The Proto-App Thesis, by David Phelps, September 7, 2023

even the strongest case for apps is, at heart, a case for protocols.

It is not a new argument. Joel Monegro’s “Fat Protocol Thesis,” written in the heyday of the app era in 2016, noted that “the market cap of the protocol always grows faster than the combined value of the applications built on top, since the success of the application layer drives further speculation at the protocol layer.” Many have tried to counter the Fat Protocol Thesis—myself included—but by 2023, it feels like all but doctrine.

the next wave of great social apps will all look like protocols—or, if you prefer, proto-apps.

The Social App Thesis, which has defined the fate of the major apps, the social apps, for the past decade. In short, there are a limited number of social categories, and the winners in each have attained distributional network effects so strong as to be unassailable to competitors.

Distributional network effects, or lack thereof, also help explain why there are two major app categories where entrants can claim share alongside incumbents without ever quite disrupting them: messaging and dating.

Chat-GPT is the first major app in nearly a decade because it’s the first example of a whole new category: proto-apps.

Chat-GPT is, essentially, one possible version of a frontend among many tacked on top of a protocol, the OpenAI Model.

proto-apps are distributional networks that can take a cut from the services they feature. They are, in some ways, the new app stores themselves.

the next great apps won’t look or operate like apps at all—they’ll look and operate like protocols. Or rather, these proto-apps will look and operate like nothing we’ve quite seen.


Aside from some examples, Phelps doesn’t focus on crypto, but that is where he’s finding inspiration:

With crypto, the foundational protocol is the blockchain itself, but any app you build on top of it can become a protocol in its own right too—because anyone can build on top of your app as well, permissionlessly, composably, reading the transaction data from your app’s open-source smart contract to write an action on their own app in turn.

My proto app hypothesis includes all the next gen open social protocols: ActivityPub, ATProtocol, and yes, the more crypto adjacent Farcaster.

Notes mentioning this note