Surprise! chatGPT o1-preview fails at simple logic!
ChatGPT o1-preview is touted as good at PhD level logic. However, while recording a short video, I found it made a mistake in a simple logic with a balance and weights. This shows that while a simple human logic that can be applied by a high-school student becomes more difficult when a mathematical solution has to be worked out. Though it got the 40KG weight problem correct, it got the details wrong as it gave wrong logic for measuring using the balance.
