Can ChatGPT Solve The Fermi Problem?
Probing AI’s Capacity for Quantitative Reasoning
Note this was conducted on ChatGPT 3.5, not GPT-4.
ChatGPT has taken over the world. With impressive feats like passing the bar exam and folding proteins, it seems like magic. However, ChatGPT is fundamentally a large language model (LLM) and there’s an ongoing debate in regard to it being…