Sunday, May 8, 2022

How Much Can Metaverse Change?

Many argue that the key difference between today’s internet and tomorrow’s metaverse is the architecture. Many argue that the metaverse will be “open,” compared to today’s “cl;osed” internet. 


Others might argue that the difference is metaverse “decentralization” compared to today’s “centralization.” 


source: a16z 


That seems a hope that will not, in the end, actually be realized as its proponents intend. To the extent that the metaverse creates property rights--and nobody denies that property rights will be created--is there historical precedent for such rights to be held in mostly-decentralized fashion?


In other words, will there not be rights holders with more volume than most others? In other words, in any mature market, do leaders not emerge? Will users and customers not gravitate to those products deemed “best?” 


If the goal is to prevent the emergence of gatekeepers, is that realistic? Are we not confusing “degree of centralization” with “degree of power” or “degree of influence” or “share of market?”


To use an analogy, it is one thing to propose creation of a “classless society.” It is quite something else to actually create it. In fact, both decentralization (people have formal rights) and centralization (some have more power and wealth than others) tend to coexist. 


Formally “open” systems can result in “unequal” outcomes. Users might have choices, but platforms can still exist. Both direct and mediated interactions can operate at the same time. 


Of course, there is another historical precedent. Revolutions happen. But classless societies do not result. One set of rulers is exchanged for another. The identities of the “ruling class” will change; but society does not become classless. 


So even if metaverses do not eliminate gatekeepers and platforms, there is a reasonable chance they will create new gatekeepers and platforms.


No comments:

Costs of Creating Machine Learning Models is Up Sharply

With the caveat that we must be careful about making linear extrapolations into the future, training costs of state-of-the-art AI models hav...