GTP-3 really rocked the HN world a few years ago, but the general public never heard of it. Prediction: people love images and this could be the moment AI generation goes mainstream.
Let’s say Art. What makes art valuable in a sense is that artists have an opportunity cost of doing that art. It’s a sacrifice of something else in their life. AI art gets inflated instantly. Everyone can get a beautiful painting in a button click. Not valuable.
Can someone come up with a use case for this?
I could use it for...inspiration, stock images for a demo site, actual images for a live site, t-shirt designs, coffee mugs, and those drop shipping products, podcast or album artwork, etc.
Yes, maybe design starts to blend, yet lots of web designers used Bootstrap and Material Design so that sites started to look the same anyway. Maybe it'll just be things looking the same except with more artistic images.
i believe dall•e is even more useful than gpt-3 because as of now, only real business model that worked for gpt-3 was copywriting.
dall•e will have branches under images like tshirts, replacing desingers or providing designers with more inspiration, stock photos, nfts & so much more...
I’m not convinced that people would pay for that at all.
It could be a helpful tool for a designer but if you’re saying someone would pay to get something done in the press of a button click I think that’s a very short term business that will be outmatched by an open source free alternative almost immediately.
Just look at how fast web designers are getting outcompeted by free frontend frameworks like Bootstrap or icon libs like Font Awesome.
Perhaps I’m just explaining the natural way of business and commodities, and most goods and services follows this pattern.
- "Get a poster of your writing prompt" -> I think some of the t-shirt/poster companies will definetely incorporate DALL-E, and let users choose from a set of designs based on their writing prompt.
- NFTs maybe? And just playing around with it for fun or views on social media. I also think somewhere there must be a fun little game one can make based on it. There was a pretty awesome rpg_generator for GPT3, maybe some kind of puzzle game based on DALL-E?
A custom shirt printing company would offer the generator. It's up to the customer to decide whether the output can be legally sold (or, perhaps, they're not selling it).
Anything you wish can be printed today, any logo of any brand. Only selling it is as a product restricted.
Copyright law doesn't only cover when things are directly sold (see piracy and the birthday song).
This comment section is filled to the brim with posts missing a "IANAL" disclaimer. IANAL.
For a real world example, I've had real company logos appear on my GAN generated images. I used labels that included terms specific to a single companies rendering technology/pipeline, and the vast majority of images that had those labels had the company logo on it. They were somewhat distorted, but it was very obvious which company it was.
1. https://www.thefashionlaw.com/ralph-lauren-files-suit-agains...
t-shirt is a good idea. many companies will incorporate it definitely.
From an artistic point of view, the value is zero. Dall•e pictures will never have any Aura to them. And for an artist to use them artistically Dall•e is “too easy”, as the artist won’t have produced or worked on the model itself.
All in all, Dall•e is the most exciting, technologically advanced, and beautifully useless AI project I’ve ever seen.
Why is DALL-E (or more likely its successors) not suitable then? It can produce a large amount of similar pictures based on a vague theme.
> Dall•e pictures will never have any Aura to them.
Debatable, I actually really quite like a lot of its art. Regardless, nobody uses stock photography for its "Aura".
Today my personal biggest use of AI is for generating photos of faces as profile pictures for tabletop rpg character sheets. I don't need them to be perfect, I just need quick, easy, cheap content that's good enough.
Tomorrow I might use Dall-E to create illustrations for my blog posts, or depict environments for my tabletop campaigns.
They might be "too easy", but they'll be cheap and quick and good enough for the purpose.
this isn't supposed to make designers useless but rather make them more creative.
dall*e can be used as an inspiration to combine multiple images & create their own. it can be used to steal color combination that is aesthetically pleasing on the eyes.
if you think dall•e is useless, you haven't seen the images shared.
the realistic ones are hidden by openai but they can generate a lot of stuff that we won't know until an open-source version comes along.
People who want cheap but unique abstract art for their house, small business owners who need ad clip art but don’t have the budget for a designer, even artists using DallE queries to generate launchpad ideas, I can imagine many reasons DallE would get used. But also don’t forget to imagine the future, this is one of the first and it’s bound to be improved until you can’t distinguish it from “real” art.
There’s a valid point in your comment about how artists currently look at AI, and a valid question about what AI produced art actually means, so I’m throwing you an upvote because I don’t think you deserved to get downvoted for sharing your opinion. Maybe more importantly, AI generated art could challenge and rekindle the age old debate about what art actually is, and bring a new perspective to the table.
You know what's really impressive, that just got out? Adobe's Premiere Pro Auto Color feature based on Adobe Sensei. Ask a colorist or a filmmaker how it works for them. Is that something that was possible because of fringe research like Dall•E? Honestly, no.
As someone who designs and makes fine furniture (not the way I make my living, but I do know the business), I know that there will likely always be some demand for custom, hand made pieces, because people assign value both to getting exactly what they want, and to the fact that something is human and hand made. When people who value those things have enough money, they'll spring for that value. But I'm well aware that the bespoke and hand made will never be more than a 3rd decimal place rounding error in the ocean of factory made furniture that fills our homes and workspaces. Images can go the same way, and will, if the price differential is right. Most images in the "image marketplace" are not art, after all, but rather anonymous commodity visuals augmenting the presentation of an idea, or product.
And I'm sure you could come up with similar areas where sometimes you just need a bunch of cheap art.
https://twitter.com/nickcammarata/status/1511861061988892675
The "ideal" article from a Google SEO point of view had 2 or 3 images spaced through the article text.
My first thought was stock images for news sites and blogs