Retrieval Augmented Generation (RAG) with 100k PDFs?! Too slow!

Jeez, guess I undershot

Not @TeoR95 's fault though, this stuffs super confusing. Reading the documentation, there’s nothing to really suggest that it wouldn’t work.