$TWTR's all-inclusive expenses were 25x higher than this. $5.6B/year = $470M/month, in support of 260M daily active users. $1.80/user-month, as compared vs. $0.07/user-month on here (I'm only counting the 30k "active").

I feel like it's a bargain social media, both in feature-set and cost.

(I don't understand the server cost breakdown either, and I don't pretend do. Just doing an empirical comparison).

The incentives of optimizing for future revenue are different than those when you are optimizing for minimizing expenses. These values are not very comparable.

Which of those two is the one optimising for future revenue?

Not trying to be a smart arse - it’s just that the obvious pick (Twitter) has been aiming for revenue from the beginning and haven’t really delivered.

Twitter.

>has been aiming for revenue from the beginning and haven’t really delivered.

In 2021 twitter had 5.08 billion dollars of revenue. Their ad revenue divided by DAU was about $20. No mastodon instance is close to getting $20 per active user per year.

I’ve made the most basic of errors. I was thinking I was discussing profit.

I find the comparison very interesting, in the broader debate of "how can not for profit services with no ads compete with giant for profit advertising based serices". I have really appreciated the analysis in this thread.

Automated backups, monitoring, a person "on call) to manage your instance and handle any issues?

The Fosstodon admins run the moderation side, but masto.host (their service provider) is paid to keep the lights on.

It's easy to say "buy a box and connect it to the internet", but that's consumer grade equipment. If my computer goes down, it only affects me. When an instance goes down, it can affect thousands of people.

I'm not saying they couldn't find a less expensive solution with AWS or Azure, but they need a solution now. They don't have the time to learn a new solution, design the system, and then implement a new backend while people are swarming to their instance. Migrating to another solution (even the above "box") means significant downtime for 50,000 people.

No, I'm sorry, these are not reasonable costs. Is there a reason why the mastodon community has utterly no clue what sort of performance/cost should support an utterly basic social media platform?

You are always welcome to use the opportunity and start a competing service in a fraction of the cost. Mastodon is indeed inefficient but you also pay a lot for peace of mind.

More to the point: the masto.host guy has been there working on making starting an instance accessible from early on. People don't seem to mind paying a little premium to support someone who's done a lot of work to jumpstart the fediverse. Starting a competing service might be easy from a technical aspect, but reputation counts for a lot here.

Instead of condescending to them, why not try to help the project be more efficient with resources? It is an open source project after all.

Yes? It’s written in RoR.

Is the ActivityPub spec just as bad, or would it be performant if written in a proper language?

There are other clients, like gotosocial written in Go, so a comparison is possible, but not sure if someone has already done that.

Pleroma is an Mastodon clone written in Elixir https://git.pleroma.social/pleroma/pleroma

>Mastodon just the most inefficiently written software ever or something?

Yes, it's notoriously bad.

The backend is written in Ruby on Rails not something like Rust, this explains why it has such beefy hardware requirements for 50k users.