Why your Ethereum project will most likely fail

Just because its possible, doesn’t mean it makes sense.

Any non-trivial project, be it a startup, a political organization or a civil society initiative, will most likely fail.

Whatever idea you have for how the world can be improved has to stand up to the fact that its feasibility and/or desirability has been continually refuted up to this point.

Most founders cannot address this fact, and this is why the primary cause of startup failure is: no one needed what you made. One can offer a plausible rebuttal to this fact on behalf of an Ethereum project: the platform did not exist until very recently. However

Most proposed projects on Ethereum cannot point to a plausible end user benefit which is primarily enabled by being on the platform.

Things like “no middlemen” and “decentralization” are not actual end user benefits, they only potentially generate benefits indirectly. Perhaps even more so than in the Bitcoin space, using the Ethereum platform seems to be a decisive end-goal for people in Ethereum projects, rather than providing a better solution to a real world problem.

Evaluating application categories

After reviewing the quite extensive list of planned apps on the platform, I have landed on the following broad categorization and sweeping judgement.

  • Infrastructure
    This is stuff like wallets, exchanges, developer tools and frameworks. These are perfectly useful ideas, but they only make sense in so far as there are actual end user benefits associated with other applications.
  • Organizational Automation
    This is stuff like board operations, voting, and in general replacing humans inside organizational structures (DAOs). We have over the last few decades seen lots of automation of critical tasks inside organizations using software. For example for doing customer and human resource management, accounting, collaboration, and so on. All of these tasks require far less human labor today than before, due to software automation. The question is, could one possibly decentralize these processes, and why would one want to? The problem is that many of these processes require lots of rich and soft local knowledge about time, place, people, risks and relationships, and this requires a thick layer of human judgement to extract, distill and codify. I would expect that existing software solutions have indeed found the optimal boundary between this human judgement and the processing and data representations delegated to the software. Smart contracts can only play a role in displacing the latter side of this boundary. But what would be gained by doing this? All principal agent problems happen on the opposite side of this boundary, where smart contracts cannot efficiently be applied by assumption, and the traditional means of dealing with such problems ultimately seem like the only approach. To put it bluntly: the major cases of misconduct and inefficiency inside organizations have nothing to do with actors breaking software execution on servers.
  • Crowdfunding
    The lesson of the very well built lighthouse Bitcoin project, which has resulted in no more than a few thousand dollars of funding over the past years, has to be that there is extremely little value to be gained by introducing trustless escrow in crowdfunding. Even if you wanted to crowdfund something very controversial, where traditional web platforms will ban you, its not at all clear that disintermediating holding funds before cap is reached is a big part of making this type of crowdfunding effective. Even worse, the project organizer could always fund the remainder of the contract privately, hence it is not at all clear that there is any actual gain of security. While it was a nice use of SIGHASH_NONE, it really solved a non-problem.
  • IoT/Smart Property
    What part of the rules for the operation of some smart property — be it a lock or a car, would benefit from being governed by consensus? There is indeed an atomic car-for-payment swap contract for Bitcoin which highlights this question very well through its comical complexity. Is there any evidence that people have a hard time with the practical exchange of a car for a payment using traditional means? Not that I have ever come across. If you want devices to be smart — in that they have electronically based ownership enforcement, then you can just use digital signatures with a server. In this scenario, the server operator cannot even influence ownership — but I doubt that would ever be a real world issue anyway. The remaining argument, that smart property is now uniquely enabled by Ethereum/Bitcoin because previous attempts were mired in double spending ownership by bad servers, just doesn’t hold up.
  • Gambling
    This is stuff like provably fair dice and lottery games, poker and prediction markets. This is actually the category I would be most optimistic about. There is a healthy ecosystem of Bitcoin gambling sites, many of which can provide provably fair games. These are mostly centralized, yet seem to function very well based purely on reputation. With Ethereum one could plausibly have more complex fair games which currently cannot be trusted to run on illegal gambling sites — in particular at higher stakes, but I am not sure how large this space is. Things liker Poker are very unlikely to work well on the blockchain. It will likely get too costly, and it will be near impossible for any public smart-contract to replicate the critical anti-collusion function that centralized services like PokerStars offer. They use a varied set of measures, including user reports, historical hole card analysis, interviews, charge backs and banning real world identities. Any public anti-collusion measure will likely be much more limited, and can be very quickly be gamed. Lastly, I have already covered my pessimism with the potential value of decentralized prediction markets.

There are other categories, like p2p-finance and finance smart contracts, but I think the general point should be clear at this point. The best case scenario for most of these projects is that there are end user benefits for other reasons, and that Ethereum is not required to deliver them. The complexity, and cost, of doing blockchain based products will in such cases eventually drive such products off chain, the same way changetip eventually went off-chain, and now recently into fiat mode.


I largely agree with Vitalik Buterin, a founder of Ethereum, when he says

First, there will be no “killer app” for blockchain technology. The reason for this is simple: the doctrine of low-hanging fruit. If there existed some particular application for which blockchain technology is massively superior to anything else for a significant portion of the infrastructure of modern society, then people would be loudly talking about it already.

and this is why your Ethereum project will most likely fail. Minor improvements which are more complicated and more costly simply will not make it. Now, don’t get me wrong, the purpose of this article is not to discourage or ridicule anyone working on any of these types of projects, but rather either

a) for the resources — human and otherwise, committed to such projects to be redeployed where there is an actual benefit to be had


b) for me to learn that I’m wrong about these projects, and perhaps the benefits of the platform in general, in which case I’d be the first to try to take advantage of these benefits for my own goals.

Founder of JoyStream, Comp. Neuro. PhD

Founder of JoyStream, Comp. Neuro. PhD