The problem with statements like "I happily use Matrix but my grandma will not, ever, not in this world" is that Matrix is more of a protocol than anything. So of course, your gradma would never use HTTP, but since browsers are now user friendly enough, everyone is using HTTP indirectly. Same can (will? Might?) happen with Matrix, where a popular applications get written on top of Matrix, and users don't even need to know about Matrix, just that stuff works. And when they move to another client, they can bring everything with them and stuff keeps working automatically.
Whatsapp got bought by Facebook which changed their strategy 100%. They used to ask users for a $1/year payment or something like that, but after being bought by Facebook, they now focus on just providing data to Facebook. I think one of the founders also have openly gone out and regretted the sale to Facebook - "At the end of the day, I sold my company. I sold my users’ privacy to a larger benefit. I made a choice and a compromise. I live with that every day." - https://www.fastcompany.com/90242892/whatsapp-cofounder-expl...
If we manage to keep the governance of the client and protocol open and participatory, I'm sure we can have clients/protocols that survive for longer. Just like we did with email and IRC in the past.
Barely any one paid the $1/year. My estimation from looking into this before is less than 2% of Android users ever paid. At most half of iPhone users paid $1 as whatsapp would switch between free or paid to get it.
The founder regretting is a joke. Facebook isn’t buying you for $19B to leave the company alone forever. The guy is worth multiple billions. No one told him when it was being sold that privacy won’t remain the same? That seems too unlikely.
So the plan is to keep on creating new protocols so that there's frequent, although not constant, stream of services that are already user-friendly, but not optimized for milking data yet?
Milking data, user friendliness and open protocols are orthogonal areas that gives different benefits. User friendliness makes it so people actually use the product, open protocols forces the service/application providers to care more about the user (otherwise they leave for something else, taking their data with them) and milking data can be done no matter how good/bad user experience is and if the protocol is closed/open.
We need to move beyond being saddened by the nature of systems. It's like being sad about gravity, evolution, or the speed of light. Network effects, social/cultural-scale and emergent behavior are the same.
If your design is subject to a failure, it's a failure of your design.
This is an empowering viewpoint. Let's stop blaming factors outside our control and find ways to create resiliency.
The problem is when a bunch of optimistic, altruistic, talented individuals just imagine we could choose to do better, like in the social dilemma. Similar environments and pressures will lead to similar outcome, like carcinization.
Somehow we still manage to be surprised and disappointed when this happens.
There are measures we can take: provider and implementation diversity should curb this. Not everyone is "grandma" so not everyone needs to use a client for "grandma".
The problem is that it doesn't keep working automatically, because federated servers are in practice a hot mess of mismatched versioning and application support.
