Web design is in its totalitarian phase. It is totalitarian in the sense that the potential market for every application is 8 billion humans and something designed for all people will almost certainly not appeal to any of them in particular. Decentralized applications built on platforms like Urbit and Ethereum have the power to change this and usher in a new age of web design.
To understand why web applications are totalitarian at present, it is good to understand both the technical structure of web applications and the revenue model(s) that support these applications. Then we can imagine the design implications of a radical shift in both of these.
Ads
The hegemonic method of generating revenue for a consumer application is advertising. Some applications obfuscate this, but the reality is that, as an application developer, you are often either displaying advertising or you are building out a pseudo-product that you hope will be acquired and implemented as a feature to support...advertising. There is also the time-honored technique of not making any money at all.
Thus, the incentive for consumer product developers is to 1) capture as many attention seconds as possible 2) display as many ads as possible during those seconds 3) charge more for each ad display. At this writing, not a single application’s value proposition to consumers is “use this application to see ads,” which means that in 100% of cases, the incentives of the business do not align with that of a user. This is a farmer-cow relationship more than a producer-consumer relationship, and the design of the user experience reflects this.
Fat Applications
This revenue scheme is supported in a feedback mechanism with the software stack, whose present configuration I’ll call “fat-application.”
In the fat-application model of product development, the majority of the value to the user (such as it is) is provided not at the protocol level (TCP/IP, filesystem), but at the application level. The protocol level has traditionally attempted to be as agnostic as possible and thus as low-level as possible. Since a average user does not actually interact with the whole application layer, but with the user interface of an application, the teams building the application have an incentive to own the entire stack between protocol and UI.
Since he hard limit on application usage is the total attention seconds of a person’s day, these applications are stuck in a zero-sum competition for your eyeballs. Thus, there is and will be constant jockeying between the applications to get you to pay attention to them and, more importantly, to increase your switching costs such that you have a disincentive to change from FaceTube to GoogBook.
From a strictly pragmatic POV, this means that there is an enormous duplication of effort that happens between all of the major web silos. Unless you are playing an elaborate privacy game, the same information about you is duplicated in the databases of every service that you use. And yet there are literally thousands of humans dedicated to “owning” this data and the software that maintains it.
Thus, we have a software stack that is built to optimize on parameters that are opposed to ours (display more ads, make ads better targeted, which increases their potential price to ad salesmen) and to lock us in once we are participants. This is supported by a colossal waste of human capital in the form of wholly redundant systems in the application layer. Every aspect of this is reflected in the UX design of Facebook.

The Design of Decentralized Applications
From the POV of a user, or an application consumer, a blockchain is functionally equivalent to an immutable, distributed database maintained by a trusted party. The real innovation is that there needn’t be a trusted party. That’s a very big deal.
This means that networking and database operations can be folded into the protocol layer, making them openly accessible without necessitating a Google or Facebook to be, at least, that trusted database maintainer. This is what Joel Montenegro refers to as a “fat-protocol.”
In the world of “fat-protocols”, the application layer (which is so thin as to be indistinguishable from a UI) becomes the differentiator between businesses. This makes sense, since there is no meaningful asymmetry between application operators, from a data perspective, since nobody can claim to have access to more than the full chain of, for example, Ethereum.
The fact that data is persistent across many “applications” means that the switching costs between applications are drastically reduced. Without a high switching costs, users are free to shop around for a UI that they like, which means that there will be more competition around UIs. This, I believe, will lead to more specialization and “customization” of the UX space.
Joel leaves open the question of the business models that will be possible here, but my prediction is that there will be many more, better designed applications in this new world. Only a couple of ad-centric companies are actually profitable and even then, I doubt their long term viability. The smart money in the future may be on solving design problems that will allow for maximum customization within a given application and more closely aligning yourself with actual user needs and preferences.
Prepare for a web at human scale.
This post can also be found on my Urbit at recessional.org