alt Hacker NewsThe standard objection: if the LLM is supposedly intelligent, why can’t it figure out on its own that this two-step process would achieve a better result?
Replies
pyrolistical • today at 7:32 AM
You don’t know what you don’t know
cubefox • today at 6:41 AM
Part of the problem is that it isn't the LLM making the image directly itself, it's the LLM repeatedly prompting edits for a separate edit diffusion model. The Gemini reasoning summary shows part of this. The style of some of the images makes it also clear that it uses an Imagen 4 derived diffusion model underneath.
[flagged]