This tweet describes a scenario where the 'nano banana pro' AI model was given a user's handwriting sample and an empty paper, and it successfully completed the user's assignment in one attempt. This implies a multimodal prompt where the user provided both a visual input (handwriting/style reference) and potentially an implicit or explicit instruction (the assignment itself) to be solved and written out in the provided style.
Inscrivez-vous pour voir le prompt complet
Inscription gratuite