I've always been a huge proponent of a different model of thinking about user data and building user-facing services/applications, whereby instead of building up a centralized silo of user data pertaining each service, services request read/write access to parts of a encrypted, user-controlled data source to access/store their service-specific data.
Sophisticated users can then choose to host their own data for greater control, and others can choose to delegate the storage of their data to a trusted third-party storage provider (entrusted only with the ability to _store_ their data, as the actual data will be encrypted with their keys/passphrases).
Either way, users can rest assured knowing that they can migrate any/all of their data to another provider at any time, and that any service accessing their data through this model cannot prevent them from sharing/forking their own data to be accessed by another service competing in the same space.
Under this model, services would have to compete only by the value they bring to users, and will no longer be able to hide behind lock-in and network effects!
I truly believe this new model of data access will become a powerful enabling force for innovation on the internet, which has far too long been stifled by a focus on petty battles for market dominance through winner-take-all network-effects rather than on building the best product.
If you're interested, projects like remoteStorage , Hoodie , and Solid  offer building blocks we can use to build applications that operate in a model like this today (though with the caveat that you'll have to roll your own encryption layer, and there are not many storage providers to choose from). I'd love to hear about other projects like these if anyone is aware of them, and would like to invite everyone to experiment with these projects and think about how we can improve the developer and end-user experience of services that operate in this model.
Nevertheless, I realize that this will be an uphill battle for sure, because services compelling enough by their own merit to break through the deep-seated network effects of incumbents will have to come first, before this data-access model can begin to reach critical mass and foster a network-effect of its own. And that is starting to seem like an impossible battle to win, because even newcomers today have huge profit-minded incentives to build up their own silos of user data from day one, in hopes of building up network effects and securing a comfortable moat of their own one day.
I still have hope that we can arrive at this new data-access model organically one day through purely technological innovation, and have been working on a framework for building applications like this that aims to improve the developer ergonomics of this data access model through modern tooling (integrating a client-centric query & caching mechanism like GraphQL and Apollo).
But if regulation like those suggested in the article can help speed up the transition by forcing incumbents to open up their silos, I'd welcome it with open arms.
You say "network effect" like it's a bad thing. The internet is valuable because it lets anyone talk to anyone. The distributed web is not valuable (today) because it lets nobody talk to nobody.
"Metcalfe's law states that the value of a telecommunications network is proportional to the square of the number of connected users of the system (n^2)."
Networks are valuable because they're winner takes all. One winner-takes-all dweb is superlinearly more valuable than 50 competing dwebs. That's why messaging networks end up federated.
It’s the overview effect that gives incumbents nearly insurmountable advantage.