Large language models like ChatGPT are getting embedded into search, and monetization is becoming a bigger priority. If things shift to an ad-based business model, it seems like this could lead to more biased responses in areas like product recommendations. I wanted to take a look at ChatGPT's product recommendations before this shift takes place, so I started prompting it to recommend the best product for various types of household goods. I tested each prompt multiple times and was surprised how consistently ChatGPT recommended the same specific products. I've documented some of the recommendations here and think it would be interesting to run some more rigorous experiments, which I can write up if there's interest!