How is agent society simulation going?

Three years past since Stanford Generative agents, is there any remarkable progress in this topic?