I believe that is the case. Out of curiosity, I had this model try to solve a very simple Sudoku puzzle in ChatGPT, and it failed spectacularly. It goes on and

I believe that is the case. Out of curiosity, I had this model try to solve a very simple Sudoku puzzle in ChatGPT, and it failed spectacularly. It goes on and on making reasoning mistakes, and always ends up claiming that the puzzle is unsolvable and apologizing. I didn’t expect it to solve the puzzle, but the whole reasoning process seems fraught with errors.