If the issue is companies scraping Twitter to train LLMs wouldn’t the easiest solution be to create an official Twitter corpus and sell access?