Is Ethereum the Wrong Technology for a Decentralized Social Network?

In the first two minutes of using a new Ethereum-based decentralized network called Akasha, I burned through 0.006 ETH on gas fees, up-voting a couple posts, and commenting on one other. At the current ether price of $700, that's over $4 spent in just a few moments, for hardly any interaction and communication.

Ethereum is an incredible technology, providing a turing complete blockchain to write unstoppable applications—programs that run on nearly 20,000 distributed nodes hosted by people all over the world. But is it the right platform for a decentralized social network, like Akasha?

There are questions of simplicity of use, value proposition, and so on, that could all be asked with regards to using decentralized infrastructure in general, but there is a more fundamental and pernacious issue that will prevent Akasha or any similar platform built on Ethereum from being able to achieve critical adoption in the near future.
Cost.

Until Ethereum achieves sharding or some other massively scalable architecture for decentralized computing, the basic cost of simple interactions will be too high for meaningful adoption by many crypto-enthusiasts, let alone the general public.

Platforms like Steemit have circumvented this issue by creating semi-centralized architectures around new consensus algorithms like Delegated Proof of Stake (DPoS), in essence creating a sort of democratic-republic to the more direct-democratic ambitions of projects like Akasha. In those architectures users delegate their vote to elect what essentially are a couple dozen super-nodes, the true computational operators of the service, that are kept in check by their voting constituents that can strip them from their hosting power.

With Ethereum, applications are unstoppable, but also relatively expensive. There are many reasons for this, but at the end of the day it can be traced to the fact that deploying to, syncing data across, and running computations on 20,000 different computers is inherently expensive.

Discounting issues of latency, there is hope to batch-commit transactions to decentralized storage providers like Sia or Filecoin (IPFS) in the future, but such an architecture would seemingly reintroduce a centralized agent to process and hold reference to such stores, or reintroduce issues of cost if the references are directly on the Ethereum blockchain. Perhaps amortizing the cost across thousands of actions bundled per foreign storage reference could bring costs into line with what end users would be willing to pay per vote, comments, post, and so on.

And maybe Ethereum will shard both storage and computation successfully, and gas fees will plummet to the benefit of all.
Unfortunately, these are not yet realities. And so, no matter how great decentralized Ethereum-based social applications like Akasha get in conventional product features, they will be severely limited from prospering until the underlying technology is cost effective enough for end users to affordably enjoy.

Say hello on twitter :) https://twitter.com/lasernite

Cross-published to medium: https://medium.com/@lasernite/is-ethereum-the-wrong-technology-for-a-decentralized-social-network-74b7d6f958ec