This is an experimental study that uses dialogue to prompt GPT4 (ChatGPT) to write a program for calculating 24 points.
GPT4 generates code based on user commands, which the user runs and provides running results or error messages to GPT4. Then GPT4 modifies the code and the process repeats.
The above process is recorded in history.md.
The final result code is final_24.py