Highly intriguing paper in many ways. There would be multiple paths to continue from here, including, say, the privacy of prompts and the demographic populatio

Highly intriguing paper in many ways. There would be multiple paths to continue from here, including, say, the privacy of prompts and the demographic population using these tools. Regarding the latter, I'd postulate that it is dominated by young men, and hence no wonder young women dominate the topics, which I'd argue is fine (with some strings attached). There you have one sociological hypothesis. But now, the real gist is: "Despite the vast volume of new synthetic images generative AI has brought us, its seems that most people using these systems are after 'more of the same'. In the process of working with these datasets and being exposed to many thousands of images, our view is that few if any AI images are memorable in the way that human art is, and its difficult to see how they can be aesthetically unique in the way that human art can." I remember reading this argument also in media (The Atlantic, it might have been). Fundamentally, we must then ask: Why is it so? That homogeneity not diversity increases despite the limitless possibilities offered? Subsequently, another sociological hypothesis: Mass culture, not subcultures? For background refer to: https://news.ycombinator.com/item?id=39161863 https://news.ycombinator.com/item?id=39161863