> If it's read-only data, I feel like hug-of-death should not be the norm.
That's what PeerTube is good at. Each video has a home site and a master copy, and each watcher helps to redistribute the video. So, if your cat video goes viral, your tiny server doesn't get hugged to death. At least in theory.
This is less distributed than bittorrent or Usenet. It's just distributed caching. Only videos currently being watched are replicated.
Now, PeerTube doesn't help with discovery. PeerTube sites only list their own videos. For this to take off and replace YouTube, a discovery mechanism (what we used to call a "search engine") is needed. Those exist.[1] (There are others, but some have bad SSL cert warnings.)
If you break the problem into discovery, hosting, and caching, all three parts are tractable. And no one part has enough authority to get uppity and act like they're in charge.
A lot of people trying to sign up probably. This degrades performance of one part of the system that is generally not highly optimised. This permeates everywhere
It sounds like you're arguing from a hypothetically perfect view with 20/20 hindsight. Scaling problems are good problems, solvable at the right time without the evils of premature optimization. Doing work without a need is wasted effort.
> It sounds like you're arguing from a hypothetically perfect view with 20/20 hindsight.
I'm just stating the obvious, which is that whether a service is centralized or federated has nothing to do with it. Mastodon would melt long before operating at Twitter scale, for example. (And to your points, I don't think that's necessarily bad.)
But, if there's any company on earth that should have 20/20 hindsight about this, it's Twitter. Post-"fail whale" era and pre-Musk, the service weathered far larger challenges without collapsing.
It probably would, but Mastodon is now at around 13M accounts, but with ca. 3x as many instances as when it hit 10M, so it's increasingly better prepared for large user surges.
E.g. signups have hit ~6x what they were earlier today but it's still only an average of one signup per instance per hour.
Great! I’m a fan and I hope Mastodon/ActivityPub continues to grow. I just wish it wasn’t prohibitively complex/expensive to host an instance with a few thousand members.
I think Mastodon will continue to be expensive to host, as the way it's engineered is not well suited to cut costs, but I also think we'll see a lot of innovation in "slimmer" services that will be "Mastodon-compatible enough" but lighter weight. Ideally we'll eventually see some sort of standardisation of a client-server API (rather than just depending on people ad-hoc reimplementing Mastodons API), but that may take longer.
eh, I think it's arguable here that making a good experience here would really help cement people swapping over. Then again unclear if twitter is going to remove the super low limit, so people may abandon it anyway.
