an Experiment with GPT

[remod]

I wanted to check to what extent a properly instructed AI could create a reasonable I Ching reading. It's an experiment, I'm not advocating using AI to consult the I Ching, just my curiosity and I wanted to share the results with you.

 

I created two custom GPTs with chatGPT and passed them the Whilelm translation (a brand new, machine translated, version directly from German to avoid any copyright issue) and the Bradford Hatcher's translation.  I also gave them the ability to cast hexagrams using the three-coins method (probabilities) and compared the outcome.
I set up the question: "Should I move from this city to solve my issue problems?" and got 48.5.

Here is how they look:

• Using Bradfortd Hatcher's translation
• Using the ML translated version of Whilelem German translation

 

If you have a chatGPT plus subscription, you can use the two GPTs for yourself. If you put the answer before the question, it will use it for the reading, otherwise will toss the coins to generate a new response.

For example the prompt:

    "48.5 Should I move from this city to solve my issue problems?"

will use the indicate response, instead of generating a new one.

 

The two readings are are quite similar, which is good sign. Some told me that the answers look "mechanical", "cold" and "non human" which is, of course, true. However, I feel it was worthwile to try and see the outcome for myself.

 

If anyone is interested, I released everything on GitHub. You need a chatGPT pro subscription, unfortunately, but the python script that is at the core of it is pretty simple. And the prompt is self-explanatory (I believe). In doubt, just ask!

 

Any feedback and comments will be appreciated.