It is interesting what the nbp model takes away from the prompt, though

Eg instead of focusing on the artist, it focuses on the location

This makes sense! I imagine it was trained in some sort of rlvr like way where you give it a prompt and then interrogate "does this image ..." (where each question examines a different aspect of the prompt)

It's obviously an incredible model. I think there's a limit to how useful another article praising it is in contrast with one expressing frustration

I would also welcome someone writing a short takedown where they fix the prompts and get better-than-2022 results from nbp

> I would also welcome someone writing a short takedown where they fix the prompts and get better-than-2022 results from nbp

NBP (and the new ChatGPT generator) are integrated with LLMs to various degrees, so seems like the obvious starting point is a reverse approach: ask them to describe the old images which has the esthetics that Fernando Borretti likes, and start generating from those prompts. If you can recover the old images, then it was just a prompting issue. ("Sampling can show the presence of knowledge but not the absence.") If you can't even with their own 'native' descriptions, then that points to mode-collapse (especially all of the 'esthetic tuning' like DPO everyone does now) as being the biggest problem.