Skip Navigation

InitialsDiceBearhttps://github.com/dicebear/dicebearhttps://creativecommons.org/publicdomain/zero/1.0/„Initials” (https://github.com/dicebear/dicebear) by „DiceBear”, licensed under „CC0 1.0” (https://creativecommons.org/publicdomain/zero/1.0/)N
Posts
2
Comments
19
Joined
2 mo. ago

  • Thank you a lot for such detailed explanation! I finally got a definitive answer to XMPP vs Matrix debate. You definitely convinced me to try XMPP, seems indeed more reliable. In another message, you also mentioned you wrote a guide on Prosody, I actually would love to check it out :)

  • That was a great video actually. I thought that Zuli was kinda cool, shame it's a paid service. Not that I mind paying, just quite literally can't. Funnily enough, they hit the same issue with broken voice calls on mobile Element client

  • Thanks for reply. Unfortunately, we can't use it, should be exclusively selfhosted service :( I do like Signal, tho, great app

  • Let's just that we all are so fucked, that we'd use anything that works. So that's not a problem to convince them to switch, we all are ready to switch

  • That's definitely possible. But the weirdest thing was the inconsistency of every issue. As an example - voice calls worked on the web client, but didn't on mobile. Client issue perhaps, tho I tried, like, almost every mainstream mobile client

  • Thanks, will check it out. If it's not too bothersome, could you specify why XMPP would be a better choice than other options? The protocol itself, I mean. There's a lot of contradicting info on each of the protocols. Some say XMPP is ancient, choose matrix. Others say matrix is a complicated mess, choose more mature XMPP

  • Selfhosted @lemmy.world

    Any good selfhosted instant messaging?

    1. Eh, not that I'm worried or not worried, kinda indifferent by now. A lot of stuff happened in the last couple of years, I think I won't be surprised by any kind of outcome.
    2. I don't even know who the guy is, haha. Tho I did my fair share of research into cognitive studies, I'm way better at applicational side of things rather than theoretical. And as far as I understand, the guy is some sort of theoretical researcher?
  • Haha, thanks, I'll pass. Too much trauma from early days of training on 1060.

    Funnily enough, even tho we didn't ask the boss to buy us all 5070, we did ask him to set us up with a new GPU for the production server. And yep, we've got one, and took the old GPU for our testing server :)

  • Nope, I definitely don't want to pressure any one to use AI. I genuinely don't understand the point of it.

    As for a possible pivot. As the saying goes "if it works, don't break it". So I work. The job itself is definitely not the worst, actually the best in terms of conditions and benefits I've ever had. I actually enjoy the process of coding itself, doesn't really matter whether I do backend or machine learning or devops. I have background in data science, and AI is basically a glorified data science, so I guess it kinda fits my initial skill set

  • Most likely the whole AI stuff will reach some critical mass, when all the promises of the big corpos can't be fulfilled, the bubble will burst. But I think after some initial chaos the field will stabilize. AI is and will be a part of our lifes for better or for worse, it just won't be this overhyped.

    As for dislike, no one really cares actually, as weird as it is. The management wants AI integration, we give them the best we can do. For my peers a job is a job.

  • I don't work for a big tech company, so I'm kinda not thaaaat woried. But definitely worried a bit.

    I also work on backend stuff in general, so if worst comes to worst I guess I'll just be transfered to another department

  • Uhh... three

  • When integrating: LLMs and speech recognition models (like whisper)

    When developing: some deeplearning stuff, what it is exactly depends on the use case. Tho, again, mostly related to text and audio data

  • I'd fucking love to :( Had to postpone expanding my home lab cuz of the rising prices

  • I started in the field cuz it's easier to gather veggies there... Like, pickles, right? Aight, bad joke. You actually reminded me that I don't even remember why I started doing AI. I do remember tho, that when the uni offered some extra lessons on AI I was like "why would I do that, I hate this AI trend, gimme cybersecurity"

  • The only time I've broken a GPU was when I've tried to launch first Kingdom Come: Deliverance on an old gtx720. It literally caught flame, albeit small. I train all my models either on dedicated test servers or, if the model is small enough, on my rtx4060 that I also use for gaming. All's been good for a couple of years

  • Nope, it's actually quite hard to ignore the topic of AI. As it happens, I work on it. So why not share the insider's perspective, so to speak? Haha

  • I actually make my own models for one job and integrate other models into our services for the other job. We also train some opensource models on our data sometimes.

  • Ask Me Anything @lemmy.ca

    I'm an AI engineer and researcher. AMA