What works well and where progress is needed

Dear Seaart-Team, let me first clarify that your Creativity-First Approach is by far the only one that feels like 21-st century to me. Embedding ComfUI as well as giving artists the opportunity to feature their own KI-Apps fits my impression of an open-minded AI provider. However, a smart business model alone is never a guarantee of mid- to long term success. It is clear that moving your engines from an SFW-world to the real world requires filling training data gaps - a process that requires time but it should not require too much of it. Several KI-Apps appear to already be ahead of the curve here. When users’ prompts involve realistic human portrayal, “surreal productions” such as frequently produced by Sono Vision and Sono Light are maybe not quite what (most) users want to see or expect to see. I think what most users - including myself - want to experience is prompt fidelity and foremost—productions that are physics and human nature aware. (If a prompt explicitly calls for including fantasy elements that is a different case.) Where you need to use surrogate constructions to obtain a realistic scene - you may use smart and realistic ones as long as training gaps have been filled. This message is not by any means a complaint but thought of as in inspiration of where to focus on in the near future. Please just move along that way - you are offering a great service!