As part of our explorations of AI for we have been working on developing AI specifically for mahamudra and epistomological enguiries. Our starting point has to use a combination of mahamudra study material and conversations with AI , chatgpt, to make a RAG ( Retrieval-Augmented Generation) store to direct the responses to questions posed to initially two AI programs.

The first AI program ‘openai ollama mistral:7b Q4_K_M’ LLM (look it up) runs locally in a home based server. You can send questions to it here:

https://constraints.org.uk/query-mistral-llm

The Second AI program ‘gpt-4o-mini’ is available from openai as as remote accessible applications prorogra, a so called API. You can send it questions here.

These implimentations of AI are optimised for sort of socratic responses. They emphasis the context of answers and will probably not even answer the question explicitly , but be more likely to give a meaningful or thought provoking lets say response.

https://constraints.org.uk/mahamudra-gpt35-query

You could try comparing the AI’s performance.