But the web is not a decentralized place. You put content on servers, and everybody downloads that content from there. There a copies or caches, yes (CDNs but also old-school http proxy caches). But if I remove content from my server (or a tweet), its no longer accessible by standard means. We build a whole set of technologies to not have a single-point-of-failure (i.e. load balanced servers + replication) but for example in CCN/IPFS you can publish content a single time from a very slow dialup line, and in theory it could be accessed by millions of users (which in todays world is only possible if you put it on a CDN or run your own datacenter to satisfy requests at that scale).

There are hundreds of companies providing managed object storage and likely tens of thousands with file storage.

And you are free to create your own. It IS decentralised just at a different granularity to IPFS.

> but for example in CCN/IPFS

That is why I a fan of those. Not sure what you are arguing. My you misunderstood my initial comment.