January 28, 2024
Over the last couple of weeks, I held the following five long conversations with GPT-4 to see how well I can extend and deepen my knowledge by interacting with the LLM. I began each conversation by asking it to suggest possible topics. I then chose topics for it to explain, and I asked follow-up questions as we went along. Some of the conversations were held over several days. At the end of each conversation, I asked it to summarize what we had discussed. My conclusion is that this LLM—and probably other LLMs that are now becoming available—can be very useful for acquiring knowledge about a wide range of subjects. Its explanations were clear and well-organized, and it responded appropriately to all of my questions. I did not spot any hallucinations, though some might have occurred if I had asked for more specific information.