nefarious

I think that might be the codecs' fault. At least for me, my headphones sound terrible in headset mode on all the devices I've tried, regardless of whether they're running Linux, MacOS, iOS, or Android.

Statcounter bases their data on web traffic. If you're browsing the web on your Steam Deck, I think that should count.

This feels short-sighted. The odds of the protest having a major and immediate impact were always low. It's not like the suits were going to have a sudden change of heart and realize they were alienating their users. The majority of Reddit's userbase weren't going to suddenly leave the site forever. But that wasn't the point.

Here's what's changed since the API changes were announced:

1. Reddit's responses to user concerns and protests have alienated even more users than the initial changes themselves, showing users exactly how Reddit's administration sees them.
2. A whole bunch of mods, devs, and contributors who put in a lot hard work improving Reddit for free are now much less motivated to do so (if they're still willing to do it at all).
3. The protest raised awareness of federated Reddit alternatives, which have grown substantially as a result. A lot of those people who helped improve Reddit for free are now turning their attention to kbin and Lemmy instead.
4. Reddit is on a clear trajectory. They've shown they will continue making user-hostile decisions and antagonizing their userbase in pursuit of further growth.

We now have an established alternative to Reddit that has reached a critical mass for growth. A lot more people are now working on making the fediverse better, and communities are forming that will attract new users on their own. From now on, every time Reddit makes another move like this, more people will move over (or get closer to moving over) and Reddit will drop in quality even more as a result. If there's ever a Digg V4 moment (maybe when they kill old.reddit), the fediverse will be much more prepared to take on the mass exodus that results.

Gotcha. I honestly just paid for Obsidian Sync, so I haven't really evaluated the other options myself. I just know I've seen people mention that it was non-trivial or impossible to set up things like Git or Syncthing for syncing.

Honestly, I should probably set up a system-wide adblocker, but I just use uBlock in Firefox and avoid apps that shove ads in my face.

Adding another +1 for Obsidian. I've tried a whole lot of note-taking apps over the years and Obsidian is the best I've ever used by a pretty significant margin.

That said, there's one caveat to note about self-hosting: if you're on iOS, there's not much of an option for syncing besides their paid Obsidian Sync service. I think there are some hacky workarounds for this, but they don't seem great.

Another option to consider is Joplin. I used it before Obsidian, and while it's not nearly as slick as Obsidian, it is fully open source, cross-platform, Markdown-based, and supports syncing with a variety of protocols out of the box. I had it set up to sync to a directory on my NAS using WebDAV and it worked reasonably well across all my devices. It also feels more Evernote-like to me than Obsidian and there's a built-in option to import an ENEX file if you want to move your notes over.

If possible, I think the karma/rep score should be completely hidden. As long as people can see a number, they're going to try and game it somehow, which incentivizes low quality posts. You can cap total karma, but people will still try to grind up to 5000 and they'll still try to get the highest comment scores they can. That encourages people to make the types of low-effort posts and jokes that often clog up Reddit threads.

The other problem with an overall rep score is that it doesn't truly represent user behavior. If 1/3 of my posts are shitty troll posts, but the other 2/3 are generic low-effort joke posts and memes that people will upvote by default, my rep will stay positive even though I'm a net negative contributor. Likewise, if I make one really popular post that gets 90,000 upvotes, my score will stay positive pretty much forever, even if I troll and harass people nonstop.

So rather than report the sum of a user's post scores, I would propose displaying a "user quality" indicator based on the average score of their recent posts instead. For example, if your average is greater than 5, you'd get a green up arrow, and if your average is less than -5, you'd get a red down arrow, but otherwise you get a neutral icon. You could have other icons for higher and lower scores, but I feel like that might still encourage people to try to game the system, so I'd propose keeping it simple and making it easy enough to get the green icon that you're not incentivized to spend any time on it.

Much like Reddit, Twitter's first mobile apps were developed by third parties. The term "tweet" even originated with one of those apps.

I was thinking the same thing. I think the user curation aspect is what made Reddit so sticky compared to old-school forums where you had to wade through every comment one by one, but having a visible karma score incentivized people to try to make the number go up.

I think Goodhart's law applies, because karma is ostensibly supposed to be a measure of how good a contributor you are, but in practice it just measures how good you are at getting people to upvote you, which it turns out doesn't require you to make quality contributions.

All of these things have already been disclosed.

ActivityPub is a public standard. https://en.m.wikipedia.org/wiki/ActivityPub

kbin is open source. https://github.com/ernestwisniewski/kbin

Lemmy is also open source. https://github.com/LemmyNet/lemmy

Google is your friend.

I had wondered about this. I figured that all of these surveillance capitalist adtech/analytics companies would have to have some metrics on this.

What would be really nice to know is how the numbers look now that the blackout has been over for a while. A 6.6% drop is pretty tiny if it only lasted a day or two.

Could've been UPS using USPS for last-mile delivery. The OP is also from feddit.de so maybe they're not in the US.