And DALL·E 3 (accessible via ChatGPT 4, and Bing [not recommend]) gets the spirit of static images of TV idents. I tried loads of combinations, with some brief and convoluted prompts for the likes of Tyne Tees and BBC1 globes. The results were not great. Hopefully someday a reliable DAN (Do Anything Now) prompt will appear for ChatGPT 4 (I have pushed it at times with interesting results).