Peertube is federated. It seems to work similarly to Lemmy. I went on a random instance and clicked “discover” and noticed that I see videos from other instances. So at least the hosting cost is distributed across instances.
The other issue then is the bandwidth. Peertube uses p2p among viewers, so if there are many viewers at the same time they can take a significant load off from the server. Instances can also cache each other’s videos to split the bandwidth cost between them.
I think these design decisions means that it is possibly viable, though it is definitely way more expensive than non-video federated communities.
I don’t think such an aggregator is required. Interoperability is smooth enough that you don’t have to think about different instances most of the time. I’ve only really noticed two points that would be confusing:
So I think what we really need to do to make this platform intuitive to people that aren’t already familiar with it is:
I have to image that most people with no intention of staying would not make accounts. Registered users are probably at least contemplating staying here for more than a couple of days. So the question for me is: will the communities here build a critical mass to sustain themselves after the blackout?
For me personally it’s very simple. I will keep using Lemmy if there is enough activity to be worth it. I don’t need it to be perfect or as active as Reddit or whatever. Just meaningful enough to warrant my time. I will keep my Reddit account for the time being, though.
This is a great video on the carbon footprint campaign:
I really don’t understand the benefit of being federated
The benefit is to prevent this from being the next Reddit. Being a nonprofit doesn’t really guarantee anything in the long term. OpenAI was a non profit and now it isn’t… Rather than trust a single entity to not abuse its power, federation aims to not give any entity all the power to begin with.
It also solves practical problems. Which single benign entity would pay for the servers and internet connections to become the new reddit? I don’t think there is one.
It depends on what your bottleneck is. For example on my system I get
$ systemd-analyze Startup finished in 11.976s (firmware) + 3.879s (loader) + 2.013s (kernel) + 157ms (initrd) + 6.354s (userspace) = 24.382s graphical.target reached after 6.316s in userspace.
The kernel boot process is only responsible for 2s of my boot time. So even if this does end up improving boot times, there’s very little it can do. The real improvement for me would be to choose a faster-booting m/b. You can run
systemd-analyze
on your setup to see if the kernel boot time is more significant for you.