That article also kinda reads like it was generated by an LLM.

No, the red flag is being ‘self-hostable’, but trying to concentrate all your users in a central, non-federated, monetised instance.

Also, we cannot verify how long they where in development because they squashed their git commits. There’s usually no good reason to do that.

I host two homeserver, one on synapse and one on continuwuity, both pretty small (tens of users), but with users in lots of large rooms. The second one was significantly easier to set up, and uses a lot less resources.

Also, element and element X work, but aren’t great. It depends on the user, of course, but I don’t think you get people by giving them the ‘dumbed down’ version.

Isn’t fluxer partially LLM-coded according to the dev?

Ignoring the issues already raised in other comments, no, open source devs shouldn’t, probably. I mean, they’re free to do whatever they want, but it’s still a centralised solution with a nitro-like monetisation scheme for their central instance, and they’ve not shown how their promised federation would work. We don’t need a bunch of new silos, and currently, Fluxer seems like it wants to become one big silo with self-hostability as a selling point.

They should keep making existing federated solution, eg. Matrix, better.

Out sound like you are outsourcing comment-writing to an LLM.

Which I would appreciate … if I had the least idea about Christian sects in the US.

If you need to keep windows around, you can also do what I do for my VMs. Download Win11 Enterprise IoT LTSC from massgrave.dev. Doesn’t have Bing search in the search bar, just local, no Co-pilot, no bullshit.

I’d prefer not having to use it at all, but some software just doesn’t play nice with wine yet.

Not the original commenter, but it feels less like a review and more like a long sales pitch. Parts of it also read like LLM-generated text. Both don’t make it feel all that trustworthy.

LLMs are stateless. The model itself stays the same. Doesn’t mean they’re not saving the data elsewhere, but the LLM does not retain interactions.

The bigger issue is that reports currently don’t federate. A report will always go to the admin of your homeserver (which might be you) not the admin of the homeserver the room you’re in is on, nor the admin of the homeserver of the other users.

Most larger (1-2k people) communities get around that by just having you ping the mods in reply to the offending content, which is a band-aid.

A spec for federated reports is apparently being worked on, but not yet available.

A discord server-like would be a space, containing channel-like rooms. The main difference is that rooms can also exist independent of spaces, if you just need a single chat for some people instrad of a group of chats.

You can set permissions for a whole space, it’s just that they currently work differently than Discord. Members have a power level, and you set the power level from which each function is available. So, e.g. Sending messages from Pl 0 (representing normal users), banning users from PL 50 (representing moderators), changing server settings from level 100 (representing administrators).

It sounds complicated, but once you get used to it it’s pretty easy.

Matrix is not

With LiveKit for calls / screen share, it is for my group. Though I’m not saying it doesn’t have issues.

will never be

Community-developed homeservers like continuwuity have gotten a lot of new support on the last few weeks. Clients like cinny are getting pretty close to a replacement ux wise (if you look at PR2599 on Cinny’s GitHub, they are working on and will soon merge support for LiveKit in a way that is very close to voice channels).

I also generally think that the only way to replace Discord as an ecosystem where you talk to many people from different communities is a federated protocol, not a bunch of new silos, one for each community.

Yup, typo. It is lit, though.

I have deployed LiveKit for my homeserver yesterday. I’m not gonna lie, it does involve a bit of work, but once it’s running it is very seamless.

I don’t see the deception. ‘vegan sausage’ ah, it’s a sausage but without meat. ‘vegan burger patty’ ah, it’s a burger patty but without meat. ‘oat milk’ ah, it’s like milk, but made from oat.

It is an easy and accurate way to describe what they are selling - a product intended to replace a specific animal product. No one is going to call a block of tofu a vegan steak - if that’s the name, it will be a product as much like a real stake as the producer can muster.

Why make things more complicated than they are?

I feel like stuff like this works for people who want to move one very specific group of discord, but to replace discord’s ecosystem in a more general sense, federated protocols like matrix are necessary.

I had a brother laser printer with a pre-heating roll that went bad. Sourcing a replacement for that was pretty annoying. But I get your point.

Why are you so adamant on this? All vegan meat replacements I’ve ever seen are very clearly marked as such, and most supermarkets around here have them in their own section.

Are you usually just stuffing your cart with things immediately after recognising a single word on the packaging? Cause that’s how you end up with shitty sausage, even if you do buy meat.

I feel like it makes sense to name replacement products in a way that makes clear what they’re replacing. Mandate putting the word vegan in front of it in the same font size if you’re really worried about confusion.

“I have helped pay for something good. More people could benefit from it, at no additional cost to me. But I’d rather they not.”