Why crypto-economics suggests that IPFS will fail

Tumblr on Mars

IPFS, the Interplanetary File System, has long been one of the most important and exciting projects of web 3. Who wouldn't get interested in a protocol that made internet communication so resilient that the first colonists on Mars could browse Wikipedia? And, while we await the first rocket ships, that same protocol can route around censorship, help billions with poor internet connections access information and provide the storage backbone for decentralised applications.

Or that was the theory. A few months ago, I'd have picked IPFS as the likely winner of the decentralised storage battle. Now, despite the project being as exciting as ever, I think there is a big chance that it will lose out to some other protocol.

Centralised funding for decentralised protocols

There is one simple reason why IPFS is now more likely to fail than dominate: it's technical architecture is designed for the age of decentralisation, but they have chosen a funding architecture from the age of centralisation.

For months, IPFS have been teasing us with the idea of a Filecoin that will be used to pay for the protocol. That's still the idea but, now, the token sale for Filecoin will be restricted to accredited (i.e., rich) investors in the US. This decision might be prompted by legal anxieties over ICOs but, nonetheless, choosing a funding architecture from the age of centralisation reduces the potential for network growth and developer buy-in.

How to grow a network protocol

Let's think about what's needed for a new protocol to become successful. You need three things:

1. Developer mindshare.
2. A network of connected services.
3. Users of the protocol (who don't have to care about the protocol at all).

To get the users, you need to attract developers, who will build out to connected services. The question, then, is how do you get the developers? In the blockchain economy, the answer is obvious: you offer them the chance to buy into future network growth.

IPFS already has a lot of traction with developers - it has 7.5k stars on Github and is the go-to decentralised storage solution for current-generation dapps. But much of that developer interest is intertwined with the crypto economy, and the underlying incentive of owning a part of the network itself. If the Filecoin sale is restricted to US accredited investors then will developers stick with the IPFS project when they can no longer buy a share in the underlying protocol? In the old, centralised, world, they might have done, simply because of a lack of choice. Now developers are more likely to jump ship, because there are other protocols that they can work with and, simultaneously, buy into as investors. Perhaps Ethereum's Swarm will get more support, or one of the standalone projects like Sia or Maidsafe will start to become the default dapp file storage because a developer has a personal interest in building a connector to one of those protocols, rather than using IPFS.

I think IPFS is playing a risky game by, effectively, removing an incentive from the decentralised economy and still expecting its decentralised technology to take off.

Invest like a crypto-economist

If this idea - that developers need to be incentivised with co-ownership - is correct, then how should we invest? On the narrow decentralised storage question, the competitors to IPFS like Sia and Storj now look more attractive. It also raises the possibility of Ethereum's Swarm developing into a winner-takes-all proposition. More generally, it makes sense to invest in protocols that prioritise developer buy-in and are careful not to jeopardise it. Ethereum continues to look strong on this front. The ecosystem of services built around Steem show how far incentivising developers can get you. The Status and Kin token launches both look like interesting opportunities to take this model and apply it in a messaging platform context.