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.

  • 3 Posts
  • 837 Comments
Joined 4 months ago
cake
Cake day: June 25th, 2024

help-circle

  • This would actually be better than most of my local broadcast TV.

    Oh, well. It’s certainly better than the following program. I believe it’s followed by the news and then some documentaries about nature and places around the world. The other channels go downhill fast. Some “breakfast” TV show with some poor souls who have to get up early to do some meaningless happy talking, recipes and maybe invite someone who once met a celebrity… Followed by a re-run of yesterday’s local news and other random stuff that’s cheap or still in the VHS deck… Idk. I suppose it’s meant for some 95yo bedridden grandmas who can’t escape. I believe no sane person with working limbs would turn on any broadcast TV in the morning.


  • Sure. I know like 5 in my city. One ontop of the main fire station, i think another one filming the inner city, several at a lake (which is a local recreation area). Someone used to capture his garden with a few trees and mostly the sky and clouds. And we used to have a bunch pointing at the traffic on the Autobahn, so you could watch traffic and brace for the ordeal before driving off to work. But they took the live feed down when the Russia/Ukraine war started. And of course ski resorts have them. You can turn on the TV and watch those ski resort webcams at like 6am on TV. Usually accompanied by some super annoying hillbilly music. 😆





  • But that’s just not true. I think you’re confusing this with some source-available licenses or these silly amendmends to licenses that make it defacto proprietary. But this isn’t the case here. This statement in the CLA doesn’t take away any rights. It gives additional ones. And it’s in addition to the AGPL. All of the AGPL applies in addition to the CLA. Every single freedom, just as if the CLA weren’t there. You can use it, modify it, copy it, etc…

    The “around a decade” of course applies. And none of that has to do with the signing away copyright per CLA. You also don’t know if Linus Torvalds is around in 5 years and keeps maintaining the kernel to your liking. You also don’t know if any of the big open source projects of today get bought by some shady investors and the next updates won’t be free software anymore… These things happen. And it has little to do with a CLA (like this one). Happens to plain standard licenses without extras, too. And it does to ones with this kind of licensing. But this really isn’t the distinguishing factor.

    I think what you mean is modified licenses. Or similar additions that render something not open source anymore. I agree, you should avoid those projects at all costs. But that’s a different story and not what this project is doing.

    My point is a different one: While focusing on some small details of a hypothetical case that I think you got partly wrong… Have you checked for any big elephants in the room? Because I don’t see anyone talking about the database / search index and whether that’s available. The website is just a very small part and just the frontend to query the database. I’d say it’s almost pointless to discuss what we’re arguing about. I can code a search frontend website in a week, that’s not the point. And completely irrelevant if it’s open source… What about the data that powers the search engine? I think that’d be the correct question to ask. Not whether the frontend is 99% or 110% open source.



  • Very good idea. I mean there are frameworks for programmers to do exaclty that, like LangChain. But I also end up doing this manually. I use Kobold.cpp and most of the times I just switch it to Story mode and I get one lage notebook / text area. I’ll put in the questions, prompts, special tokens if it’s an instruct-tuned variant and start the bullet point list for it. Or click on generate after I’ve already typed in the chapter names or a table of contents. Or opened the code block with the proper markdown. So pretty much like what you lined out. It’s super useful to guide the LLM into the proper direction. Or steer it back on track with a small edit in its output, and a subsequent call to generate from there.




  • I’m not sure if you people are paying attention to the right thing. It’s fairly common to do this. And it doesn’t mean they can take away anything. Everything will still be AGPL and still available. Someone is then going to fork it and maintain it as it happened with lots of other projects. This just means they’re also able to also sell it under different conditions, including your patches and contributions.

    I think what you should pay attention to is, whether the search index is open or closed. That’s something with significant impact. Not if they’re able to monetize your small bugfix without paying you. I mean that’d be nice, too. But not a super big thing unless you contribute a substancial amount of code. I mean you get a whole open source search engine in return for signing away your copyright. And it doesn’t change anything for the people using the software. For them it’s still AGPL. And the maintainer could stop developing the software at any point, anyways. Could (and does) also happen to projects without a CLA.






  • And learning from the dataset is kinda the whole point of LLMs, right? I see some fundamental problems there. If you ask it where Alpacas are from, or which symptoms make some medical conditions, you want it to return what it memorized earlier. It kind of doesn’t help if it makes something else up to “preserve privacy”.

    Do they address that? I see lots of flowery words like

    Integrating privacy-preserving techniques often entails trade-offs, such as reduced accuracy or increased computational demands, […]

    But I mean that’s just silly.


  • Hmmh, yeah that further narrows down the list of potential candidates. I can’t give any good recommendation. To give some insight: Most people say their first time wasn’t good sex at all. It’s unfamiliar, you don’t know what to do, you’re stressed out and it doesn’t feel good etc. It’s certainly exciting and something to remember. But usually not enjoyable. And keep in mind losing virginity is a one-time thing. You can enjoy the look or habits or jokes of someone each day they’re around. The virginity thing is something you do one day and then it has lost it’s meaning in a way. Idk. Make of this what you will. I can sympathize with someone saying they don’t want to experience their first time with some random person but with someone special. I think that’s valid. The other way around is a bit more tricky. You can’t really expect that from someone. You can try. At the same time be aware of your high standards and expectations. That doesn’t mean you have to lower your standards. But you could squander a chance at meeting your potential partner if you’re not open to it. And these things happen if you’re focused on small details and that makes you unable to look at the whole picture of who someone is.

    And the last thing, we all can’t look into the future. Statistically, your first partner won’t end up being your spouse. It’s a nice romantic dream to marry your first love. But more something from a movie. So if you’re going for that, that could also turn out to be a fruitless endeavor. In any case, you’ll know in hindsight. But I really don’t know how much effort to put into making a fist relationship perfect. Maybe it’s a good idea to strive for it, but not be entirely crestfallen if it turns out differently.