WEB3 NETWORKS ARE COLLABORATIVELY MANAGED, BUT NOT NECESSARILY COMPLETELY DECENTRALIZED!
Shermin Voshmgir: April 27, 2022
Since many refer to the Web3 as the decentralized Web, I would like to spend some time critically reflecting this term. Personally, I am not a fan of the word “decentralization” in the context of the Web3 for various reasons:
First of all, the term “decentralization” describes more what the Web3 opposes rather than what it allows us to do, namely to collectively manage web applications, their underlying data and the governance rules of all social and economic interaction of the network including the token generation rules. Rather than speaking of the “decentralized web” I would therefore suggest referring to it as the “collectively managed web,” where no single entity (private or public) has unilateral control over our data. In the Web3 data is collectively managed, and individual participation is rewarded with network tokens. Individuals are privately incentivized to contribute to a public infrastructure. This is an absolute paradigm shift not only regarding how the Web3 is managed in the backend, but also how individuals can be rewarded to contribute to the upkeep of public infrastructure in general through token rewards and it all started with Bitcoins reward mechanisms “Proof-of-Work.”
THE TERM “DECENTRALIZATION” IS SOMEWHAT OF AN “ANTI-WORD”
It categorically opposes “centralization” and implies somewhat of a more democratic Web. It also implies that fully automated systems need no manual intervention and course correction. While automation and decentralization both have their merits, no large scale social, economic or political institution can be fully decentralized and automated without human intervention. Protocol upgrades and course correction are necessary in any Web3 based network. In the event of unexpected events, manual and often rapid intervention is needed and might not be able to be provided by fully decentralized systems.
DECENTRALIZATION IS NOT BINARY
Furthermore it is important to acknowledge that — depending on the type of Web3 network — the levels of decentralization vary. Often, the practical implementations of these networks are less “decentralized” than the initial founders had envisioned. The Bitcoin network — for example — is not fully decentralized but has become de facto “controlled” by a handful of large mining companies, a limited number of full node operators, and a handful of core protocol developers. The same is true for the “de facto” power structures in many other Blockchain networks. While none of the different network actors usually have exclusive power over the network, points of power concentration have developed over the years in all blockchain networks or other Web3 networks. Such power structures and points of centralization are a political and economic reality not only in the Web3 but anywhere institutions and individuals pursuing their own best interest collide. This needs to be considered when designing the initial token governance rules of any decentralized application/Web3 network/DAO.
WHEN CAN WE IN FACT SPEAK ABOUT DECENTRALIZATION?
If the network is controlled by 5, 50, 500 or 50.000 different people and institutions? What aspect of network management do we consider to be decentralized: the code development? The policy development (governance)? The effective maintenance and value contributions of network participants? Which aspect of the network is decentralized and to what degree? We need to be much more specific when talking about “decentralization.” The reality of Web3 networks has shown that decentralization is not a binary state but most likely evolves along a gradient, on a scale of 0 (full centralization) to 100 (fully decentralized). Depending on the type and purpose of the Web3 network, different levels of decentralization might be optimal. This is why I dedicated two whole chapters of the book (Token Economy) to describe the intricacies of DAOs Governance.
Furthermore there are many people who argue that fully decentralized systems also have downfalls, and I tend to agree. We can find testament of this assumption in the history of Government, Markets, Biology and Engineering. Direct democracy has proven — for many reasons — not to work — at least in large societies. Speed of decision making and decision overload that it imposes over its citizens are some of many issues of “decentralization of government,” which is why most democratically run societies have resorted to various forms of representative democracy instead of direct democracy — where certain points of centralization exist. In terms of markets, fully unregulated aka “decentralized” aka “free markets” also have their downfalls which is why governments across the globe have introduced laws to regulate economic interactions. Since individuals and the companies they own and operate, tend to act in their own best interests, they often ignore the greater good of society and resort to tactics such as forming cartels, “bid rigging” or predatory pricing. The merits and downfalls of more or less regulation has been the subject of many political debates over centuries if not millennia, and similar political debates are the topic of many Twitter feuds around the “right” type of governance of Web3 networks.
While it is likely that the future of the Internet might be more decentralized, this does not mean that we will get rid of centralized systems altogether. Centralized systems, or at least a certain degree of centralization, have advantages and will likely prevail, at least for specific use cases. Which is why I would like to turn the attention to the fact that Web3 networks and their applications are collaboratively managed. The right level of collaboration needs to be determined based on the purpose of a network, and the values of the participating community and in consideration of the reality of socio-economic dynamics.This is an excerpt of the upcoming Third Edition of my book Token Economy. It is a text I recently added to the first chapter of the updated book which should be released later this year.