I love this. This reminds me of Data for Star Trek: The Next Generation. The most sophisticated android in the quadrant had only a couple of limitations compare

I love this. This reminds me of Data for Star Trek: The Next Generation. The most sophisticated android in the quadrant had only a couple of limitations compared to humans. One was his inability to use contractions, and the other was the inability to understand jokes. It seems like ChatGPT shares the latter limitation of the fictional Data. Second. I’ve seem this language model describe it self like the subject in Searle's Chinese room thought experiment. Only “passing” the turing test because it has been trained to do so. However, this shows that ChatGPT doesn’t actually pass the turing test. In fact asking a language model to explain jokes might be a good way to conduct the Turing test going forward.