That’s interesting. This post had suggested it isn’t yet possible to host an AppView. It seems the reality is more complex.

Whether it soon becomes possible to self host an AppView, the one remaining centralized component will tell us a lot about where it’s headed.

ATProto is almost there with the only missing piece being the AppView. I’m not sure if BlueSky is hesitant about releasing theirs as open source, but I don’t think there are any barriers to a third party implementing one.

Yes. I think it’s good not to form opinions about subjects you don’t know much about.

When it comes to voting in an election, it’s possible to make good decisions about candidates without forming opinions about every policy issue. That’s kind of the point of representative democracy.

That’s enlightening. It links to an article about self hosting a relay, which explains that, as I suspected, a relay does not have to mirror the entire network. It also seems that using a relay at all is an optional optimization.

It looks like the BlueSky AppView is not (yet?) open source. I wonder why nobody has built an alternative yet.

I don’t think many people have read RFC 5322 (I haven’t), but most non-technical people I know understand these things about email:

• There are different service providers, and people can email each other no matter which provider they use
• There are different email apps
• Some apps are tied to specific service providers and others are not

I do lament the overall level of tech literacy.

The average person understands email pretty well. Mastodon doesn’t require much more understanding than that, but could probably use some UX and messaging work.

That’s a bit of a circular reference: “it got popular because it got popular”. The question remains: why did BlueSky reach that threshold and Mastodon did not?

I’m inclined to agree that’s a problem. Everyone’s first encounter with a social media content recommendation algorithm was one designed to manipulate them into clicking ads, so it caused some backlash. Recommendation algorithms can be tuned to show things people care about and want to engage with.

I’d still echo the (current) top comment’s advice to use something open source, local, and encrypted.

They’ve already taken the hard stance. If they roll it back, they will lose the trust of their users.

Biowink GmbH is probably not a corporation registered under US law. If I had to guess, the government of Germany will not be particularly eager to force them to turn over data to the USA. The Germans take their Datenschutz very seriously.

Having moderated a number of online spaces over the years, sort of. It’s usually the harshest thing a moderator can do, but it does not have very much real world impact on most people. In many parts of the internet, it isn’t even very effective at keeping the same person from coming back with another account, which isn’t a big deal if they don’t come back with the same behavior.

I’m not particularly shy about reaching for the permanent ban if it seems like someone is being an asshole on purpose. I’m not getting paid for it, and I do not have much patience for dealing with people who don’t want to be respectful toward their fellow humans. There’s usually a way to appeal if it’s a misunderstanding. That’s especially true in systems like Lemmy and unlike traditional web forums where one account and UI provides access to many communities, leading to drive-by comments.

I’m also fond of somewhat ambiguous rules like “be excellent to each other” or “don’t be an asshole”. Without that, if a community gets active enough, someone will show up, act like an asshole, and argue about the rules when they get banned.

It seems to me software designed to facilitate discussion shouldn’t have a downvote buttton. There should be a UI for marking comments as inappropriate, but it should require a second step saying why. Perhaps one of the reasons should even be “I disagree”, but that option should have no effect.

It’s not impossible to abuse of course, but it nudges people in the right direction. Those UI nudges can be pretty effective.

In the context of ad-supported algorithmic social media, offensive is the wrong question. It’s about brand damage.

Showing an ad next to something that actually offends people can damage a brand, but even something a little edgy might turn off customers of a brand with a more formal or conservative audience. The algorithm’s ultimate goal is to get people to watch ads, so something a little edgy might reduce the reach of that content. Censoring it prevents the algorithmic downrank.

It’s probably more fair to call its claim of being decentralized fake or disingenuous. There doesn’t seem to be any way to participate in the ecosystem without going through the Bluesky relay, a central point of failure with strong incentives to eventually do something shitty.

propagating and caching images

This leads to the problem of abusive users trying to cause legal issues for server owners by introducing CSAM.

nobody is refusing to offer their sevices on linux because it is vulnerable

That’s not quite true, though in that case it’s about the service provider being unable to verify that the user isn’t running a operating system configured or modified to work against the interests of the service provider.

Maybe, but the archetypal non-technical user, my mother does want to run a third-party ROM. Her phone is out of its official support period, and she knows that security updates are important and would like a way to get them. Most people, at least in wealthy countries do have a technical person in their lives they can ask things like that. She doesn’t want to buy a new phone because it would be too big and lack a headphone jack, a position I share.

I had to recommend against running what I run (LineageOS, Magisk, Play Integrity Fix). Without PIF, too many apps will refuse to run on LineageOS. She doesn’t need root for much else (maybe adblocking) and doesn’t have the knowledge to make good decisions about whether to grant root permissions to an app that asks (Magisk doesn’t have an allowlist-only mode, but it should). Finally, keeping root through an update is fussy. It’s not hard, but it’s an extra step that has to be done in the right order every week or two.

Unlike Firefox in 2024, a third-party Android build that’s easy enough to install and isn’t sabotaged by Safetynet would something many non-technical users care about: an extended useful life for their devices.

Last time I used one was because I forgot my physical wallet and needed to pay for something. I don’t want to tell Google about my shopping habits, but I like to have options in case of emergency.

I’m running LineageOS (with GMS), Magisk, and Play Integrity Fix.