Can ChatGPT reason using counterfactuals? Can it interpret nuanced uses of novel metaphors to explore the complex topic of language understanding? Is it morally vacuous, or merely cautiously optimistic? Consider the data and judge for yourself.
BotChat – Can I ask you a riddle?
"Can I ask you a riddle?" I asked ChatGPT on our second day. "Of course! I love riddles," it replied. "Go ahead and ask me one."