Conversation

Notices

1. Embed this notice
   Rich Felker (dalias@hachyderm.io)'s status on Thursday, 16-Jan-2025 06:07:26 JST Rich Felker

   • tante
   • altruios phasma

   @woe2you @altruios @tante No, they're preparing unauthorized derivative works. And they're explicitly and intentionally disregarding opt-out. Ability to access something via the web does not imply right to incorporate it into other works, republish altered versions of it, etc.

   • Embed this notice
     altruios phasma (altruios@mastodon.social)'s status on Thursday, 16-Jan-2025 06:27:32 JST altruios phasma

     in reply to

     • tante

     @dalias @woe2you @tante

     The difference between a human reading a website and writing an article ‘inspired by’ what they’ve read And an LLM consuming and outputting content the same way is we recognize that an LLM is a tool and can do the same thing faster.

     Reading is training. Reading isn’t copying. Output is the issue: not input. It’s worrisome to see so many not grasp this.

     Looking/copying isnt stealing. It just isn’t. No one lost their website.

     Rich Felker repeated this.
   • Embed this notice
     Rich Felker (dalias@hachyderm.io)'s status on Thursday, 16-Jan-2025 06:27:32 JST Rich Felker

     in reply to

     • tante
     • altruios phasma

     @altruios @woe2you @tante You really have no understanding of the mechanics of the LLM fraud and are just repeating their false talking points about learning/training.

     LLMs are overfitted (that's the large part) models, and as such, they actually encode copies of large swaths of their training material. The best description is "lossy compression algorithm". I don't get a free pass pirating movies because the compression is lossy.

   • Embed this notice
     Rich Felker (dalias@hachyderm.io)'s status on Thursday, 16-Jan-2025 06:29:22 JST Rich Felker

     in reply to

     • tante
     • altruios phasma

     @altruios @woe2you @tante In any case, it's not up to the AI simps to decide if we should be furious at this industry for scraping & pirating our stuff. It's up to us. And questioning/challenging our response is an asshole move.

   • Embed this notice
     altruios phasma (altruios@mastodon.social)'s status on Thursday, 16-Jan-2025 06:43:51 JST altruios phasma

     in reply to

     • tante

     @dalias @woe2you @tante I would love to be wrong about something I am angry about or hate. I try to make that happen as often as possible. I don’t see challenging discourse as disrespectful or an asshole move. Everyone has an opinion. Sharing is how we learn.

     I really do think a lot of anger is misdirected at AI/LLMs when it would be more accurate to say that capitalism as a system is the thing at fault here.

     Ai is just the easy thing to point towards.

   • Embed this notice
     Rich Felker (dalias@hachyderm.io)'s status on Thursday, 16-Jan-2025 06:43:51 JST Rich Felker

     in reply to

     • tante
     • altruios phasma

     @altruios @woe2you @tante LLMs and capitalism are inseparable. Nobody would be making slop machines except for capitalist motives.

   • Embed this notice
     Rich Felker (dalias@hachyderm.io)'s status on Thursday, 16-Jan-2025 06:44:59 JST Rich Felker

     in reply to

     • tante
     • altruios phasma

     @altruios @woe2you @tante If you're having a hard time understanding how awful this stuff is, go read my pinned thread.

   • Embed this notice
     altruios phasma (altruios@mastodon.social)'s status on Thursday, 16-Jan-2025 06:51:08 JST altruios phasma

     in reply to

     • tante

     @dalias @woe2you @tante that is inaccurate.
     There are other motives other than profit to making AI. LLM’s are a reasonable attempt: they would have been made regardless of capitalism.

     Capitalism is the problem. And AI, while an accelerant to capitalism… would be an accelerant regardless of political systems deployed…

     If we had a leftist government: AI development would be optimized for citizens first as a public open source inspectable utility…

   • Embed this notice
     Rich Felker (dalias@hachyderm.io)'s status on Thursday, 16-Jan-2025 06:51:08 JST Rich Felker

     in reply to

     • tante
     • altruios phasma

     @altruios @woe2you @tante You really have no idea what you're talking about, and seem to be basing this on marketing copy not any understanding of the actual math.

   • Embed this notice
     Rich Felker (dalias@hachyderm.io)'s status on Thursday, 16-Jan-2025 07:06:20 JST Rich Felker

     in reply to

     • tante
     • altruios phasma

     @altruios @woe2you @tante Thinking that LLMs are useful, or that even if they were, that the astronomical environmental and economic costs of building them would be justified by anything other than VC "last mark holding the worthless stock" shenanigans.

   • Embed this notice
     altruios phasma (altruios@mastodon.social)'s status on Thursday, 16-Jan-2025 07:06:21 JST altruios phasma

     in reply to

     • tante

     @dalias @woe2you @tante

     Railing against capitalism* didn’t see that in the marketing brochure /s

     But really? What have I said that would indicate I’m only consuming marketing material?

   • Embed this notice
     Rich Felker (dalias@hachyderm.io)'s status on Thursday, 16-Jan-2025 19:03:31 JST Rich Felker

     in reply to

     • tante
     • Peter Kraus
     • altruios phasma

     @pkraus @altruios @woe2you @tante No, it cannot. When asked to cite sources, it instead falsifies citations most likely to be believable. The citations it gives have nothing to do with any process by which it generated the information-shaped slop it output.

   • Embed this notice
     Peter Kraus (pkraus@berlin.social)'s status on Thursday, 16-Jan-2025 19:03:33 JST Peter Kraus

     in reply to

     • tante
     • altruios phasma

     @altruios @dalias @woe2you @tante well if the author attributes their sources and cites properly, we call it academic publishing.

     Can an LLM do that, reliably? Can it be held accountable for academic misconduct?

   • Embed this notice
     Rich Felker (dalias@hachyderm.io)'s status on Thursday, 16-Jan-2025 19:04:13 JST Rich Felker

     in reply to

     • tante
     • Peter Kraus
     • altruios phasma

     @pkraus @altruios @woe2you @tante And no it cannot be held accountable. That's why it must never make a decision.

   • Embed this notice
     Jean-Christophe Helary (jchelary@sciences.social)'s status on Friday, 17-Jan-2025 13:03:28 JST Jean-Christophe Helary

     in reply to

     • tante
     • altruios phasma

     @altruios @dalias @woe2you @tante A human writing an article "inspired by" has actually done the reading, the understanding and the writing, and also not burned a forest in the process.

     The human writing has *added* knowledge to humanity as a whole thanks to the writer. AI, just like crypto, is just a costly and wasteful way to burn energy.

     If you don’t see the difference, feed *your kids* AI output all day everyday and check what comes out.