Web3 is a hot topic these days. No, I do not mean Web 3.0 (so-called "The Semantic Web"), but Web3 - the (somehow vague) idea of building the next stage of the Internet upon "a fully decentralized online ecosystem based on blockchain".

According to Web3 zealots, the Internet as we know right now (Web 2.0) may be running on decentralized, open protocols (like HTTP), but it's also primarily controlled by "a handful of super-powerful corporations - think Google, Facebook, Amazon - subject to the nominal control of centralized government regulators".

I have some mixed feelings about that, do you?

What is centralized?

First of all, indeed, there are strongly centralized aspects of the Internet as we know it:

  • DNS is centralized
  • PKI is also somehow centralized - because of the chain of trust mechanism and existence of CAs
  • scarce resources (like IPv4 addresses) are also managed centrally
  • certain types of (financial) transactions are bound to physical identity, which is managed in a centralized manner(s)
  • key steps of (online) payment processes are under governance and control of certain centrally appointed institutions

I do see some patterns here. Centralization is used for the sake of:

  1. various aspects of security (e.g., identity validation, eliminating the ambiguity, etc.)
  2. custody over common standards (and how they are respected)
  3. administration of resources that have to be in the shared pool or (for whatever reason) are scarce

Necessary or not?

For me, all of those reasons are valid and, in the end, are all about "outsourcing trust". There are some areas of our lives that require a certain level of stability/predictability (worth paying for - either in the form of monetary premium or operational inefficiency). Personally, I expect that:

  • no one will hijack/override my domain
  • the money in my savings account doesn't get stolen, and its value remains reasonably stable overnight and prone only to market- or global economy-related factors (instead of Elon's tweets ...)
  • if someone manages to steal from me and the funds get transferred, the recipient is not only traceable but also their identity can be revealed (if the law requires that)

I'm fine with PAYING for these guarantees - because my imagination is good enough to imagine reality without them.

In these cases, my trust is based upon law regulations, appointed institutions, government guarantees, various consortia, and cooperatives with a particular mandate. I am good with that - as they are all ACCOUNTABLE. Yes, they may not be perfect (what/who is?), but I find it more reasonable to improve them evolutionarily instead of wiping them off (w/o a viable alternative).

Betting on a fully decentralized model means that the accountability is all gone, and I need to trust: algorithms, whitepapers, the most popular cryptocurrency exchanges (or their equivalents). And only within a certain range/scope because it's fully decentralized - while I use protocol/standard X, my neighbor may be fully into Y. Who could forbid him?

Do you think such a reality would work better for the "average Joe"? I don't think so.

Is there another option?

Centralization is an instrument. A tool. Or a mechanism. It has a price, but it works pretty well in some scenarios - exactly in the ones that require stability and predictability. Ditching it is simply naive - it will bring questionable advantages to the masses, except the small subset of dodgers, tricksters, and people aware enough to take advantage of the ignorance of others (who just follow the flock).

The perfect model is probably (as usual) at some sort of equilibrium (between centralization and decentralization) - and it takes what's the best from both models, for the benefit of all the parties involved.


No, I don't have anything against Web3. Actually, I'm a believer; I just imagine it differently - in parallel to the current monetary system, in a supplementary role. IMHO Web3  has a huge potential as (what some people call) an Internet of Value - the next evolutionary step from what Ether does in Ethereum transactions these days:

  • seamless ("in the background") payments for information we consume on the Internet (which would depend on fixed exchange rates - e.g., by binding them to another currency or some formula)
  • traceability of information sources/references for data - which, of course, would require a new form of self-contained, "web3-signed" information