I thought that generative image AI wasn’t able to generate the same person multiple times in different pictures and contexts. When did that change?