hi guys, if @khoomeik solves Sanskrit OCR, this will put the Vedic literature in the training data. the Vedic literature is INSANE. this has big ramifications for "predictive language models" i was thinking about this, and how tiny cultures are being obliterated en masse.