I’m wondering the same, language models have been on the training for years, they already have all the reddit data they want, why would anyone spend a lot of money now for something they already have?
Unless there’s something new we’re not yet aware of, I don’t see how “scraping AI” is a valid reason for what spez is doing.
This is what I don’t understand. Data from before ChatGPT is useful but nothing after it is. Now you can’t tell for sure if anything on reddit is made by AI or not, which is critical for training AIs. It’s like the “bomb pulse” after the first atomic bomb tests which resulted in a release of carbon-14 in the air.
Spez is just flailing trying to make as much money off reddit as possible and trying to find any justification to do so, no matter how illogical it is.
What money is there to make off of language model AIs? They’ve been using reddit comments/interactions for a long time now. They’re too late.
I’m wondering the same, language models have been on the training for years, they already have all the reddit data they want, why would anyone spend a lot of money now for something they already have?
Unless there’s something new we’re not yet aware of, I don’t see how “scraping AI” is a valid reason for what spez is doing.
This is what I don’t understand. Data from before ChatGPT is useful but nothing after it is. Now you can’t tell for sure if anything on reddit is made by AI or not, which is critical for training AIs. It’s like the “bomb pulse” after the first atomic bomb tests which resulted in a release of carbon-14 in the air.
Spez is just flailing trying to make as much money off reddit as possible and trying to find any justification to do so, no matter how illogical it is.
That’s why they made it so expensive, to make money of new data