Might be true, honestly. I'm on NixOS using the proprietary drivers for my 3080 and 4090. No issues, took one line of configuration. I do have to stay on X11 unfortunately until Wayland supports the real drivers at least (though i hear that's being worked on, maybe already working?).

For all of NixOS' pain, it really does make some things awesome and simple.

I did that at work and people had a heart attack lol. Which i thought was silly because the mod literally had no code in it besides module file declarations. /shrug

Expecting every different platform to all move in lockstep is not feasible so a situation where some support downvotes but some do not is going to have to be part of the process. Even if it was just kbin to kbin initially, that'd still be an improvement.

Yea, that is what i was thinking. I myself am working on some ActivityPub stuff and while compat is important to me, i also plan on potentially exposing new features in this manner. If other software wants to federate with it, cool, if not, whatevs. It's data propagation, as with everything in ActivityPub it's optional that consumers use it as desired.

Is ActivityPub not open to arbitrary extensions? I thought it was a very loose spec and the only concrete spec is an agreement between instance types. But i've not looked at the protocol closely yet, despite developing around it hah.

Weird, i can. Firefox, Linux.

Yea, i'm working on my own Fedi software and i'm struggling with the point of boosting in the link aggregator context. It's an odd overlap with Reddit-style reposting to appropriate subs, but based on the user.

It makes sense in the Twitter UX, but i struggle to find it's place in the Reddit UX.

Yea, i'm not sure how much benefit of doubt we should be handing Mark Zuck of all people. There's few people in the world who make their intentions more clear than him. Not that i'm trying to paint him as evil, i'm not and i don't think he is, but i also see no reason to expect self-run instances to offer an olive branch to him.

We should be vary paranoid about Embrace Extend Extinguish in these communities.

I'm (as a dev experimenting with the fedi) more interested in the features each platform adds to the fedi and ensuring we have consistency. Which is probably included in what you said, i'm just being explicit.

Ie regardless of if it's an upvote or a star, hypothetically it should have a similar effect in all platforms that federate with that value. If it modulates score in an algorithm, it should "mean" the same thing to the fediverse so that it has similar outcomes in all platforms. A silly but extreme example of this failing would be one site upvotes and downvotes meaning opposite things on two implementations - it would be chaos hah.

I hope (but don't yet know until my implementation is further along) that Like and Boost are functionally different. I also hope no ones implementation is conflating them.

Either way interesting times and i'm excited to see where we take it :)

That brings into question what downvotes are, then hah.

What are your thoughts on reactions instead of rewards?

I'm on mastodon.art and donate $1/m, same story. It's only a handful of people that donate (hundreds, but still - much smaller than the thousands of users) and we cover the bill.

Though, this is also why i'm experimenting with custom Fediverse instance software that prioritizes low cost operation. I think Fedi would be better off if it wasn't a huge lift to figure out hosting. There's enough challenges in hosting instances, it would be nice to reduce as many as possible.

So you're running a local DB and backing up the external drive? Neat. What about static storage? Images/videos/etc? How are you handling backup?

I'm interested in this space because i'm writing an experimental ActivityPub client and one of my goals is to make it "as cheap as possible" to run a micro-instance. I plan on exploiting a single directory, sqlite and file storage, that users can just rsync or backup however they wish. Cheap is tough hah.

DB would be a big one, no? Not sure what Kbin is using, but DB costs tend to be meaningful as well, especially since depending on how the app is designed the usage requirements can sort of parity the app hosting requirements itself. Plus storage, but that can be more variable depending on how robust you want it to be. Ie S3 is pretty robust but it's expensive, etc.

I actually like them when there isn't a time limit. Chivalry 2's battle-pass-thing is nice for that reason. It's just something i can throw on while i play normally with no grind

Yup, it's like email but take away recipients. Yea, there's sorta recipients, but you don't really know who it's federated with/etc. We (foss devs) need better optics here. UX is difficult, though i welcome ideas.

Seems kinda wonky. I'm working on writing my own flavor of an ActivityPub client. Is there somewhere you know of that this is being discussed, developed, etc? Iirc ActivityPub itself doesn't define this, right? I figure what we're talking about here is just the data model some apps have created within the ActivityPub spec - and by keeping it consistent apps can federate with each other. I'm curious on advancement of that model to introduce new concepts and features but retain some compatibility.

Not, at the moment. Configuration is pretty good, it has really good defaults (LSP/RA integration/etc), but Plugins are still being designed.

Same. Helix and RA.

This is the way. Tons of people will stay on Reddit, Reddit will be fine. Just look at Facebook, how long it has lasted and what it has done, etc. Reddit is the new Facebook (platform, not company) and tons of people will be fine sticking with it. That's okay.

The beauty is we have options, and we're working to improve those options.

Which is exactly why this is better than Reddit. It'll definitely have growing pains and long term UX warts, but not being siloed and leashed to someone like Spez's decisions is a really great feature of the Fediverse.