A bunch of players in this space now, some of them YC companies.
I’d love your opinions on this - from what I’ve seen, all companies doing this today are using similar methods (and many are reliant on third party AI APIs, like Dall-E, or are using licensed nets and have nearly zero control of further development).
Lots of issues with the existing implementation (feature misalignment, feature loss, feature hallucination, warping/distortion, very odd looking faces, unrealistic bodies, and much more), but there are an awful lot of attempts at selling this current iteration even though the output is off-putting right now.
If you have real life examples to the contrary, I’d love to see them — but the real question here is will we see this tech replace a large percentage of fashion photographers for large brands any time in the coming year? Or will it take longer?
