Mei needs a shelf-audit answer that compares visual evidence instead of merely describing photos. Before: unordered photo dump After: context, reference, close-up, exception The after version gives each image a job. Context sets the scene, reference defines the standard, close-up carries the evidence, and exception narrows the cause. The after version preserves comparison order. The model can reason over change instead of generating separate captions. The after version reduces token waste. Extra images are useful only when they add evidence, not when they repeat the same angle. Order turns multiple images into an argument. multi-image evidence sequencing Moved from image…
Read the full lesson
Sign up free: one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in