A software developer and Linux nerd, living in Germany. I’m usually a chill dude but my online persona doesn’t always reflect my true personality. Take what I say with a grain of salt, I usually try to be nice and give good advice, though.

I’m into Free Software, selfhosting, microcontrollers and electronics, freedom, privacy and the usual stuff. And a few select other random things, too.

  • 0 Posts
  • 238 Comments
Joined 10 months ago
cake
Cake day: June 25th, 2024

help-circle

  • hendrik@palaver.p3x.detoFediverse@lemmy.worldFirst draft woes
    link
    fedilink
    English
    arrow-up
    2
    ·
    edit-2
    4 days ago

    I think it needs to work across instances, since we’re concerned wit the Fediverse and federation is one of the defining mechanics. Also when I have a look at my subscriptions, they come from a variety of instances. So I don’t think a single instance feature would be of any use for me.

    Sure. And with the cosine similarity, you’d obviously need to suppress already watched videos. Obviously I watched them and the algorithm knows, but I’d like it to recommend new videos to me.



  • Wasn’t “error-free” one of the undecidable problems in maths / computer science? But I like how they also pay attention to semantics and didn’t choose a clickbaity title. Maybe I should read the paper, see how they did it and whether it’s more than an AI agent at the same intelligence level guessing whether it’s correct. I mean surprisingly enough, the current AI models usually do a good job generating syntactically correct code one-shot. My issues with AI coding usually start to arise once it gets a bit more complex. Then it often feels like poking at things and copy-pasting various stuff from StackOverflow without really knowing why it doesn’t deal with the real-world data or fails entirely.


  • I’ve also had that. And I’m not even sure whether I want to hold it against them. For some reason it’s an industry-wide effort to muddy the waters and slap open source on their products. From the largest company who chose to have “Open” in their name but oppose transparency with every fibre of their body, to Meta, the curren pioneer(?) of “open sourcing” LLMs, to the smaller underdogs who pride themselves with publishing their models that way… They’ve all homed in on the term.

    And lots of the journalists and bloggers also pick up on it. I personally think, terms should be well-defined. And open-source had a well-defined meaning. I get that it’s complicated with the transformative nature of AI, copyright… But I don’t think reproducibility is a question here at all. Of course we need that, that’s core to something being open. And I don’t even understand why the OSI claims it doesn’t exist… Didn’t we have datasets available until LLaMA1 along with an extensive scientific paper that made people able to reproduce the model? And LLMs aside, we sometimes have that with other kinds of machine learning…

    (And by the way, this is an old article, from end of october last year.)







  • hendrik@palaver.p3x.detoSelfhosted@lemmy.worldOff-grid hosting
    link
    fedilink
    English
    arrow-up
    12
    arrow-down
    1
    ·
    edit-2
    16 days ago

    Some people do it. For example we have this solar-powered website: https://solar.lowtechmagazine.com/

    You’d need an energy source like a solar panel, a battery and some computing device. Like a single board computer (Raspberry Pi) you can also run webservers on smartphones, or even a microcontroller. The server part works without an internet connection. But you obviously need some way to connect to it. A wifi (router) or a computer connected via an ethernet cable.

    The tech isn’t too complicated. Just install nginx if you have a raspberry pi, open a wifi and put your website on it. If you choose a phone, try Termux and a supported webserver. Both Linux and smartphones are designed to even work without an internet connection ;-)


  • Not really. I could use some good selfhosted search engine. I mean all the existing projects (which is just YaCy, to my knowledge) are a bit dated. Nowadays we only got metasearch engines and we’re relying on Google, Bing etc.

    But I don’t need any chatbot enhancements. That’s usually something I skip when using Google or Bing because it doesn’t work well. The AI summaries tend to be wrong, and it’s bad at looking up niche information, which is something I need a search engine to be able to find. The AI just cites the most common slop, or at best the Wikipedia article. But I don’t really need any fancy software to get there… So for me, we don’t need any AI augmentation.

    And I think the old way of googling was fine. Just teach people to put in the words that are likely to be in the article they want to find. That’d be something like “Rust new features 2023” or “homelab backup blog”. Sure you can strap on a chatbot and put in entire natural language questions. But I think that’s completely unnecessary. We have brains and we’re perfectly able to translate our questions into search queries with little effort… If somebody teches us what to type into the search bar, and why.


  • Yes. Plus the turing machine has an infinite memory tape to write and read. Something that is in scope of mathematics, but we don’t have any infinite tapes in reality. That’s why we call it a mathematical model and imaginary… and it’s a useful model. But not a real machine. Whereas an abacus can actually be built. But an Abacus or a real-world “Turing machine” with a finite tape doesn’t teach us a lot about the halting problem and the important theoretical concepts. It wouldn’t be such a useful model without those imaginary definitions.

    (And I don’t really see how someone would confuse that. Knowing what models are, what we use them for, and what maths is, is kind of high-school level science education…)



  • It’s a long article. But I’m not sure about the claims. Will we get more efficient computers that work like a brain? I’d say that’s scifi. Will we get artificial general intelligence? Current LLMs don’t look like they’re able to fully achieve that. And how would AI continuously learn? That’s an entirely unsolved problem at the scale of LLMs. And if we ask if computer science is science… Why compare it to engineering? I found it’s much more aligned with maths at university level…

    I’m not sure. I didn’t read the entire essay. It sounds to me like it isn’t really based on reality. But LLMs are certainly challenging our definition of intelligence.

    Edit: And are the history lessons in the text correct? Why do they say a Turing machine is a imaginary concept (which is correct), then say ENIAC became the first one, but then maybe not? Did we invent the binary computation because of reliability issues with vacuum tubes? This is the first time I read that and I highly doubt it. The entire text just looks like a fever dream to me.


  • Yeah, seeking support is notoriously difficult. Everyone working in IT knows this. I feel with open-source, it’s more the projects which aren’t in a classic Free Software domain, who attract beggars. For example the atmosphere of a Github page of a Linux tool will have a completely different atmosphere than a fancy AI tool or addon to some consumer device or service. I see a lot of spam there and demanding tone. While with a lot of more niche projects, people are patient, ask good questions and in return the devs are nice. And people use the thumbsup emoji instead of pinging everyone with a comment…

    I feel, though… I you’re part of an open source project which doesn’t welcome contributions and doesn’t want to discuss arbitrary user needs and wants, you should make that clear. I mean Free Software is kind of the default in some domains. If you don’t want that as a developer, just add a paragraph of text somewhere prominently, detailing how questions and requests are or aren’t welcome. I as a user can’t always tell if discussing my questions is a welcome thing and whether this software is supposed to cater for my needs. Unless the project tells me somehow. That also doesn’t help with the beggars… But it will help people like me not to waste everyone’s time.





  • I’d go with the Full Disk Encryption. You can be sure everything is encrypted that way. Any additional complexity adds ways to mess up and compromise security. Entering the password is a bit cumbersome. But that’s part of the deal. I just carry my computer keyboard to my NAS and enter the password each time I need to reboot. Which doesn’t happen that often. There also used to be some tutorial somewhere on how to put a Dropbear SSH server into the initrd so you can enter the password over network.