The 1999 dream, still waiting
In 1999, Tim Berners-Lee wrote about a Semantic Web in which computers could analyze all the data on the Web: content, links, and transactions between people and machines. The vision was that day-to-day mechanisms of trade, bureaucracy, and daily life would be handled by machines talking to machines, and the people long promised would finally materialize.
Twenty-six years later, we have the technology to make that work, and almost no one uses it. The reason is human, not technical. Adding schema.org markup to a blog post is a chore, and a chore that yields no visible benefit to the person doing it. The author writes the post for humans, hits publish, and moves on. Going back to add JSON-LD to mark up the book title is homework, and almost nobody does the homework.
The block is the hook
Spolsky's pitch is a clean inversion of the usual order. Instead of asking authors to add semantic markup, the editing environment offers them a block. A search box appears. The author types the title of a book. Autocomplete offers a match. The author clicks. A block gets inserted into the post, displaying the book in a format the author likes, and carrying the structured data behind the scenes. Same idea for an address, a recipe, a person, a Burning Man theme camp. The author does work to insert a block because the UI helps them do that work. The semantic markup is a byproduct.
The list of structured data types is essentially infinite. The challenge is that no single editor covers the long tail. WordPress has hundreds of block types, but not thousands. Notion, Trello, Mailchimp each have their own. If the same author wants a book block in WordPress, a Notion page, and a Mailchimp email, they need three different ecosystems, and no incentive for any developer to build the same block three times.
The Block Protocol's bet
The Block Protocol is an attempt to make a block worth building once and shipping everywhere. A block built for the protocol carries its own behavior, its own appearance, and its own structured data, and it can be embedded into any editor that supports the protocol. The protocol itself is free, open, and vendor-neutral. Commercial and private blocks are allowed, but the spec is not a moat for any one company.
The theory of adoption is that no single human on earth has to cooperate. Each user opts in by adding a block to their editor. Each developer opts in by publishing a block to the protocol. Network effects accumulate without coordination. The WordPress plugin is the wedge: it makes embedding Block Protocol blocks as easy as inserting any other WordPress block, and WordPress alone powers a substantial fraction of the public web. A block built once for the protocol is now usable on a meaningful share of the internet the day it ships.
What's actually shipping
The WordPress plugin will be free and widely available in February, alongside version 0.3 of the Block Protocol specification. The plugin is also useful as a starting point for developers who want to ship a custom block on WordPress without writing PHP or learning the WordPress plugin SDK. The barrier to building a block drops from "learn the WordPress ecosystem" to "follow the protocol."
The real measure of the project is what happens after February. If the protocol's first thousand blocks are useful, the open web starts to look more like the original Semantic Web pitch. If they aren't, this becomes another interesting technical artifact that didn't earn the network effect it needed. The mechanism is in place. The content is the variable.