A new AI image generator is going to change everything
Thursday Theses: A project called Dall-E will be a powerful tool when it releases to the broader public
Soon, the idea itself will have the burden of creativity.
That’s the promise of a new AI that generates images from text prompts with photorealistic capability.
Projects like Wombo have been available for several month and have been able to deliver creative art ideas, but you couldn’t call the results life-like.
No, watching people use Dall-E has convinced me there’s something here to see. And I think it’s going to be massively impactful.
I’ve been following DALL-E for several months, and now it’s beginning to get traction in the news (granted, this article is about its less sophisticated counterpart).
Finally, on Monday, July 11, I got access to DALL-E.
I’ve written about my appreciation of Julia Margaret Cameron before.
Here’s an homage of her style, centered on the Arthurian character of Dindrane, the Grail Maiden.
Imagine the possibilities:
works of art
stock images
product designs
architectural designs
There’s have a wide range of possibilities with this tool.
Still, I’m the kind that’s not satisfied with just being obsessed with this present tool.
There are other routes I foresee this technology taking.
Isn’t it fairly inevitable this will extend to video too?
Isn’t it pretty much a safe bet next we’ll turning textual stories into full films.
I can already tell, as soon as this becomes available to me, I’m gonna be hooked.
But what do I mean by “the idea will have the burden of creativity”?
The execution of the idea will be left to the machine.
Your prompt has to be creative enough to have an interesting outcome.
Each person has the capacity to be an artist, whether it’s as a painter, designer, or even (pseudo-)photographer.
You just have to have a creative idea (and a specific one, as you’ll learn at times).
This will revolutionize our world.
I’m not a hype guy. You know this.
I’m stating this from a point of fact, regardless of whether the outcome is positive or negative.
Naturally, there is going to be good and bad.
And the technology is not perfect. Sometimes the choices it makes are a bit… surprising. That is to say, they are not always as related to the initial creation after the course unfolds.
Some ideas don’t work as well as others.
Like, no matter how hard I tried, it couldn’t wrap its head around: “centaur picking up a burger and fries at a drive-thru window.”
But alas, perhaps not everything is meant to be.
I did, however, get a mascot I’ve been thinking about inventing for my brand: a quokka.
The quokka they gave me, who I’ve named Carlyle, has more character.
Because unlike the big smiling quokkas we all know and love, Carlyle looks distraught.
Regardless, this is a great sketch of my new porkpie-wearing friend.
The keys are fairly simple: state your subject, style, and specifics.
Just like I couldn’t get it to make me a centaur as the subject of one project, it also doesn’t completely grasp every style.
When I asked it to give me something in the vein of the old Filmation cartoons, it had a hard time with that.
It offers variations in what it provides:
And there are some styles that it’s quite good at replicating. Check out this watercolor:
I’m not sure it totally gave me what I was after in terms of a Renaissance sculpture garden gnome though:
There are other applications of this technology.
Check out this, which is connected to my most liked tweet ever (which is a bit sad, really):
It does a heck of a job imagining the impossibilities.
And that’s ultimately what we’re after.
What are the implications?
Will it cheapen the role of artists?
Well, that remains the perennial problem.
But we do live in a world where we can appreciate Amish furniture and Mennonite doughnuts — it’s the homemade, handmade quality that we appreciate the most.
For the artists of the future who do it in the eventual old-fashioned way, I do believe a deep appreciation will remain.
I think there more examples beyond the Anabaptist world.
The handcrafted never goes away.
MIDI instruments will never replace real life orchestras in the ways that matter most, and digital apps will never totally do way with good ol’ acrylic painting.
I doubt that we’ll only ever want AI-generated stock photos.
There may come a day when taking the more difficult route to doing photography becomes the rage of Instagram and YouTube.
This does seem like its implications could be more intimate to the creative process than either of those examples, so I do have to check my optimism a little bit.
There’s always room for wariness, but it’d be impossible not to wonder about the possibilities when considering a tool this powerful.
Does everyone become an artist now?
Will good art be able to float to the surface in the coming flood?
We’ll have to invoke Alexandre Dumas for that one: “Until the day when God shall deign to reveal the future to man, all human wisdom is summed up in these two words,-Wait and hope.”