Earlier quoted context omitted.
Perplexity are also flagrant copyright violators
Because they summarize and cite sources or because their models were trained on copyrighted materials? Summarization and training should be tranformative, and the user questions add that element of novel purpose to the original materials that should make the output non derivative. But most of their responses are one time use, nobody is ever returning to them.
No, because they plagiarize
https://www.wired.com/story/perplexity-plagiarized-our-story...