As part of our explorations of AI we have been working on developing query systems specifically oriented toward Mahamudra and epistemological enquiry. Our starting point has been a RAG (Retrieval-Augmented Generation) store built from Mahamudra study and practice sessions, which directs AI responses toward the view and language of the tradition.
This work has evolved considerably. The current implementation uses semantic chunking and dual retrieval — finding relevant passages both by the content of your question and by its contemplative register. The two results sometimes agree, sometimes diverge, and both are worth attending to.
The query is available here:
These implementations are optimised for open-ended enquiry. They will not always answer your question directly — more likely they will surface something relevant from the tradition that opens the question further. Bring your own discernment.