I really enjoyed this perspective! Lately, I’ve been asking "what would the model cite?" before "what would rank?" because the quality of training data changes everything. It’s become a bit of a hobby for me:

https://wiki-nest.win/index.php/Why_Does_My_JavaScript_Site_Look_Fine_to_Users_But_Not_to_AI_Crawlers%3F

I really enjoyed this perspective! Lately, I’ve been asking "what would the model cite?" before "what would rank?" because the quality of training data changes everything