Hacker News new | ask | show | jobs
by amram_art 178 days ago
The problem is not in the image models rather the training data and its context. "British museum" for MJ is the image source, "British museum" is the setting for Nano Banana.