totally agree, Prosody is great, and in a lot of ways, it's been a real source of inspiration.

I'd actually feel pretty safe with a LLM project if spec'd against the protocol documentation as opposed to a rookie's handcoded implementation. If security is the concern, I don't think using random reimplements is very safe to begin with.

> Something that is supposed to be very private and secure

Not sure how using an LLM for development of the tool has any affect on this. From a security perspective, I think that would've had more validity a year or two ago, but these days, I'd argue LLMs practice better security in generated code on average than a normal developer.

From the privacy angle, I have no idea how having an AGENTS.md (for development purposes) has any impact on privacy.

I don't want to "burn [your] karma to the ground", I just want to understand this angle, because I'm not sure I see how it applies.

The security comes from E2EE - which happens in the clients not the server

Do you not use AI for coding?

So true. The existing clients are actually the only reason I haven't asked all my friends to join my server. Only the ones who will put up with jank.

Gajim is _okay_ on desktop, but it leaves a LOT to be desired. Everything else feels way too bare-bones to feel comfortable recommending to Discord expats. Fluux Messenger is the closest I've seen, but it's sluggish for me. Maybe that's just Tauri on Linux.

Conversations is good on Android (bit barebones feeling), and I haven't tried out voice with it. IDK what the situation is in the Apple ecosystem.

I really, really wish I had the time to just start building a strong suite of XMPP clients. Every day Discord shoves a new ad in my face I get closer to biting the bullet.

from what I know, developers don't like xmpp to build clients for it, especially more complex ones.

I use Dino on the desktop and the experience has been almost as good as Conversations on mobile. For what it's worth.

Matrix is way more expensdve to run by its protocol design

> Looks like matrix is in this niche now.

Matrix is almost as old now as XMPP was when Matrix first came out, and yet, after all those years and grandiose promises, and loads of venture capital, it doesn't strike me as approaching the problem in a better/sounder manner than the protocol it was set to replace (quite the opposite, actually, Matrix, the protocol and its implementations are worse on almost every metric).

Matrix still does not allow me to choose a display name for 1to1 conversations. It renames the room for both parties. So I have to live with 10 “Johns” in my contact list.

i remember that too. XMPP is still alive and kicking, just a lot quieter these days. :)

Eh, I've used both a lot and matrix is kinda flailing from my perspective. At least for casual use. If you are a client of Element maybe the experience is better, but with their "next-gen" battery friendly client they've been working on for years, I can't even get notifications working through public unified push servers when using matrix.org. Meaning that just to get started using it I have to run my own matrix or up server.

I'm a big fan of Tuwunel. Run Synapse myself since I've had a server running for about 7 years now, but if I had to start over, Tuwunel all the way.

There's also Continuwuity, which I think is more popular, but I don't fully understand the difference. I think it's a community split issue? As someone who maintains a client, Tuwunel seems to match the spec better, but I'm sure folks have better reasons to pick one over the other.

All the stuff trying to implement a protocol are usually better off using a language that has sum types imo, and I think that's a big enough argument to dismiss C.

Considering that LLMs more often than not forget stuff when writing code, I definitely prefer Rust's automatic memory management system compared to balancing mallocs and frees throughout the code.

You could do RAII in C++, but the compiler won't warn on unsafe code.

Basically any deterministic process that can look at your code and guarantee a certain useful property is worth it's weight in gold with AI driven software development processes because it significantly reduces the review burden.

Meanwhile the C ecosystem is currently collapsing from AI finding critical RCE vulnerabilities in infrastructure projects written in C.

Hyperbola GNU and C99 half-lover there. C it's fast, but for networking clients/servers Rust it's far more secure, even more with current AI vulns. Under amd64 it should be almost mandatory. Half of Gnome would be far better with gtk4-rs. For legacy i386 users, it's either Go (aerc for email rocks) or Algol68.

C is a loaded gun pointed at your foot regarding memory management. Rust directs your gun somewhere safer. That is why.

I'm keeping an eye on Iroh https://www.iroh.computer/

> ...I have been absorbing whether tried-and-true old tech like these are going to revive in the AI era, just because it's "locked in" in the training data. AI is going to keep some of these torches burning forever maybe.

I agree with you 100%! With so much open source software and know-how material that has been available for AI to train on, it seems inevitable that such tools will be used that much more often...and then as the AI trains on newer content, it will continue a virtuous cycle of open source software that will keep being used, and further improved...and as you stated "burning forever".

why not use the bajillion other servers and use this vibecoded one? why not make a great client with iced for example instead? are you just rewriting in rust using AI for no good reason just to farm attention?