PM_ME_VINTAGE_30S [he/him]

Anarchist, autistic, engineer, and Certified Professional Life-Regretter. I mosty comment bricks of text with footnotes, so don’t be alarmed if you get one.

You posted something really worrying, are you okay?

No, but I’m not at risk of self-harm. I’m just waiting on the good times now.

Alt account of PM_ME_VINTAGE_30S@lemmy.sdf.org. Also if you’re reading this, it means that you can totally get around the limitations for display names and bio length by editing the JSON of your exported profile directly. Lol.

  • 1 Post
  • 19 Comments
Joined 1 year ago
cake
Cake day: July 9th, 2023

help-circle
  • It can use ChatGPT I believe, or you could use a local GPT or several other LLM architectures.

    GPTs are trained by “trying to fill in the next word”, or more simply could be described as a “spicy autocomplete”, whereas BERTs try to “fill in the blanks”. So it might be worth looking into other LLM architectures if you’re not in the market for an autocomplete.

    Personally, I’m going to look into this. Also it would furnish a good excuse to learn about Docker and how SearXNG works.


  • LLMs are not necessarily evil. This project seems to be free and open source, and it allows you to run everything locally. Obviously this doesn’t solve everything (e.g., the environmental impact of training, systemic bias learned from datasets, usually the weights themselves are derived from questionably collected datasets), but it seems like it’s worth keeping an eye on.

    Google using ai, everyone hates it

    Because Google has a long history of doing the worst shit imaginable with technology immediately. Google (and other corporations) must be viewed with extra suspicion compared to any other group or individual because they are known to be the worst and most likely people to abuse technology.

    Literally if Google does literally anything, it sucks by default and it’s going to take a lot more proof to convince me otherwise for a given Google product. Same goes for Meta, Apple, and any other corporations.

















  • To quote the linked build doc:

    The recommended way to build ungoogled-chromium is by consulting the repository for your supported platform (links here).

    So if I have any build errors, I’ll probably just be told to go install the package for my repo. If not, I’ll probably just waste the dev’s time with some idiosyncratic issue with my own machine or toolchain.

    Now I’m on Debian stable, so there have been a few programs I’ve compiled because the one in the Debian repo was too old or unsuitable for my purposes. But in general, I trust the Debian packagers not to do nefarious stuff with the files in their repo. I also trust the open-source community to collectively and loudly shit their pants if this trust even looks to be violated.

    Compiling from source requires non-trivial computer science knowledge, namely familiarity with the overall steps of how compilers translate human-readable code to executable binaries. Thankfully, most projects I’ve compiled have decent instructions where you just read the commands and drop them in, but some programs aren’t well-documented. Some compiles are harder than others. For example, Ardour requires an absolute shit-ton of dependencies. It’s a rather complex compilation. So far I’ve run into no dependency glitches, but considering it is a music software that needs near-realtime performance to be useful, you can probably understand why I’m paranoid about stability. How do I know that, a few years down the line, some of the dependencies will decide that I compiled it wrong and my recording session gets borked? For this reason, Ardour actually uses this as a selling point to buy copies of their software: you can compile it, but it absolutely sucks to do, and you can’t be sure that you did it correctly until it’s too late.

    Most people don’t actually understand how computers work. Sure we know how to use our desktop environment, but really we’re engaging with the abstraction that tool provides. Going into the console breaks that abstraction.

    So let’s say that I put you in front of a computer with your favorite desktop environment (you would have to pick one, or a window manager), but I don’t tell you what the actual operating system is. It could be anything, Linux, BSD, ReactOS (or Windows if we throw out the desktop somehow), TempleOS, anything. All the main apps like a browser, file explorer, settings GUI, office applications, have already been installed, as well as the applications you use. Do you think you could use this mystery computer? Probably, because you know how to use computers that follow the desktop metaphor.

    Most people who use computers are familiar with the desktop environment [1] and will therefore have a much easier time transitioning between systems who each use the desktop metaphor. Compiling programs from source exposes the user to details that unmask the complexity of their systems. For most users, this causes unnecessary anxiety that does nothing to help convince them that the operating system is accessible for general use.

    Now while I think there is a great deal of value in learning to compile from source and to do so in some cases… generally, I just can’t be arsed to do so. I’m fine delegating a modicum of trust to others because then I get to focus on the stuff I’m interested in. There’s a .deb file where I can install the program and move on with my life.

    [1] Those who aren’t, like young children, are likely familiar with at least smartphone GUI idioms. They will probably be more comfortable on a PC whose desktop resembles that of a smartphone.