Web 1 — Everyone sees same content when they open a website.
Servers just had filesystem and nothing else. Creation is limited to people with some learning curve. Ran from 1991-2004. Still a significant part of current ecosystem under CDNs.
This is analogous to newspapers and other such media. We can only consume, and everyone consumes same data.
Web 2 — Internet as we know it today. People can interact with websites, create their own content, have personalised websites, etc. Data and software to collect data is owned by central entities who run servers (backend) and hence gave rise to monopolies like Google, Amazon, etc. They've control over my usage and hence I can't delete data or history.
This is far from non-computerised world shopping/knowledge sharing experience. I can go to a general store, buy something and come back without letting anyone except shoopkeeper and I knowing about it. And the shopkeeper doesn't keep a list of who bought what, either. But internet companies did keep all the data, and used it to fuel more technology enabled solutions.
Web 3 — People can interact with websites and have the ability to control their data. A crowd sourced ledger with enough controls (diff ways to do it, like majority votes or algorithms, etc.)
So, people can decide whether the data stays or not. Servers don’t run on central entities, instead the database is shared across and multiple people can serve/own.
For example, if I order something from Amazon and want to delete it from history without a trace forever, I should be able to do it.
However, practically speaking, Web3 still needs servers. And people with money has the governing power over algorithms, servers, and the ecosystem. It might make a difference overall, but it's likely to impact a small percentage of internet.