What are some of the pitfalls that led to a centralised web being built on top of the deecentralised internet protocol?
I'm thinking a lot about this in terms of building on top of NOSTR and where things could go wrong.
Discussion
The web was never decentralized.
Er, yes, it was.
I was there, building my first site on a tenner a month shared space with Demon having read "Learn HTML in a Week". I started on Gopher and WAIS and even Telnet.
Centralisation happens because people like it. They get big services at low costs because of economies of scale.
The same thing will happen to relays. I run nostr.naut.social. It's 4000sats, on most of the lists of Paid Relays, it's a UK Relay so all UK Nostrians should join for that paid relay speed. It covers about half it's cloud costs and NONE of my time.
This is fine, we're all exploring, but roll on 3 years and most people are going to be on the very reliable high performing FREE Google Relay because it's FREE and IT WORKS.
Nostr, like everything else built by people, will be what it will be, and that is almost certainly not going to be anything remotely like what the idealists dream of WHERE those dreams clash with human nature.
Well, but how did your data packets get from A to B. The telecom infrastructure was even more centralized in the early days.
Well if that's where you start, EVERYTHING has always been centralised and always will be. Going right back to hunter gatherer
I worked in the surveillance industry. The internet was a military technology and the agencies never let go of it.
Google and co all where built with that money.
The centralization is a consequence of government having a money printer and thus being able to heave centralized services to a scale that is almost impossible to compete with in a free market way.
But in the long run the market always wins and humans tend to drift to decentralization over time.
The pendulum has only seldom swung far into centralization of civilization as it is now.
There is hope.
Affordability and technical know how to run web and email servers to host content
Also, getting an email account back then wasn't so easy. It's what led to Hotmail and yahoo getting a centralized foothold by providing free accounts and X amount of storage.
Hopefully this doesn't repeat with nostr and lightning but it's well on its way.
Personally I don't host a nostr relay yet, and all my lightning capability that can accept zaps is currently custodial
I think it was just ease of use.
Cult of personality
servers, cloud.
Well, how do we escape the centralization of ISPs? I mean all that bandwidth has to originate from somewhere. 🤷🏻♂️
Most people value quality and easy of use over decentralisation.
Great question, and a good intention to think about deeply.
I think you may find some very helpful insights in the Xanadu project, what the internet should have been...
#[2] did a great presentation about the death of dencetralization in smtp
https://planb.lugano.ch/the-death-of-decentralization-in-smtp/
Four things I can think about:
* technical knowledge barrier (most people have no idea or the will to setup, run and maintain servers)
* ISPs split upload/download speeds favoring download by far making it less favorable to self host
* incentives: people unfortunately do not value privacy enough and as soon as companies like google started to provide great centralized services (starting with gmail), everybody jumped on it and never looked back as decentralized/privacy focus solutions were just more complex to work with
* network effects: most people just use what other people use. Once centralized solution got big enough, it just got bigger from there.
Me thinking one pitfall is if a general gossip protocol is added to nostr big corporate centralized relays could capture the routing market.
No historical basis to go off of. The industries around the internet moved fast and provided immediate conveniences to the public while the companies made big money fast. Little thought went into the effects of how it was built. imo
Effectivity, scalability.
There is only limited amount of use-cases that heavily benefit from decentralization, immutability and censorship resistance.
Same way as timechain is great invention for the purpose of money. But probably not for anything else...
The thing that will prevent a lot of people from using nostr is lack of ease of use
None of this is that difficult, but it’s a barrier most people won’t want to figure out. Some ppl have a hard enough time figuring out how to add a pfp to FB. They’re not going to go to nostr.build
But as more people become educated and grow up with this technology the more these things will become easier to use, and the more things that will be built to increase ease of use
In 2000, there was only a handful of ppl in the world who really understood how the internet worked
Now how many are there?
I don’t see building centralized services on top of decentralized services as a problem.
The economy/market is decentralized, and companies are centralized entities that exist on/within it.
The perception of centralization of “Web2” really comes from a minority of Web2 services like social media and search. The internet is vast, but those services greatly helped adoption.
If some entrepreneurs envision new ways to leverage Nostr I’m all for it.