The real answer is that size is a trade off. If you make entities too big you get all the problems of a monopoly, single point of failure, monoculture, etc. If you make entities too small you start having problems coordinating and knowing who to trust.

What you want is something in the middle. Email is a good example.

>Yes and this assertion is laughably wrong: >History shows the exact opposite. For example, at one point every business above a certain size had its own electric generator, and developed their own electricity on-site. In other words, decentralized power. That was replaced by contracts with central generation providers because that centralized model was so much more efficient.

Your comment assumes that efficiency is the most important thing. Addressing power imbalances and improving local resiliency are also important.

In the 1960s and 70s for many people computers meant centralized power and oppression. Computers were huge and expensive and therefore centralized, and only big government and big business (i.e. The Man) could afford them. A lot of the appeal of desktop computers in the 1980s was that they were personal: you could run the computing tasks that you wanted, on a machine that you controlled.

Though rather than home users, it was business that drove the adoption and commercialisation of the pc. That was through networks of desktop machines running software like lotus which could be configured by users, replacing mainframes that required costly administrators and programmers. It was largely about cost efficiency and agility, particularly as the financial sector was undergoing big changes at that time.

Maybe for people who already liked and used computers, which was a tiny minority. For most Americans, a desktop PC was their very first direct experience with computing at all.

Today what do most people use their computers for? Accessing the services of big service providers like Google, Apple, Facebook, Twitter, Amazon, etc.