We had a large language model as a guest on our movie review podcast

This was done with gpt 3.5 turbo, there's some preamble, then the model starts contributing at 4:37.