this post was submitted on 03 Aug 2026
467 points (91.2% liked)
Technology
87575 readers
3593 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related news or articles.
- Be excellent to each other!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
- Check for duplicates before posting, duplicates may be removed
- Accounts 7 days and younger will have their posts automatically removed.
Approved Bots
founded 3 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
I mostly agree, but even if LLMs were 100% correct all the time (they're not), there are still some pretty serious ethical issues surrounding it. It is up to each person to decide if they find that acceptable or not.
I disagree with the moral panic about AI stealing things. They're not Xerox machines. It does exactly what humans do except faster. Before AI, this YouTuber was reading a bunch of stuff, "stealing" things from it, and putting it back together in a different way, just like AI, except he actually understands what he's reading.
I am concerned though about the resources it uses. But I don't see that as an ethical issue. That's more of a logistical issue. If they can make AI that doesn't drink up Lake Michigan and doesn't make the electrical grid scream, then more power to it.
My biggest concern whether people are for it or against it is that they don't think it's Skynet, or Data from Star Trek. It's just a big dumb machine, a gigantic MENACE matchbox computer.
My complaint is at least before the individual stealing the idea and recombining it was the one getting paid.
Now that's just a SaaS that benefits a handful of overvalued superscalers. They get to sit on their horde of data and charge for access.
This makes it sound like it can do anything humans can do, but you're actually wrong in a lot of different ways. Generative AI is a reconstruction machine. It takes the things it "knows" (because they don't know anything, they're just lots and lots of coefficients in giant matrices), and attempt to reconstruct from an incomplete data point.
They are actually a lot closer to Xerox machines than they are to humans. If you train gen AI on one thing, it will only know that one thing, and it will only recreate that one thing. It could never do more than that. Humans just do not fundamentally work that way. Like, if you want to say most stories are recombinations of other stories, fine, but that is wrong. There is an origin that was created by a human. Not everything is stolen.
There are many, many issues with gen AI, from the resources it uses, to its use in surveillance, to the fact that it is still wrong on fundamental ideas. It's just kinda shit technology and the people that think it is awesome are usually the dumbest fucking morons I have ever met, and that is terrifying. Man, just learning the math and science behind it and how it works should be the bar, but none of them know how it fucking works or why it cannot improve without using more and more resources.
They literally stole the training data. Everything after that is fruit of the poisonous tree.