Phil Stevens :tinoflag:RE: https://mastodon.nz/@gristc/116009021016084895
"Essentially all LLMs are built on scraping the Internet, badly, often completely trashing servers as they download the entire site without using any of the polite techniques developed by archivers and search engines over the past 30 years. They're all built on copyright infringement. Which is a felony, as anyone who has crossed Disney well knows."
All GNU social JP content and data are available under the