I run my own peertube instance like I run my own mastodon instance. Mastodon is cool, your private instance can just link up to any other instance and get all

I run my own peertube instance like I run my own mastodon instance. Mastodon is cool, your private instance can just link up to any other instance and get all the data. If you search for something, your instance will proxy the search back to the originating server. The experience is only trivially more annoying than just using one of the big servers. Peertube is not like that. Notionally it supports all the same features, but when all the details come together, it just doesn't work. Say you have a private instance and you just want to browse videos. How do you find remote content? You can have your server 'follow' another, duplicating its timeline into yours... If the remote server allows it. As a user, you can simply follow any channel anywhere and it will populate into your server's local timeline. But what about search? You have to use a third party search index, and it's opt-in. Or you can go to each remote instance you know of and search there directly. Assuming you have a bunch of content coming into your server somehow, how do you find something to watch? You don't! Most of the filtering and searching is useless. Your individual user subscription feed is the only thing even remotely reliable. As much as I want to love PeerTube, I really, really hate it. Discoverability is so bad that the whole platform is nearly unusable. Realistically I'm sure that the best results you can get is just from signing up to the largest instance you can find. Which really defeats the entire point of it being distributed and federated. It really doesn't matter how many videos exist on the service if no one can find them. PeerTube is going to have to put some serious effort into this problem or they're going to fade into irrelevancy in a few years.