hendrik @ hendrik @palaver.p3x.de Posts 7Comments 2,171Joined 9 mo. ago
Well, they learn from the training material and then they apply what they learned. That's not exactly copying. Like me learning to code with textbooks and then being able to do it myself. LLMs are supposed to do something similar, generally they don't reproduce their training material verbatim. But it's complicated. And I believe we have some court cases and the legal system needs to find out how to apply copyright. Plus the big companies just steal stuff. It's not like me buying the books for university, Meta just downloads all of them via bittorrent. Which is definitely illegal. But i think the difference between learning, copying, being inspired by something is more nuanced. And if something like "fair use" applies, there isn't much an author can do. I guess LLMs are able to memorize stuff as well. And I don't think that is okay. I'm not sure if we have examples of that happening, that'd make a copyright case a bit easier.
I often recommend Mistral-Nemo-Instruct. I think that one strikes a good balance. But be careful with it, it's not censored. So given the right prompt, it might yell at people, talk about reproductive organs etc. All in all it's a job that takes some effort. You need a good model, come up with a good prompt. Maybe also give it a persona. And the entire framework to feed in the content, make decisions what to respond to. And if you want to do it right, and additional framework for safety and monitoring. I think that's the usual things for an AI bot.
If you manage to do it, maybe write some blog post about the details and what you did. I think other people might be interested.
Is memory that small, connected externally, or does that SoC just end up being a large package, with that much RAM on it?
Fair enough. I'm a bit unsure whether that happens on Lemmy. My old posts and comments rarely get any votes, interactions or corrections after say two weeks. These people must either be completely passive, or no one reads it after that. With a few minor exceptions. But you're right. This has happened to me, too. So you have a point here.
I halfway agree, but the issue with that is, that's not what happens in reality. In reality these things don't run on renewable energy. And not utilizing datacenters at capacity is just a waste of resources. And they could find people who donate their voices, which would be fair... But they're not doing that. So I think half the arguments still apply. It is innovation though, we shouldn't be opposed just for the sake of it. It needs some proper argumentation.
I kind of dislike it. I mean it's a good thing if they read it. If not, it just takes 5 minutes out of my day if I come up with a good nuanced answer here, and that's time I'm not going to spend answering other people's Linux questions. But it's alright, you made it completely transparent that this is a re-post. And it's a good thing to diversify. People often just ask in one big community, or even discuss everything in the super big technology communities even we have dedicated ones for certain specific tech topics.
Make Lemmy a nice place with a nice atmosphere, nice people and meaningful information and discussions. Once it shines, people will want to come here.
Why re-post this? I hope bpt11 reads the answers here or it's kind of a waste of time.
What kind of models are you planning to use? Some of the LLMs you run yourself? Or the usual ChatGPT/Grok/Claude?
It is like I said. People on platforms like Reddit complain a lot about bots. This platform on the other hand is kind of supposed to be the better version of that. Hence not about the same negative dynamics. And I can still tell ChatGPT's uniquie style and a human apart. And once you go into detail, you'll notice the quirks or the intelligence of your conversational partner. So yeah, some people use ChatGPT without disclosing it. You'll stumble across that when reading AI generated article summaries and so on. You're definitely not the first person with that idea.
I really don't think this place is about bot warfare. Usually our system works well. I've met one person who used ChatGPT as a form of experiment on us, and I talked a bit to them. Most people come here to talk or share the daily news or argue politics. With the regular Linux question in between. It's mostly genuine. Occasionally I have to remind someone to tick the correct boxes, mostly for nsfw, because the bot owners generally behave and set this correctly, on their own. And I mean for people who like bot content, we already have X, Reddit and Facebook... I think that would be a good place for this, since they already have a goot amount of synthetic content.
Hmmh, the 4090 is kind if the wrong choice for this, due to its memory bus width... For AI workloads and especially if you want to connect lots of memory, you kind of want the widest bus possible.
I think there is quite some overlap with Lemmy. You can do images and threads here. And they even show up in a tree-like structure. I'm not sure about ephermal, I don't think Lemmy is as advanced. But other Fedi software have ephemeral posts.
So, given your requirements, I'd say this is that place. Communities would be your topic boards, and posts would be the threads. And user accounts and moderation would be the improvements to make it less toxic.
Lemmy and Fediverse software have a box to tick in the profile settings. That shows an account is a bot. And other people can then choose to filter them out or read the stuff. Usually we try to cooperate. Open warfare between bots and counter-bots isn't really a thing. We do this for spam and ban-evasion, though.
Ich weiß nicht so recht was die Frage ist... Bezüglich der Jobs, für die man Leute bezahlt? So ungefähr alle: Bäcker, Supermarktverkäuferin, Altenpfleger, Politikerin... Mein Punkt war, dass man das für Soldaten auch so machen kann. Schreibt man eine Stellenanzeige, bezahlt denen 40-50k im Jahr und vielleicht bildet man die auch gut aus.
SATA is more reliable than USB. And SSDs are faster, but more expensive than a harddrive. In the end all of this is possible. You need to see how much space you need and what you can afford. Unless you need it to be quiet/silent or super fast, a regular harddrive might do.
Can we add whether something is Free and Open-Source Software to these infographics? It'd be massively useful to someone like me, and some of the projects seem to be FLOSS in contrast to your regular (paid) cloud service.
I use OsmAnd and Open Street Map data as well. That's crowd sourced and not under control of any company. Though, as other people pointed out, it doesn't do recommendations. It's map data, navigation and sometimes opening times, accessibility etc of a place.
Wobei sich um die Alten und Kranken kümmern, die Gesellschaft und das Land mit sozialen Dingen nach vorne bringen auch ein Beitrag ist. Also abseits der Bundeswehr gibt es auch allgemein noch viele Möglickeiten aktiv das Land zu schützen. Auch so sachen wie Entwicklungshilfe und Unterstützung anderer Länder hilft ja in erster Linie gar nicht erst in eine Kriegssituation zu kommen. Vielleicht fokussiert man sich mehr darauf, weil das ist auch üblicherweise effizienter und billiger als dann schlussendlich einen Krieg austragen zu müssen. Deshalb würde ich im Allgemeinen sagen, schickt die Leute eher dorthin als zum Grundwehrdienst an der Waffe. Das heißt jetzt allerdings auch nicht, dass es nur damit getan wäre...
Also ich wäre dafür, dass man Leute bezahlt für den Job. Das ist ja in anderen Branchen auch so üblich. Dabei gewinnt man auch noch mehr Vorteile. Man muss sie nicht zwingen, man kann sie länger und besser ausbilden... Ich denke das ist keine kognitive Dissonanz. Leute mit 18 einziehen und die 9 Monate hauptsächlich herumpimmeln und saufen lassen, so wie es früher üblich war, ist halt nur nicht die Optimallösung.