In this blog post, the author discusses the process of verifying AI confidence levels, particularly in the context of an AI that operates as a scientist. The piece explores the relationship between datasets and hypotheses, emphasizing the need for critical assessment of AI-generated confidence scores.