I urge users to become participants and contributors. As Kongar has written, we have access to sophisticated hardware, but we're not deploying it. If mobile-computer operators contributed to a mesh-network we'd be in a position to discard subscriptions for telecommunications access. Before this can happen the expectations of users needs to shift from contributing capital (paying for a mobile plan for instance) to contributing resources: running voluntary software on their machines.
The requirements needed to participate in mesh-networks are not just technical knowledge. If a user deployed a program like Briar they'd notice something rather swiftly: participating in mesh-networks demands more resources than not participating. The most obvious effect of this is that the mobile-computer's battery will drain faster. Carrying a spare battery seems to be common already and that's without participating in a voluntary mesh-network. The users' expectations would need to shift before they could adopt such software.
I'll refrain from commentary on Apple's iOS (though I expect it to be similar), but Google's Android is hostile towards user-software that needs to run constantly in the background. It will terminate those programs, even if the user has granted it 'permission' to run in the background.
Desktop and laptops provide a much more fertile environment in comparison. It is the norm to have long running user-programs. Even commercial operating-systems do not find anything amiss because it is the norm here. For the user in such an environment I do not see any possible objection to running such programs other than an increase in data usage, but even then these programs often allow for deployments that have a fixed quota.
Why then do we not see mass adoption of software like i2p, hyphanet, GNUnet and so forth? I exclude tor because it reinforces the status quo that must be changed. It allows its users to exit the tor-network and access the web. It's a good technical precaution against network surveillance, but the majority of that network's userbase want to exit the network. To contribute to the tor-network is to run an exit-node and that is not financially feasible for the vast majority of internet users. It is not democratic, but exclusionary and elitist. User-contributed nodes that run on the hardware they have, with the data connections they have, will see their nodes deprioritized in routing if not outright dropped. The other projects I've listed do prioritize peer-to-peer communication whilst tor's network has just become a way to access the web anonymously. I2p is technically superior to tor's for in-network communication; it supports tcp and udp. You can torrent within i2p's network. You can not torrent in tor's: you lose udp.
In my eyes the major issue of the current 'web' is the rent-seeking economy that upholds it. Technically we all have the tools to self-host a blog, e-mail, lemmy instance or whatever it is you want to share. We have a glut of cpu-capacity, storage and bandwidth. The hindrance isn't in the property we own, but in the capital expenditure demanded from a perpetual rent-seeking economy that's evolved around the 'web'. Users are behind cgnats and need to rent whitelisted servers. Users aren't allocated a static ipv4 address and need to rent one. Users then need to rent domain names and before Let's Encrypt they had to pay for a SSL certificate that had an expiration date--a perpetual expense. Completely arbitrary, artificial scarcity to create an economic system of rents. It's all absurd.
Software like lemmy isn't revolutionary either. It's retroactive. It wants to shift conditions back to when web-masters had privileges. Privileges that were taken away by the new-style of business that superseded them. If you want to run lemmy you need to pay your dues to the domain registrars and whitelisted server admins. The revolutionary act is to run i2p or gnunet and do whatever the fuck you want, but once you realize Alice and Bob can communicate with one another without paying tolls to DNS registrars or whitelisted server-admins nor subjugate themselves to that third-line--the forum administrators--lemmy becomes superfluous.
Users shift responsibility to the server-admins that run the lemmy instances they interact with. In exchange they give them all their content. That content has value. Lemmy is still a value extraction mechanism because everything on the 'web' has devolved into economic rent-seeking. To exist on the web is to participate in that system. It is the 'web' that needs to be ditched. Of the alternatives GNUnet is the design that makes the most sense, though it is still juvenile.