Except that it already has been. They've already scraped it, and can refer back to either the archives, or just scrape Reddit like they do with other websites if they want to pull more information.

They didn't pay before, why would they bother paying now? Worst case is that they just exclude Reddit (like they did Twitter), and train from other sites instead. It's no great loss.