Mark Tomczak @mtomczak@qoto.org

Alright. :)

New server's been up and stable for a week, so I think it's time to move out. Time to test the account migration part of this system.

Catch y'all at @mark :)

Self-hosting a Mastodon server, thus far: the good, the bad, the ugly.

**The good:**

- Migrating from another server is very straightforward. Pieces of your account can be exported as CSV and consumed by another Mastodon server.

- Hackable server. I'm running a variant that adds some features (I've also bumped my character limit, because that's a function of the server features and not the protocol). Although it's Ruby on Rails with JavaScript as the UI layer, it's not deeply, deeply difficult to hack around on.

**The bad:**

- Setup is a chore. You can use Docker, but at the end of the day it's far from what most people without server admin experience would consider "turnkey." You'll still have to do the part where you push a bunch of nginx rules up and set up some security certificates by hand. In my case, I'm behind Cloudflare and that adds a bit of additional complexity because it's off the getting-started script (though other people have blazed this trail and helpfully put together guides for it).

**The ugly:**

- Stuff I think of as "par-for-the-course open-source usability sharp edges" are there. Little things that won't stop you from using it but indicate that nobody's paid to do the UX, like the fact that in the import UI and the export UI, the order of the "block list" and "mute list" options are flipped, increasing the odds that if you're ripping down CSV files by hand and then slamming them into the other UI, you'll swap the two. Fixable? Probably easily. But with all these eyes on the product nobody empowered to fix it has noticed or cared.

... all in all though, this is *way* low on the list of most painful open source projects I've ever tried to set up an install of, and I'm very, very impressed.

@trinsec So far so good! I really need to dig in and read the source code to grasp all the nooks and crannies but I like what I see. Hasn't said itself on fire yet, although there are a couple of things about it that differ from the base mastodon setup instructions (I'm going to try and get those blogged sometime this week, time permitting).

I am trepidatiously setting up my own account on my own Mastodon node: @mark. Not officially switching over yet because I want to confirm it won't set itself on fire, but feel free to toss me a follow.

@jcreed They is really cool. I'd bet curious to know what options.

@lauren Oh yeah. This is why I think the question of copyright pedigree for image generators is ultimately moot; it's fighting the last war.

Assume there's a sea-change and it is decided you can't scrape the web to build an image generator. Okay. Then the twenty largest media and ad companies in the world will *pay* ten thousand artists to feed their beasts with full rights signover. It'll only cost ten thousand, and yes, those ten thousand will be making the machine that shifts work away from all their peers... But tell me you *can't* imagine finding ten thousand artists who would take that deal.

It only takes one generation of scabs to lock-in the machine for quite some time and fundamentally change the way media is created.

... I was excited about this until I compared notes with a friend who does "say potato" to distinguish humans from AI voice recognizers and we realized we were inventing CAPTCHAs for voice callers because **this is hell the future is hell we're in hell.**

@jcreed

New strategy for identifying robocallers is hitting them with a #bluey quote.

If they continue on-script right past "Coconuts have water in them," the next message they get is "Add this number to your do-no-call list."

@linuxgal before we get too high on our horses, I would suggest people consider the type of data that can be gleaned from the vast amount of pulling of packages and libraries that the average Linux user must do to make the machine serviceable.

Are we so deeply confident that every single package source we rely upon is not just collecting aggregate data on us?

A "You wouldn't download a car" dedicated meme editor.

https://youwouldntsteala.website/editor.html

@retr0id Blessings to you for this.

@kevinrothrock Even when they're world leaders, folks *gotta* stop tweeting on no sleep.

@Dragon @ben @theinstantwin

Yes, you have the thrust of it. News is dying, and the audience you're describing on the other app won't pay to consume it (not in nearly the volume needed to allow people to make journalism a career) and their views can't be audited in a way monetizeable via advertising.

So there's little value seen in a newsroom spending a nonzero amount of money (in setup and admin) to chase a projected zero dollars.

I suspect this status quo to maintain until and unless someone proves that you can use Mastodon as some kind of loss-leader to induce people to spend money on the business side of the newsroom.

@bhollis Smells like the work of the tire self-actualization lobby.

@Dragon @ben @theinstantwin

> I can only assume this newsroom doesn't have a tech section. Which is odd.

Why would that be odd? Most newsrooms arose from the print space, which has no particular overlap with electronic communications. They're often late to the game and working with minimal resources to get in.

Meanwhile, newsrooms born online found it difficult to generate revenue and have been going through a collapse / consolidation cycle for at least a decade.

On the one hand you have an industry unwilling / unable to spend money on technical folk, and on the other an industry unable to make sustainable revenue.

@Dragon @ben Until there's a box you can buy that sits on your desk, plugs in with one plug, self-administers, and hosts your whole Mastodon solution with at most one additional employee added to staff, I think I'd pull the scare quotes off of 'effort' in that sentence.