I’ve never written a public post of this nature and I find that I am most comfortable doing things in private — I’m a normal person who’s excited by projects like Ethereum and Bitcoin because of their promise of empowering individuals, and I admittedly enjoyed pleasant surprises and benefits as an early participant for both networks. I am hoping that this post serves to help contribute, if only a little, to the network and the increasing number of people joining and participating in crypto-projects.
In my opinion, the past year in Ethereum has been a greater rollercoaster than even the price rise (and fall) of Bitcoin in 2013. The launch of various projects like The DAO (and its crash), Digix, Augur, etc. has been incredibly exciting, and wealth is being created left and right with crowdfunds and ICOs. It’s wonderful to see, and the ecosystem is still blooming with progress and innovation despite various difficulties that mostly arose from the crash of The DAO.
However, while the continued creation of new projects is something to be desired and encouraged, I find that common-folk “investors” are quite keen to jump on investment opportunities, often with the mindset of hoping to get lucky and strike it rich quickly. ICOs that have been hyped up with pretty graphs get filled within hours, sometimes minutes, and there are many, many people waiting to throw their money at the various lottery tickets.
(I’d put a shut up and take my money meme here)
Throwing money at something itself isn’t a vice, but I don’t think there’s enough material out there to help people determine whether something is a worthwhile project; rather, perhaps it is too easy to fool people into thinking that a project is worth investing into. This is not exclusive to Ethereum-based projects. The hundreds of coins listed on coinmarketcap.com and the predictable pump-and-dumps of several hyped coins serve as clear examples of this pattern. What concerns me more however — and the motivation for writing this post — is that there has been a trend of an increasing number of projects within Ethereum which prime themselves for an ICO success with several characteristics that may be well-intended, but in my opinion, fall short: lofty, attractive mission statements; pretty, material-design websites; “whitepapers” that are becoming less and less technical; immediately tradable tokens; and less and less complete codes released.
I believe one of the projects out there that succumbs to my concerns listed above, is the Golem project.
(I have nothing against the team, and I’m sure that the team is working very hard. I do feel bad that I’m picking them, but their crowdsale is in only less than a day away, and I probably won’t have much impact on their ability to raise money anyway. Nonetheless, any project should be open to criticism, and perhaps you, the reader, will find my criticisms valid and contribute to raising awareness for greater due-diligence before investing in crypto projects. A positive review of the Golem project can be found here.)
This was a much longer introduction than I thought it would be, but herein lies my criticisms/concerns for this project.
The Business Case:
Golem aims to create a decentralized IaaS/PaaS/SaaS provisioning system wherein participants are able to either offer their computing resources in return for payment within the Golem network or use an available computing resource for a price determined in a purely decentralized marketplace. The white paper for Golem positions its business case and need largely based on the idea that the price for computing resource prices from existing service providers are still inefficiently priced and full of hefty margins.
Interestingly, their defense to the claim that the market is an oligopoly is a link to the Wikipedia article on oligopolies. But the market isn’t an oligopoly — the commodity of this market, computing power — is not a finite resource! You, me, your mom, and companies are free to acquire computers, servers, etc. for whatever our purposes. The reason for “the cloud computing revolution” is because computing resource is now readily available everywhere and will continue to decrease in price. Furthermore, the price of computing resources is decreasing not only from improvements in technology, but actually also because of competition among companies. As an employee for one of the largest cloud companies, I’m aware that there actually aren’t high margins on computing and infrastructure resources. The main reason why companies like AWS, Azure, Google, IBM, Oracle, Rackspace, HP, Savvis, GoGrid, and even Verizon (there are actually many cloud providers, contrary to what Golem claims) are able to offer successful and popular cloud computing services is not just because clients find it cheaper, but because of the additional benefits of using a service provider that take away the inconvenience of managing their own infrastructure/application.
I find it doubtful that there really will be a price advantage over choosing the Golem Network (for IaaS at least). Regardless, for the purposes of continuing this discussion, let us grant that the computing resources in Golem will be cheaper than the ones continuously being set by the existing vendors. Can the Golem Network services be competitive?
First let’s take a look back at the market that the Golem project wants to target.
In their Medium post, Golem discusses its market like below:
Just to get it out of the way, I’ll put the 1% Fallacy out there; blanket statements about the large size of the market they’re entering and implicitly suggesting they’ll be able to grab even a small slice of it, generally should not be convincing to anyone.
But perhaps they are able to. After all, a decentralized computing resource system for which people can get paid for is a pretty novel idea, and something that prima facie appears achievable via an awesome platform like Ethereum.
Then let’s take a look at this market. The 50 billion-dollar industry that Golem is targeting does not come from average consumers or users that other crypto-projects are targeting. It’s not from those average-Joes or a select few random users who happen to use EC2 for small VMs to play around with. The market exists because it is a market largely for businesses, from SMBs to Enterprises ranging from your local credit union to Goldman Sachs.
And as stated before, the reason why service providers like AWS and Google are popular is because of the level of their service and the abstraction away from uninteresting/difficult aspects that they are able to provide its customers. Thus for Golem to succeed, the Network must offer services strong enough to meet the requirements of not only regular individuals, but of Enterprises that require IaaS/PaaS/SaaS services.
