Ask HN: How to Build an LLM of the Buddha?
4 comments
You can easily create a custom version of ChatGPT which focuses on whatever teachings you want. I'm not familiar enough with the teachings of Buddha to know what to tell it to make it for you.
eg there's https://chat.openai.com/g/g-WxckXARTP-astrology-birth-chart-... for astrology
eg there's https://chat.openai.com/g/g-WxckXARTP-astrology-birth-chart-... for astrology
Fascinating question.
I think the problem is that the 15-17k pages of the Pali Canon are already hopelessly sectarian. You won't get ChatBuddha, you'll get ChatTheravada.
As an aside, this answers the oft-posed question: how could Ananda possibly have recited the entire canon after Master Gotama's death? Answer: there was a lot less of it then.
More useful is to discard the sectarian suttas, read the rest and do what they say. In my experience it is highly effective. When you realize you must be "a lamp unto yourself" then you're almost there.
I think the problem is that the 15-17k pages of the Pali Canon are already hopelessly sectarian. You won't get ChatBuddha, you'll get ChatTheravada.
As an aside, this answers the oft-posed question: how could Ananda possibly have recited the entire canon after Master Gotama's death? Answer: there was a lot less of it then.
More useful is to discard the sectarian suttas, read the rest and do what they say. In my experience it is highly effective. When you realize you must be "a lamp unto yourself" then you're almost there.
The whole point of Buddhism is to become a Buddha and not waste any effort in the idolization of an imagined external Buddha, be it in image or scripture. This is a matter beyond words or mental believe. So what is the point of an LLM'ification of Buddha/Buddhism?
you wouldn't happen to be related to Francisco J. Varela, would you?
Some ideas that pull me to thinking about this:
- LLMs build a world model and are not 'just' parrots
- The Buddha spoke and people wrote it down, maybe the 'original' world model can be extracted from the writing?
- The Pali canon has 15-17k pages, sounds like a lot but it's a small dataset compared to how much data is used to train models these days
- There's orders of magnitude more commentary that could be used, but does it detract/dilute the 'original' world model?
- Say we dump everything written about buddhism, there's probably redundancy but maybe enough material to get 'enough' data to train