. But more importantly, take note that processors CAN reject transactions. And this can and will be used for censorship.

Proof of stake is better for the environment, but even worse for censorship and consolidation of the network

Proof of stack cements the power of the creators of the pyramid scheme. It’s very clever. Very cynical. And very anti-libertarian.

Permission from nobody? Or permission from everybody?

To do something on Web2, you need permission from a small number of competing organizations. Emphasis on competing.

Yes FANG is too big. That has nothing to do with Web2. FANG can and will be disrupted, like Microsoft and AT&T, and Standard Oil before them.

To do anything on Web3, you need the permission of a majority of a syndicate. This is a truly terrible idea. And its effects will be the complete opposite of what its proponents contend.

Why not a Web of Agents instead?

Web2 has a much more pressing limitation: it’s a Web of Documents. And not everything we want to include on the Web is a document.

A document can’t act on your behalf. A document can’t automate an enterprise.

Documents are inherently historical. Documents are the wrong metaphor for modeming what’s happening in real-time.

The Web has always had Agents

They’ve just been limited to User Agents. The technical term used for a web browser in web standards is literally the phrase “User-Agent”.

Users have become obligate bureaucrats

This truly sucks. It sucks to have to do so much damn paperwork to manage your digital life on the Web. We’re living in Terry Gilliam’s Brasil, but with an internet of tubes instead of vacuum tubes, and GDPR notices instead of TPS reports.

Users are saddled with micromanaging every aspect of their digital lives—which, mismanaged, leaves to gross privacy violations and security vulnerabilities—because the passive web of documents can’t do shit until somebody asks.

Let’s give autonomy and agency to the Web, instead of artificial, power-consolidating scarcity

You, as an individual, are massively outmatched by FANG. What if you could give your data the agency to fight back on its own, on your behalf? Wrap your data in a Web Agent, and it can gate access to your data for you. It can guard your privacy for you. And it can do so every second, of every day. You can level the playing field against FANG, and even tip the scales in the other direction.