In my experience it doesn’t require that much cherry picking if you use a carefully crafted prompt. For example: “ A professional photography of a software developer talking to a plastic duck on his desk, bright smooth lighting, f2.2, bokeh, Leica, corporate stock picture, highly detailed”

And this is the first picture I got: https://labs.openai.com/s/lSWOnxbHBYQAtli9CYlZGqcZ

It got it a bit strong on the depth of field and I don’t like the angle but I could iterate a few times and get a good one.

Additionally, wherever it classically falls over (such as currently for realistic human faces), there will be second pass models that both detect and replace all the faces with realistic ones. People are already using models that alter eyes to be life-like with excellent results (many of the dalle-2 ones appear somewhat dead atm).

Even this image is just an illusion of a perfect photo, which is a blur for most part, see the face of duck. I had access since past 4 5 days and it fails badly whenever I tried to create any unusual scene.

For the first few days when it was announced I use to look deep even in real photos in search of generative artifacts. They are not so difficult to spot now, most of the times anyway.

NB: when you share links like that, nobody who doesn't have access can see the results

sure they can, just tried in incognito

I didn't even need incognito.

Even the high quality stuff still can't do human faces right.

This one surprised me when it came out, felt more ‘human’ than lots of stock photos: https://labs.openai.com/s/AsRKFiOKJmmZrVDxIGa75sSA

They avoided using real human faces in the training data.

If the price is low enough, you can have humans rank generated images (maybe using Mechanical Turk or a similar service), and from that ranking choose only the highest quality DALL-E generated images.

If someone can make money doing it they might.

Heck: If the cost to entry is prohibitively low they might do it at a loss and take over the site

It's a lot better than you are claiming. Mind if I ask if you have access personally?

Yes I have. And I realized it as soon as I started experimenting that mind blowing results are mostly cherry picking.

It's very good at generating art style images. These kind of images are mostly amazing most of the times. But the Photorealistic images only work with cherry picking.

> And I realized it as soon as I started experimenting that mind blowing results are mostly cherry picking

Me and you must have very different definitions of "cherry picking". For prompts that fall within it's scope (i.e not something unusually complex or obscure) I get usable results probably 90% of the time.

Can you give me some examples of prompts that you tried where you found good results difficult to obtain?

I get bad results on unusual prompts, you are right about that.

It did generate good dslr like face closeups, as good as Nvidia does, most of the times but not always. Sometimes there are weird artifacts and face does not make sense.

Dslr style blurry photos are mostly good. From the looks of images I follow, imagen is probably more believable. Don't know how much cherry picking goes on there. See this thread [1] for example. I failed to generate image like this (honey dress) in dalle2.

[1]: https://www.reddit.com/r/ImagenAI/comments/w3saku/creating_i...

Give it a few years. I'd be exiting if I owned a stock site