Probably a dumb question, but wouldn't Reddit be very easy to scrape instead of using an API at these prices? I assume the various LLMs' datasets are based on web scraping similar to Google for content that doesn't have a public API.

Also, it seems to me that Reddit could gain some goodwill here by supporting the flagship competing client apps with a lower rate while still charging a higher price for those trying to leverage its data for ML, similar to how Google provides much of the funding for Mozilla despite being a competing browser.