I learn a lot from talking to large language models (LLMs) like Gemini, ChatGPT, and Claude. Frequently I’ll ask the LLM to quiz me on what it’s taught me, in order to test my understanding of what I’ve learned. I find that doing so helps me learn the material better, promoting longer term recall. Even with this strategy in place, however, I still forget a lot of what I’ve learned from these AI assistants.