Too late. It's pretty much required nowadays in most regions (not something I agree with, people who don't want computers should be able to choose to entertain themselves in other ways). Computing is now the only established way to have pointless arguments with people around the globe. Also the only established way to attend school in lots of places. It's really difficult to have homemade computing without first "lightly disestablishing" the silicon companies holding trade secrets.
Did you train them without any data centers though?
Unless video games are bad, too?
Not really, but whenever choosing a game to play, I am biased towards some of the lesser resource intensive titles like TeeWorlds, Need for Speed: Underground and The Ultimate Shovelware Title: Homebrew Edition rather than, for example, Counter-Strike 2, Forza Horizon or The Ultimate Shovelware Title: Unity Edition.
He didn't have the same reaction at that specific point of time. That does not mean he was fine with it, but even if it was, maybe people change? Insane conclusion, I know. Stop with the black and white thinking.
Project is permissive. Company takes it and distributes binary. Nerds don't use it and look for the og, since the license forces the company to disclose that they just did that in a way or another. The end.
Project is copyleft. Company takes it and distributes source via GitHub file uploading a zip named after a date or some dumb shit. The original maintainer is now bullied into adapting the code. Sooner or later, the company now has influence over the project. Which is how we ended up with that Debian decision the top post is about, and with completely undocumented drivers right inside of the official Linux tree.
Black and white thinking. Just like fascists (including tankies).
LLM bad, data center not inherently good. Humans could get rid of data centers if they get their shit together and distribute instead of just decentralize, similar to how they can establish anarchism the same way. BUT if the ideal were to be assumed impossible (because of course you would), data centers are good at a certain quantity. Peeing is fine, everyone does that. If you pee 8 times a day, go see a doctor.
You can can inspect open source software. You can compile the source into binary. You cannot inspect open binary. Well, you can, but it's difficult. Can anyone ever inspect the fucking weights in a useful way? They are just bias numbers.
I feel like it's useful to have choices here. Like a "<-- you said this guy sucks" label separate from "hide" (placeholders), "ignore" (no placeholder when 0 replies) and "block" (no placeholder at all).
Netsurf hasn't seen any love lately. It's still being developed on git but no releases. Same with Dillo.
Dillo rejects scripting. Makes perfect sense to me. It also rejects automatic redirects, which also makes sense but the web has already been established around that as given so it can't pass the most passive anti-bot.
The state of scripting support on Netsurf is uhm, quite interesting. Scripts can write to the page parse time (like a snowflake generator) but can't do ANYTHING after the page loaded, EXCEPT text fields and text areas. And scripts can always read but not write checkbox state. There aren't any button events, either, but if you check the state of a text field repeatedly you can do interactivity. What is also interesting is that the new way of finding elements (getElementById) is functional, but the old way (document.formName.elementName) ISN'T. So that means that basically nothing existing functions (apart from again, the snowflake generators from geocities). But you can do the old style text animations if you specifically program it to use the new method despite being old style shit. Netsurf also passes "GoAway" and the most basic setting of Anubis (the passive redirects).
Both are good enough to browse "old Lemmy", although I haven't tried to login since my instance doesn't have it and the owner ghosted me (hi @milan@discuss.tchncs.de).
There's also a half public domain half permissive (BSD) text mode browser, https://chawan.net/ . Extraordinarily good concept! However, it avoids some of the bloat by forcing bloat onto your terminal emulator and having it render images, and my terminal emulator does not bother itself with such tasks and emulates a terminal, so image heavy sites like Lemmy fucking suck on it. No, it cannot just weight different characters from some public domain font (spoiler: if converted, all of them by international conventions are public domain) by brightness and by individual pixel similarity to an actual image and then just display those. That would make too much sense. I have high hopes for it, though. Also, I find the keybinds really weird.
I have not tested Chawan's scripting support much at all, all I can say is that it's certainly better than Netsurf. And that everything I've written in the worst fucking programming language ever for Netsurf (not released yet, very soon I hope) runs on Chawan. However, they are unusable since textboxes open your preferred editor. Makes perfect sense, a browser shouldn't be a text editor. But like, that breaks my gimmick that makes it work with Netsurf, damn. Checkboxes can be written to, for once, tho no access key support. The JS is still not enough to login to Lemmy (the default bloated frontend).
Fun fact, most Tetris clones have a 4D table of all pieces and their rotations. Or multiple user selectable. The actual game generates them (look up "tetris recomp site:github.com").
I read so fast that with long subtitles I have too much time to think about me being useless