River

@riverpunk@defcon.social
65 Followers
110 Following
497 Posts

Computer Science student. I run a blog where I talk about technology, political science, and anything interesting.

https://riverseeber.net/blog/

US Politics 🌎
Programming 💾
FOSS 🔓
Leftist ☮

Pro-community and would love to see a world that doesn't normalize self-isolation so much 🫂

Feel free to interact as much as you want. "We are on a social platform after all!" -- @stux

DMs are open for anything as long as you're a human not a robot 👋

Websitehttps://riverseeber.net/
PronounsAny
Music Scrobblerhttps://www.last.fm/user/riverpunk
BogosBinted
Signal (DMs open)https://signal.me/#eu/AptqX0EjfLFSQmM3zw6vBYFUhw9ESHWm3c1OW3CQyFzAbyQJNU9Pv44j_snofvL3
BookWyrm accounthttps://books.theunseen.city/user/riverpunk
River1h ago
Nick!13h ago
Are you a gamedev with imposter syndrome? Check out the patch notes for PEAK!
River2h ago
Felicitas Pojtinger3h ago
Incredibly gorgeous day today in #Vancouver
River1d ago
lenazun1d ago
social networks are forcing me to quote Zizek re: "Don't fall in love with your suffering. Never presume that your suffering is in itself a proof of your authenticity. Renunciation of pleasure can easily turn into pleasure of renunciation itself"
River1d ago

I bought a Raspberry Pi and a tiny little display screen for it like a year ago. Never was able to get the screen to work. It used a ribbon cable DSI thing. I found a thing on a forum somewhere, and I had resolved that the screen must've just been incompatible with the Raspberry Pi platform -- something to do with bad drivers, or else literally just not compatible at all.

Well I was trying again this morning, and I realized that the issue was I had just literally been plugging in it wrong. I was plugging in in to the wrong side of the ribbon clasp, so either the pins weren't making contact, or else not contacting correctly.

My bad, I guess lol. Time to resume my Cyberdeck project  

River1d ago
Ben1d ago
It really bothers me when bigger accounts lift content from other accounts and don't credit. Like who's that for? Why not just boost the original to all the people who follow you instead of cutting off the legs of someone in your own community. Such a dick move. Fedi is about all of us and about lifting one another up. This ain't fuckn Bluesky or Twitter. If you're just after clout, go chase it somewhere else.
River2d ago
Show thread🅰🅻🅸🅲🅴 (🌈🦄)3d ago

@selzero I love how advertising companies and social platforms (which are largely one in the same) are like "but it's anonymous, aggregated data".

That doesn't matter when the directives are things like "target emotionally vulnerable teen girls with ads that make them feel worthless".

River2d ago
Neil Brown2d ago

I wish that more FOSS / ethically licensed projects would be more forthright about asking for donations of help *outside tech stuff*.

There are plenty of people who value FOSS who have skills to contribute, be it legal, accounting, comms, marketing, graphic design, project management, and loads of other things, and plenty of projects which could benefit from some volunteering time in those areas.

Finding a match can be difficult, so seeing people asking for help is excellent.

#FOSS #volunteering #ProBono

River2d ago
Em 2d ago

"Your prompts remain private while using our AI" isn't the pro-privacy statement most AI companies seem to believe it to be.

What about the data of the subjects you used to train your AI model? Did you ask consent from all of them? What's that? Your product isn't possible if you ask for consent first?

Well, it's definitely not private then.

#Privacy #NoAI #Consent

River2d ago
Enara ❔2d ago

Duckduckgo made a poll:

https://voteyesornoai.com

I'd have done it before implementing AI on the search engine, but better late than never.

Are you YES AI or NO AI? Vote now.

Big Tech doesn't care if you want AI or not. They should.

River2d ago

Exactly. The solution to "AI" isn't ever going to be stronger copyright. The solution is to institute better privacy laws, and to actually enforce antitrust law that would prevent monopolies being big enough to do this kind of crap in the first place.

RE: https://social.coop/@copiesofcopies/115905590434593333

Aaron Williamson (@copiesofcopies@social.coop)

Training ML models on copyright-protected works is probably fair use. LLM creators have banked heavily on this, and so far have won on it. I understand why many people don’t like that result, but it’s the same basic doctrine that gave us great things like useful search engines (remember those?). To get a different result for LLMs, we’d probably need to fundamentally undermine the fair use doctrine, and I’m not remotely convinced that’s a baby we want to throw out.

social.coop