Reddit is used to train AI… Twitch is used to train AI… Facebook is used to train AI… It just occured to me the easiest and best way to stick it to these people… get off the computer and have real life interactions. These people are busy building AI that can take over our digital lives… so take that back and make it a life they can’t take over…
I know easier said then done, it also depends on others, its expensive out there… I just woke up with a little bigger smile today realizing I could do something thats as easy as leaving reddit and could stick it to them… even if it just made a fraction of a difference.
See you tomorrow lol
The Fediverse is also used to train AI, it just doesn’t profit from it like the others and the data is gathered involuntarily
I was reading a thread about another one of these fedi scrapers just this morning: https://olifant.social/@oli/statuses/01KZXMMFKAHN4ZDZN9B8K3QM0T
Just remember kids, everything you say on fedi is extremely public, even your dms.
Ironically i cannot see the message behind that link.
Not Visible
There’s something at this URL but it’s not visible to unauthenticated web visitors. This could be for a variety of reasons:
It's a followers-only or direct post. It's a public or unlisted post but the author doesn't show these on the web. It's a reply to a post that is itself not visible.To try viewing it anyway, search the following URL in your fedi client of choice, and see if your fedi server will show it to you:
https://olifant.social/@oli/statuses/01KZXMMFKAHN4ZDZN9B8K3QM0T
URL
If you believe this was an error, you can contact the instance admin. Provide them with the following request ID: 000037yeaw14600frvag.
The DMs are not really public. They are only visible to admins of either instance.
Haahah, modern Luddism, same as the old Luddism.
LLM’s are in, in full swing. You’re not putting the genie back in the bottle.
Sad that you think this has any meaning, like people who think the thorn character somehow befuddles LLM.
You’re not putting the genie back in the bottle.
What do people mean when they say this? You’ve given up on regulation and accountability? I haven’t.
We need to fix some of our glaring problems before they kill us. By and large, “AI” is a harmful, wasteful scam. LLMs are not thinking machines.
LLMs - current AI isn’t going anywhere.
Home run models are going to stay even when big companies fold over the resource demand.Sure, but generally the problems with the slop machines include people thinking they’re alive, people not learning how LLMs actually work, giving your data away, and centralizing more power to the Epstein class.
If you’re hosting your own slop bot, you generally know, at least at a surface level, that they’re not smart, and they don’t think, and that they aren’t alive. You know how system prompts work, you know how different models achieve different things. You also don’t give your data away to the Epstein class for profit.
This isn’t to say it would solve all problems or create some problems of their own. But I’d much rather have everyone hosting their own slop bots than giving money to the Epstein class.
My main problem is the ethics of their training data.
And the insane wasted amount of resources. And Imho the resources are not that insane if you only look at a single “small business/family” needs. If GPU and RAM prices would be normal - I would have my own that can run a 25b model - and I would be happy with that.Yeah, stolen data is one of the problems it doesn’t solve.
But I wasn’t saying it takes an insane amount of resources to run locally. I’ve done it on a tiny B50. I’m saying that most people who are able to know what to install and configure to run them locally are generally more likely to be educated to know those things about LLMs.
Regulation and accountability are one thing. Having a fucking pipe dream and saying some version of “if we all band together, they’ll fail because no one will be using it” is another.
You are not going to stop people from using AI. Thinking there is literally any scenario where people will just stop using AI, or stop creating content that AI can be trained on, means you have an utter misunderstanding of basic human nature. It’s like a child thinking they can stop all wars if they just explained to everyone that killing is bad.
That’s what this post is. Use sites like reddit. Don’t use sites like reddit. It doesn’t matter. Maybe it makes you personally feel better, but don’t delude yourself into thinking it makes a difference. You would have more luck trying to fight the tide by standing on the shore with a sign that says “Tide, please don’t come in today”.
I see posts like this as pathetic desperation. Someone upset that technology is growing in a direction they don’t like, so they are performing worthless actions and deluding themselves into thinking it means something.
The post repeatedly specifies that this is a way to stick it to AI companies, not destroy them. You “stick it” to an entity when you do something that reminds yourself that they don’t have full control over you. They might not even know or care that it’s happening, but you do, and that’s enough.
You can stick it to your parents by sneaking out and drinking, you can stick it to your boss by slacking off whenever they’re not watching, and you can stick it to AI companies by living a fulfilling life outside of the digital space.
same as the old Luddism
You mean actually having a point about how centralised use of new technology empowers the rich to steal from the poor and should be more strictly regulated? Because that’s what it was.