Thanks for the feedback! And yeah, it's absolutely something I've been thinking about. I'm not sure I'll even publish it once I have it working because I'm worried about a bunch of nonsense like CSAM or bigotry, much less the more mundane issues of not spam.

And that's the rub, building a good community is hard, especially on a digital platform, and requires a very different skillet from building good software. I'm not sure I'm cut out for that part, but I can learn from the issues lemmy runs into and try to solve them with technical solutions, namely quality moderation tools.

That's especially challenging in a decentralized system, and it seems to have caused a lot of people to leave, so I'm trying to have a good solution out of the gate.

Since it's decentralized, I can't force anyone to recognize any given moderation without breaking the whole point of decentralization (i.e. nobody has control, even me), so my plan is to rely on a web of trust type system. For example:

1. you flag users you trust
2. users report content for violating certain rules
3. if enough people in your web of trust flag content, you won't see it

I'll probably include a default list, but users would be free to choose their own moderators if they think mine suck. But I have no idea how well that'll work, but once I get a prototype working, I'll post it somewhere (probably here on lemmy) to solicit feedback.

I think this idea is different enough to get people interested, and hopefully robust enough to keep people on the platform. To get content, I'll probably bridge it with lemmy or something so it'll look like another instance (again, not sure if that's feasible or even wanted). It's early days, and the more frustrated I get with lemmy, the more I'll work on it.