I should have elaborated a little more on my line of thought: if it doesn't require constant maintenance then we don't
have to have single companies that do this work for millions of people. Other social arrangements are possible even with just the software we have today. It's kind of a continuation of my line of thought at
https://futureofcoding.slack.com/archives/C5T9GPWFL/p1650464577515109.
Though now that I write that out, it occurs to me to ask how much the maintenance burden goes up when doing it for someone (just one other person) beside oneself or outside one's household. There's also the privacy calculus that happens inside human heads at the prospect of trusting someone else with your data. I still don't understand the reasons for herd behavior ("if it's good enough for everyone it's good enough for me"). Perhaps the current social arrangements are a wise reaction to current software.
Oh, another big one is federation. Hosted software works either if it's completely private or completely public, but two people hosted in different places probably have challenges in "peering" data with each other.