With the recent Twitter team announcement of bluesky, a research effort looking at creating a protocol standard out of Twitter, this piece is a timely look at a topic on a lot of minds in tech on the risks of the mega platforms, and what to do about it.

There are some great details here explaining the differences between the two. The idea of newer communications platforms morphing into networks of disparate systems with shared protocol standards certainly gets my decentralization nerves tingling:

A protocol-based system, however, moves much of the decision making away from the center and gives it to the ends of the network. Rather than relying on a single centralized platform, with all of the internal biases and incentives that that entails, anyone would be able to create their own set of rules—including which content do they not want to see and which content would they like to see promoted. Since most people would not wish to manually control all of their own preferences and levels, this could easily fall on any number of third parties—whether they be competing platforms, public interest organizations, or local communities. Those third parties could create whatever interfaces, with whatever rules, they wanted.

More freedom without diminishing the overall power of the technology (at least too much) is always preferable to a world where we try and create mass consensus when that’s an impossible achievement.

I’m glad to see that later in the piece there’s analysis of the business model issue. Companies like Facebook and Twitter didn’t shy away from protocols in favor of centralized walled-gardens because for technical reasons, or because protocols aren’t advanced enough. It’s the simple fact that aggregating the largest audience possible is more efficient for monetization. It’s not necessarily in any one inventor’s interest to work on the protocol problem — why not make a new technology proprietary and closed instead of working on a protocol that advantages a marketplace of competitors?

Protocol standards are also painfully slow to evolve and expand once they gain wide adoption, for logical, and mostly appropriate, reasons. Getting features added to something like NNTP, SMTP, or HTTP takes years if it happens at all. For core communications infrastructure you can make the case for stability and compatibility, but that must be balanced with new innovations and ideas.

I agree broadly with decentralization for technical and sustainability reasons, but it remains to be seen how exactly to get from here to there.