I was playing around with Bing this afternoon; that’s pretty cool technology! Had it generate a few covers based on classic series like Outrun, Sega GT, that sort of thing.
Figured I’d give a different prompt a try as well. So here’s “Retro-futuristic image of a white haired older man, dressed in a white 80’s suit and sunglasses. Stepping out of a futuristic, red Lamborghini Countach with white interior. Background is an 80’s Miami street with palm trees and art-deco buildings, sun-drenched, with neon lights on the buildings.”
That’s a vibe for sure.
This is awesome! It’s so neat to see other people’s taste and what Bing and the others do with it.
I’m not even sure if ‘thank you’ is an appropriate response here, since the tech did all of the heavy lifting :D
It’s very impressive for sure. I’ve seen what others made with it, so I figured… I’d see what ‘it’ can do. So, I just input some prompts based on stuff that was in my head, and it actually did quite well with what I imagined. It also did quite well with the style I imagined on other prompts. Stuff like ‘sunny, colorful, retro-futuristic’ gave me this rich, dreamy, 50’s advertising vibe that can give you an interesting vibe.
Of course, there’s plenty of silly things to be found in the resulting images if you look closely. This one looks quite error-free, others had some odd perspective issues, cars driving on the wrong side of roads or empty, people walking out of place or facing wrong directions, etc. Here’s one that I made with a prompt like ‘people boarding a retro airliner while an attractive stewardess waves at us’. Looks generally OK, except the plane has no wing and the stewardess appears twice as large as anyone else in it. A few others had similar surreal qualities.
It is interesting to think about generated images. It really is just seeing what it does with the words we put together, but each user comes away with a little bit different style from the choice of words, the ideas each of us are interested in, and which variant from the responses we choose.
So while you can’t take much ownership over the computational work that was done to make the image, I think you can take ownership of the style and thought that you put into it.
I like to think I’ve got at least an impressive enough vocabulary to work with as a professional writer :D
Though it does feel weird to be that descriptive to a piece of software. It’s almost like you’re telling it a little story about the thing you saw in your head and it’s trying to literally ‘paint the picture’. Bing actually encourages you to write more in-depth prompts, and I actually ran out of space on some of them, while I definitely had potential details to share with it. I suppose that means I’m at least giving it enough to work with :D
Yes, it’s weird and so cool how being good at describing things suddenly became a good catalyst for the computer to spit out whatever pictures we have in mind.