I can think of half a dozen trivial ways to filter this, most of which are probably already being done on training sets. This isn't going to come anywhere close to starving the rat. Nothing will, they'll just build "better rats."
That said, I'm glad it won't. Humanities future will involve AI, and the luddites won't be able to stop or slow it. They'll just make it more expensive at worst.
Today's AI's are the worst they will ever be, and nothing anyone does today can change that.
> Today's AI's are the worst they will ever be, and nothing anyone does today can change that.
That's not at all as certain as you put it. "Model collapse" isn't just a theoretical possibility, it's happening right now, noticeably.
I wouldn't be surprised if we look back at this time as when models were the best they ever would be. Similar to how few people today would argue that the web is definitely past its peak.
I can think of half a dozen trivial ways to filter this, most of which are probably already being done on training sets. This isn't going to come anywhere close to starving the rat. Nothing will, they'll just build "better rats."
That said, I'm glad it won't. Humanities future will involve AI, and the luddites won't be able to stop or slow it. They'll just make it more expensive at worst.
Today's AI's are the worst they will ever be, and nothing anyone does today can change that.
> Today's AI's are the worst they will ever be, and nothing anyone does today can change that.
That's not at all as certain as you put it. "Model collapse" isn't just a theoretical possibility, it's happening right now, noticeably.
I wouldn't be surprised if we look back at this time as when models were the best they ever would be. Similar to how few people today would argue that the web is definitely past its peak.