By the 2017 "crypto mania" the infrastructure was barely there. Imagine that even a single app/smart-contract on Ethereum could not scale beyond 500K users without choking the entire network. Cryptokitties comes to mind.
Just building infrastructure is not a magic solution. You need to iterate over developer tools, give developers the right tools, have educational resources, raise awareness around why decentralized apps are important etc. That's generally the developer traction phase where I believe Blockstack is now (we've had 170 independent apps/startups built on top -- most in the last 6 months).
With an active community of developers organically building decentralized apps and playing around with tools, you can see a lot of experiments but can still end up with a graveyard of apps that no one uses. That's the user traction phase which I do not think we're in yet for Web 3 / decentralized apps. I remain confident that as the UX of these apps gets better and as security/privacy becomes more and more imp, we'll start seeing "killer apps". Some of these will be "crypto native" meaning they are not just "decentralized X" but they'll have some crypto native functionality that was simply not possible in web 2.0 and the traditional client/server model.