I'd rather they were destroyed, but practically speaking that's impossible, and this sounds like the next best idea to me.
this post was submitted on 22 Dec 2024
1470 points (97.4% liked)
Technology
60062 readers
3387 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related content.
- Be excellent to each another!
- 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, to ask if your bot can be added please contact us.
- Check for duplicates before posting, duplicates may be removed
Approved Bots
founded 2 years ago
MODERATORS
I don't think it should be a "punishment." It should be done on principal.
Not sure making their LLMs public domain would really hurt their principal, their secret sauce is in the code around the model.
And yes, I do recognize that you meant "principle".
I want to have a personal llm that learns all my interests from my files and websites visited. I just want to ask it stuff that I don't have to remember.
I'm working on something along these lines for myself, I think of it like using AI as a filter to create a bubble of good Internet around me
I think that'd be ok, even with this proposal, as long as you don't sell that LLM for public use. It's fine it I draw a picture of Mickey Mouse in my notebook, but if I try to sell that picture I could get in legal trouble.
Isn't that similar to what recall is?
Yes, except without Microsoft spying on you
Exactly. I don't want a service, I don't want to pay for a service, I don't want to send my files for free to get stuck for later ransom like Google did with email. I just want to purchase a product called a computer and load up a program in it that runs locally and gives me access to my data.
They don't mean your data, silly. They don't give a fuck about that.
They mean other huge corporations data.
intellectual property doesn't really exist in most of the world. they don't give a shit about it in india, bangladesh, vietnam, china, the philippines, malaysia, singapore...
it's arbitrary law that is designed to protect corporations and it's generally unenforceable.
it’s arbitrary law that is designed to protect corporations and it’s generally unenforceable.
It's arbitrary, but it was designed to protect individuals, but it has been morphed to protect corporations. If we reset the law back to the original copyright act of 1790 w/ a 14-year duration, it would go a long way toward removing power from corporations. I think we should take it a step further and perhaps make it 10 years, with an optional extension for another 10 years if you can show need (i.e. you're an indie dev and your game is finally making a splash after 8 years).
they don’t give a shit about it in india, bangladesh, vietnam, china, the philippines, malaysia, singapore…
Unless it's their intellectual property, whereupon it's suddenly a whole different story. I'm sure you knew that.
Examples?
China: https://futureworld.org/mindbullets/china-sues-us-for-ip-theft/
Philippines: https://www.rappler.com/business/industries/110820-etude-house-files-charges-lazada/
Decided to stop looking after that since that's three examples.
So true. IP only helps the corps and slows tech development. Contracts, ndas, and trade secrets are all you really need to keep your ideas safe. If you want your country to develop fast, get rid of any IP laws.
load more comments
(3 replies)
I used whisper to create subs of a video and in a section with instrumental relaxing music it filled on repeat with
La scuola del Dr. Paret è una tecnologia di ipnosi non verbale che si utilizza per risultati di un'ipnosi non verbale
Clearly stolen from this Dr paret YouTube channels where he's selling hypnosis lessons in Italian. Probably in one or multiple videos he had subs stating this over the same relaxing instrumental music that I used and the model assumed the sound corresponded to that text
Wouldnt that give people who is it for bad things easier access? It should be made illegal to create if they dont legally have access to that data
The "illegally trained LLMs" they're taking about are trained on copyrighted data that they didn't have permission to use, this isn't about LLMs that have been trained to do illegal things. OpenAI (chatgpt) is being sued because there is a lot of evidence that they used copyrighted content for training, like NY Times articles. OpenAI is so profitable that they'll probably see these lawsuits as a business expense and keep doing it. Most people won't sue anyway...
i know that by illegally trained LLMs they are talking about training on copyrighted data(by legally have access to, i meant that they are legally allowed to train AI on it).
Its ridiculous that companies can just ignore laws
Oh, I'm not sure what you meant in your first comment then?
view more: next ›