Two Ideas for Humans Learning from LLMs | Dark Hacker News