This is not an AI story. This is better.
Four open source video player projects. Competitors. Each with their own community, their own architecture, their own egos. Combined: 75,000+ GitHub stars and tens of billions of video plays across the web.
Video.js. Plyr. Vidstack. Media Chrome. Four teams, four approaches to the same problem.
They merged. Into one project. And rewrote everything from scratch.
The result: 88% smaller than the previous version.
How this actually happened
Video.js has been around for 16 years. The original creator, Steve Heffernan, stepped back years ago. The project grew, accumulated dependencies, gained features nobody used, and became what all long-lived open source projects become: a sprawling codebase that works but nobody wants to touch.
Then Heffernan came back. Not to maintain the old code. To start over.
But instead of starting over alone, he reached out to the teams building the competing projects. Not to acquire them. Not to hire them. To collaborate. To take the best ideas from four different approaches and build something none of them could build individually.
That conversation, four rival project leads deciding to work together instead of competing, is the rarest thing in open source. It almost never happens. Ego prevents it. Sunk cost fallacy prevents it. “But my architecture is better” prevents it.
These four got past all of it.
Why 88% smaller matters
An 88% reduction in bundle size is not an incremental improvement. It is a fundamental rethinking of what a video player needs to be.
The old Video.js was 400KB+ minified. The new version is under 50KB. On a mobile connection in India, that is the difference between the video loading and the user giving up. Across tens of billions of plays, it is petabytes of bandwidth saved.
This is what happens when you remove 16 years of accumulated assumptions and ask: what does this actually need to do?
Why I care about this
In a world drowning in AI hype, this story has zero AI in it. No LLMs were involved. No prompt engineering. No training runs. Just four teams of human engineers who decided that collaboration was more important than competition.
They looked at the same problem from four different angles. They took the best solution from each. And they produced something 88% smaller than any of them had alone.
That is engineering. Not the Silicon Valley kind where “engineering” means “scaling a product.” The real kind. Where you make something better by making it less.
Video.js v10 beta is available now. videojs.org
[Draft: Awaiting Carlos’s twist]