For individuals, I wonder how many of them will really find the Golem Network useful enough to provide meaningful dividends to the investors of the project.
First, Golem defines three groups that will exist in their ecosystem: the Requesters, the Providers, and the Software Developers.
At an individual level, I’m skeptical that there will be a large number of Requesters. For IaaS, there simply isn’t a large enough number of individual people who want/need to purchase computing power for a VM image. That is the case now and will continue as such in the future, not because it’s expensive, but because most people achieve their compute requirement needs from their personal gear. Those developers who do like to spin up an EC2 instance so they have a sandbox on a separate machine may find Golem useful, but that turns to a discussion on the technical capabilities platform. A robust IaaS requires a number of different systems configured to work together beyond “just” a virtual machine. There must be tight integration and support for complex storage, networking, and web services topologies that can address various availability and recovery requirements. Golem makes interesting references to the IPFS project and others like Swarm but the additional dependency on developing technologies make the Golem project all the more risky.
But Golem states that it wants go beyond IaaS/PaaS:
There are various questions that arise regarding the Application Registry itself, but I will skip discussion of it because there isn’t much to work from, aside from the fact that it serves to outline the rules of publishing an application on the Golem network as an Ethereum smart contract.
Regarding the Providers, there may very well be a large number of people willing to offer up their unused computing resource to be used by others in order to be paid — again, the idea is pretty cool. But with computing resources are already cheap, will the average Individual Provider actually make any meaningful money? Will the reward be profitable enough to encourage one keep his/her machine running and take care of any maintenance requirements?
The case for a significant number of average users/individuals using the Golem Network in both the Requester and Provider level seems weak to me. What about for businesses?
As mentioned prior, for their reference to grabbing a share of the 50 billion cloud-solutions market to be relevant, business and enterprise entities must use the software and solutions on the Golem Network. If not, Golem team’s discussion of the IaaS/PaaS/SaaS market size is irrelevant and a moot point.
For businesses, the switchover cost from AWS to another service will be prohibitively high. Beyond even security concerns, there is no pressing reason to migrate a company’s ERP, Financials, BPM, HR, and other suite of applications into a decentralized application architecture that would likely be less efficient.
Furthermore, let’s look at the timeline of the projects, which is contingent upon their receiving of the full 820,000 ether.
Based on even their most liberal projected timelines for the features of each “Golem” the total time is a staggering 93 months — that’s almost 8 years!! Granted, the technological feats from implementing Golem would be quite an achievement, but when my company first began launching our IaaS/PaaS platform, the development of comparable services (and arguably many more services beyond that) listed on the Golem’s roadmap took around 3 years. Yes, we have a lot more developers but perhaps that points to the my overall concern of whether the Golem team can really build a competing product that people would be willing to switch over from AWS and many other cloud solutions that will continue to arise and develop for the next 8 years:
- What benefit does the decentralized IaaS/PaaS/SaaS offer businesses over the existing cloud computing platform?
- Would those benefits outweigh the cost of time and effort of switching over?
- Will the product ever match the quality of the integration and development services offered by many open platforms like Cloud Foundry and OpenShift?
- How will the Golem Network be able to deliver specific business requirements such as SLAs?
Golem had originally been a project intended to be submitted as a proposal to The DAO. It’s a shame that The DAO did not work out because I believe the fee structure is a much more appealing structure for the token holders than what they would get from the ICO.
After The DAO crashed, the Golem team switched to a crowdfunding model, where 1,000 GNK would be issued for every Ether.
They explain the new “economics” of the token here, but I believe a crowdfund for a project like Golem is not an adequate incentive structure to ensure that the team continues to work on developing a performing product, especially for a project as ambitious as Golem.
This is exactly the problem! They say it as if it’s a good thing, but there is no incentive for the creators to continue to build the Golem Network with this framework. All they need is a successful funding of 820,000 Ether, which they get to keep 18% of, and that’s it. This follows same trend of video game developers releasing incomplete products, getting wealthy, and not improving the game. I believe it’s human nature, not an attack on the Golem team members’ characters. The current incentive structure does not ensure a proper development lifecycle.
The current deployment of Golem is a basic CGI rendering platform where runners of Golem can render .blender and .lux files. Aside from an unconvincing performance benchmark testing where we can’t verify the actual metrics performed by the multiple connected nodes running Golem, there isn’t much out there to ameliorate my concerns about whether their grand aspirations are achievable. And frankly, I question whether what they currently have is deserving of a lump sum award of (at the current price) ~$9 Million USD.
This post serves to explain why I personally believe that the Golem token crowdsale is not worthwhile; it is not a personal or hostile attack on the Golem team. Any initiative (especially one that aims to crowdfund an amount of 820,000 ether) must be open to criticism, and investors should research carefully before investing. I hoped to provide the other side of the coin for those who are wondering whether they should invest. As always, corrections to my points and other comments are welcome.
The crowdfund is happening in less than a day and many people are hyped about it so the Golem team will probably obtain the wealth that they aspire. Personally, I’d prefer they (1) re-work the incentive system for both themselves and the token holders and (2) hold-off the crowdsale until after the hard-fork, but it’s their project.
Good luck to the Golem team, the crypto-investors, and the Ethereum ecosystem.