I read something recently that gave me food for thought. An alternative has to be 10x better for people to switch; and the core experience is what convinces people, not cool extra features.

I don't know how universally applicable it is, but I wonder how Mastodon stacks up in that. Personally I think it's 10x better, but is it really? Or more importantly, are we communicating clearly that it is?

@Gargron I think instance selection is a huge roadblock.

Keep laughing, but: dynamically generated instances as a default sign up, including a dynamically generated cute name and logo. Once the person count hits Dunbar's number, a new instance automatically spins up.

Pooling new people together gives them the opportunity to meet each other and create their own culture.

Of course this would just be a default option for someone who doesn't already have a specific instance in mind.

@kai @gargron and then you've defeated the point of federation

@mhu2141ai @Gargron yes, centralization is a point of concern. Maybe the new auto-generated instances could expire and force people elsewhere after a time limit.

My concern is that the onboarding process is simply too arduous for all but elite or highly motivated people.

@kai I apologize for the tone -- it's just that things looking to maximize user count / "engagement" have been universally terrible IME and selecting for people who make a baseline effort helps. If people aren't thinking about who moderates their instance, their experience will probably end up suffering for it anyway, and -everyone- loses if you have a flood of people who don't know wtf they're doing

@mhu2141ai no need to apologize; I'm no social media philosopher or anything.

I just remember a reddit april fools' day where you got placed in a chatroom with one other person and then the rooms got joined so it was 4 people and that kept happening and you could vote when to stop growing the community. It was really interesting. I think a similar experiment would be interesting with activitypub.

@kai Yeah, as an experiment and not a general purpose feature that sounds fun

@mhu2141ai it solves other onboarding problems though 😉

@kai I'll reiterate: anyone who can't spend 10-15 minutes selecting an instance deserves 0 onboarding, and onboarding them would likely be a detriment

@elomatreb @kai I really do not think this is half as difficult as you're making it

@elomatreb @kai Honestly I see no issue even with that, nor any way to automate it sanely if there was

@elomatreb @kai Yeah, I'm willing to concede a fair bit of what I said for that to remain...