If the majority (vast majority) of users of the internet rarely break out of social media and you posit that social media is largely 'dead', then isn't the internet effectively dead for the vast majority of users?
Would those smaller social media platforms then be large, end up commercialized some how, smaller ones pop up, and something else take over until we repeat the process?
Access to the real, genuine Internet people and places will be made invisible; protected by the gargantuan SEO-fed lipid-berg of AI-generated content, keeping the social media peasantry ever-corralled in the cattle pen where they shall be kept happy and fed by their keepers.
The problem isn't really with the internet, but a handful of websites that have come to mediate the internet for a lot of people (despite appearances to the contrary, the two are not the same).
This is a critical distinction, because the former is a problem like "the water is too wet", like you can't really fix that. You can build new digital infrastructure though. That's a solvable problem.
And then that new infrastructure will be subverted just as the current one has been. I don't think this is an issue that can be solved through engineering. I think the problem lies in different domain.
Can't we just build a new infrastructure at that point? The current infrastructure worked well for about a decade. If we have to rebuild every decade then that's not so bad.