• 0 Posts
  • 8 Comments
Joined 1 month ago
cake
Cake day: May 20th, 2024

help-circle

  • It’s also difficult for developers to publish to Linux because of the wide variety of different Linux systems.

    I disagree there. The issue is that in Windows people bring over their own version of libraries they compiled on (the millions of .dll files) and you can even look in your Uninstall Apps settings where there’s a bunch of MS specific runtime bundles to see that’s even an issue in the MS ecosystem.

    In Linux, developers have relied on the library versions just being there. It is, I’d argue, the most compelling reason package managers basically had to come into existence. On the flip-side this can cause issues where there is some version on the system by the package manager that replaces another version. And something not a part of that package management system isn’t a part of those dependency checks and if they don’t put the libraries with the binaries…well it is just luck if you have them all or if other versions can support those library calls in the same way still.

    In Linux that is all those .so’s in /var/lib and stuff.

    You don’t really see many proprietary things using package managers and those that do are packaged by someone else and are in some sort of repo that isn’t part of the vanilla install because of legal caution.

    Companies that made their money on porting games to Linux prior to Proton basically causing them to shutter Linux porting would put their .so’s in with the game bundle themselves, just like you see happening in Windows when .dll’s are inside the actual program’s folders.

    However, the more that this sort of dependency management has become abstracted by development suites that take care of this for the developers, the less they understand about it.

    Flatpaks actually take care of this and it is one reason they are so popular. They figure out (well that’s a simplification) those library dependencies, sandbox the apps with those dependencies so the library paths don’t interfere with other flatpaks or the base system itself. People complain about this as a con because “the download is BIGGER” even though flatpak doesn’t install the same runtimes over and over again, so once they are there, the download may still be bigger but the installed storage isn’t.

    Anyway, yes Linus Torvalds complained about the “Linux fragmentation” issue but it was about DE’s not the state of the development ecosystem itself as I recall, though the rant is very old, so maybe I don’t remember all of it.

    Wider application support would be a start.

    Sure, but that’s not a Linux problem, that’s a developer problem. Linux supports application development just fine. It is a kernel and the surrounding ecosystem is the operating system after all. It is developers that don’t support it. That isn’t really something Linux in and of itself can effectively solve. Users have to increase and developers supporting applications for Linux will also increase. The classic Linux Chicken and the Egg problem but it is capitalism and that’s just going to be how it has to work.





  • It really depends on what model you want to run and how much training is bundled with it. You can pretty much run any model if you have enough disk space but of course GPU + VRAM is preferred for a ChatGPT like fast response. Otherwise, running on an older CPU and RAM is going to be noticeably slower, especially with complex models with a lot of training data to trawl through.

    There are some pretty lite models out there but the responses will be more barebones and probably seem ‘less informed’.

    Give GPT4All a try for your first time. It makes install, configuration and usage point-and-click while being fairly straight forward. For the presented/featured models, it presents a small summary and VRAM recommended, though there are many, many other models available from inside the UI.


  • Not that I’m defending it but the data and the model itself on Recall stays all local and encrypted, according to Microsoft. It also says it won’t use it for ad targeting or will sell the data. Of course, the caveat is that is what they are saying right now and may not be saying in the future. We’ve obviously seen strategies where gradually things move down the spectrum as it continuously normalizes.

    With MS we’ve seen the “Start” menu advertise Candy Crush forever and then “recommended apps” and it isn’t a far step to show “sponsored recommended apps” and then just “sponsored content” as things continue to become more normal for everyone, especially if its for the “Home” version or whatever. People will just argue to pay whatever for a Pro license.

    Going to full blown ads now though? It’ll piss the consumer off. Do it gradually over a decade? There will be some rumblings, sure, but it probably won’t matter. By then they might be able to give you a “free” cloud VDI (with lots ads from the OS) with less ads and CPU/GPU power based on subscription tiers and you just need to buy a cheap $30 thin client and everyone will just be OK with that.


  • For those that don’t know, they are going to release something called FreeLlama which might be FOSS (no public info as to what the license actually will be).

    Winamp says that they still want to control ‘what features’ go into winamp and it’ll remain proprietary. I assume they really just want people to contribute interesting things to FreeLlama and then put the contribution into Winamp.

    The license probably won’t be FOSS because they probably aren’t going to want anyone contributing to own copyright to the code that they are committing.

    It is odd because FOSS contributors aren’t really known for being OK with this sort of thing in the past, so I doubt they’re going to get much out of it. Maybe it’s a Hail Mary and they’ll end up blaming people for not freely giving up their devtime and creativity to a company that wants to make money on it.