Playing with DALL-E 2

Re:Playing with DALL-E 2 by stickfigure on Wednesday April 13, 2022 @12:37AM (#31011229)

ELIZA and markov chain generators were curiosities without practical application.

DALL-E 2 is useful now. It's easy to imagine it replacing 99designs, Getty Images, and most of the digital art services on fiverr. I can't imagine what is going to happen in the world of print-on-demand tshirts.

And of course porn... what a market.

[ Parent | on HN ]

Re:Playing with DALL-E 2 by kingcharles on Wednesday April 13, 2022 @12:57AM (#31011334)

Magazines and newspapers are regularly paying thousands of dollars for a single abstract illustration for an article. DALL-E 2 is going to replace a huge proportion of that. Not only the cost saving, but you can get an image minutes before you go to press.

[ Parent | on HN ]

Re:Playing with DALL-E 2 by dx034 on Wednesday April 13, 2022 @05:20AM (#31012624)

And what if the image is mostly comprised of an existing artwork with copyright? Not sure I'd take that legal risk as a newspaper.

[ Parent | on HN ]

Re:Playing with DALL-E 2 by stickfigure on Wednesday April 13, 2022 @11:30AM (#31015857)

Then it will be rejected by the person using DALL-E, or someone else in the publishing chain? Maybe an automated reverse image search AI?

I don't think anyone (yet) imagines all the humans at the NYT will be replaced by GPTX+DALL-E. We still have editors even though humans author the articles.

[ Parent | on HN ]

Re:Playing with DALL-E 2 by dx034 on Thursday April 14, 2022 @02:44AM (#31024027)

Does that exist? Can you determine how mcuch of a DALL-E picture is identical to copyrighted material? And what's the threshold anyway for copyright to apply? I'm not sure that's a process with set rules yet.

[ Parent | on HN ]

Re:Playing with DALL-E 2 by whywhywhywhy on Wednesday April 13, 2022 @06:07AM (#31012899)

Until it can deal with a round of feedback without completely changing the imagine then illustrators have nothing to worry about.

[ Parent | on HN ]

Re:Playing with DALL-E 2 by mkl on Wednesday April 13, 2022 @08:03AM (#31013649)

It can already do that to some extent. You can erase bits you don't like and give it a new description to fill them in with.

[ Parent | on HN ]

Re:Playing with DALL-E 2 by chimprich on Wednesday April 13, 2022 @08:05AM (#31013667)

> DALL-E 2 is useful now.

If you look closely at most of the images, they don't look quite right. There are usually artefacts, or slight misunderstandings of the brief etc. Its about 80% of the way there, but I think that last 20% is going to be a lot more difficult.

I'm sure there are some cool applications for this - maybe if you need a quick and cheap image for your newsletter, personal website or an experimental game for example.

For a serious commercial application I would think it would be easier and safer to pay someone.

[ Parent | on HN ]

Re:Playing with DALL-E 2 by stickfigure on Wednesday April 13, 2022 @11:23AM (#31015762)

> I'm sure there are some cool applications for this - maybe if you need a quick and cheap image for your newsletter, personal website or an experimental game for example.

I think this understates the market. For every New York Times article, there are millions of newsletters, analyst reports, tshirts, websites, logos, etc. Many of them are abstract eye candy and don't need photorealism.

Actually, even the high-budget articles often have pretty abstract art. I read this Wired article yesterday, and all the illustrations could easily have been generated:

https://www.wired.com/story/tracers-in-the-dark-welcome-to-v...

[ Parent | on HN ]

Re:Playing with DALL-E 2 by speleding on Wednesday April 13, 2022 @05:00AM (#31012515)

Indeed, you can just wait for someone to try this on porn. (Yeah, I'll admit I'm curious to see what happens... for science!)

[ Parent | on HN ]