AI Generated Imagery

[Max Peck]

DALL-E does not grok Mary Poppins.

[hepcat]

At least it didn’t put her in a sari.

[Unagi]

DALL-E does not grok Mary Poppins.

That’s really quite atrocious

[Carpet_pissr]

DALL-E does not grok Mary Poppins.

That’s really quite atrocious

[Daehawk]

Its ruins all the faces on whatever it makes for me.

[Max Peck]

How come everything it draws for me has its faced mushed up? Its kinda freaky. I mean its NEVER clear.

At least one face that it can draw well is Henry Cavill. But it still messes up the eyes.



Just never ask it to draw a perfectly normal human hand.

[Daehawk]

Yep totally normal hand..nothing to see here. im starting to think it uses drugs or is slowly going insane.

[Isgrimnur]

[Max Peck]

[Jaymann]

I made this with a description to face generator. More realistic, but some images took over an hour to generate.

[Chraolic]

Wow, it's pretty good at art mashups. The middle left one pretty much nailed the style I was going for.

[Blackhawk]

Now I'm singing "Riddle me this, Batman" to the tune of "If I Were a Rich Man."

And now the Batman in my head is on a gargoyle, and with a Superhero Cafe voice, is singing "Sunset, Sunset."

[AWS260]

[Rumpy]

I guess stealing and eating all those hamburgers has made him grow an extra leg.

[Hrothgar]

[AWS260]

We can fire all the comic book artists -- AI has it covered.

[Max Peck]

I feel like DALL-E understands Rob Liefeld better than it does Jack Kirby.

[AWS260]

I feel like DALL-E understands Rob Liefeld better than it does Jack Kirby.

Checks out.

[milo]

No feet visible. Checks out.

[Max Peck]

AI art is challenging the boundaries of curation

In just a few years, the number of artworks produced by self-described AI artists has dramatically increased. Some of these works have been sold by large auction houses for dizzying prices and have found their way into prestigious curated collections. Initially spearheaded by a few technologically knowledgeable artists who adopted computer programming as part of their creative process, AI art has recently been embraced by the masses, as image generation technology has become both more effective and easier to use without coding skills.

The AI art movement rides on the coattails of technical progress in computer vision, a research area dedicated to designing algorithms that can process meaningful visual information. A subclass of computer vision algorithms, called generative models, occupies center stage in this story. Generative models are artificial neural networks that can be “trained” on large datasets containing millions of images and learn to encode their statistically salient features. After training, they can produce completely new images that are not contained in the original dataset, often guided by text prompts that explicitly describe the desired results. Until recently, images produced through this approach remained somewhat lacking in coherence or detail, although they possessed an undeniable surrealist charm that captured the attention of many serious artists. However, earlier this year the tech company Open AI unveiled a new model—nicknamed DALL·E 2—that can generate remarkably consistent and relevant images from virtually any text prompt. DALL·E 2 can even produce images in specific styles and imitate famous artists rather convincingly, as long as the desired effect is adequately specified in the prompt. A similar tool has been released for free to the public under the name Craiyon (formerly “DALL·E mini”).

[AWS260]

No feet visible. Checks out.

I posted that image on the QT3 forum and someone made the exact same comment. Liefeld's legacy is secure.

[Max Peck]

[AWS260]

Ha! I laughed out loud at that one. Quite a bit.

[Max Peck]

Like I said, DALL-E gets it. Two thirds of the images of feet have no feet.

[Max Peck]

[Max Peck]

AI-generated Baldur's Gate character portraits put players' to shame

Choosing a character portrait is one of the key decisions to spend hours mulling over in any decent CRPG. That’s why it’s awesome to see someone using the power of publicly accessible AI to generate fantasy mugshots for classic examples of the genre such as Baldur’s Gate. Indie app dev and documentary editor Alex Hay has created some great examples, which he’s shared on Twitter with Baldur’s Gate: Enhanced Edition devs Beamdog and their CEO Trent Oster.

Hay used one of artificial intelligence research lab OpenAI’s image transformer models, called DALL·E 2 to generate portraits in the style of the first two Baldur’s Gate RPGs. There’s more examples of what Hay’s been generating on the Dungeons&Dalle Instagram account, or here on Twitter. Hay used this as a prompt for DALL·E 2 to generate images based on the legendary hamster-loving barbarian Minsc, for example:

A portrait of a giant bald barbarian with purple face paint. A hamster sits on his shoulder. The scene is at sunset and he wears armour made from leather. Epic fantasy art, sharp, award winning on Artstation, 4k HD

[Isgrimnur]

To see this content please enable targeting cookies.

[Max Peck]

To see this content please enable targeting cookies.

I've don't see that because I just block their scripts to begin with.

Of course, it's possible that I decided to block their scripts because I saw that. It's a mystery lost to the mists of history.

[Max Peck]

[Hrothgar]

So Bill Harris has been posting about Midjourney.

PC World has more information about the open beta and the cost to continue using it after your free tokens run out.

For now, beta users will receive about 25 free images as part of the Midjourney beta. After that you’ll have to pay either $10/mo for 200 images per month, or a standard membership of $30 per month for unlimited use.

There’s only one catch with Midjourney: prompts and the resulting images are generated via Discord, the social networking/chat app. To be part of Midjourney, you’ll need a free Discord membership. (PCWorld’s beginner’s guide to Discord can help you get settled if you’re new to the service.) The following link will sign you up for the Midjourney service: discord.gg/midjourney.

Here's one of Bill's:


Here's one from PC World:

[JetFred]

After getting just giddy over Dall-E Mini/Craiyon, now Midjourney has me so floored that I don't even know what to say. It won't be long before you won't even need a game with your AR/VR headsets, just procedurally generated real-time environments by the AI. The only barrier to that now would be processing power and size, same as always.

Anyway, yes I plan to actually give some of my money to Midjourney. Ho-lee crap.

[Max Peck]

Craiyon has a very suspicious Minsc-sized hole in it's understanding of the cosmos...

[JetFred]

Yep, gave my money to Midjourney and I'm more addicted to it than I've been to any game other than Planet Coaster and Mini Motorways in a long time.

[Max Peck]

[Isgrimnur]

Odd balls

[JetFred]

A friend and I cannot stop making D&D locations, items, creatures, and portraits with MidJourney. These selections are typical, not exceptional. I'd estimate about an hour or two of tweaking to get these ten images. Click for full.

[Max Peck]

Noice!

[JetFred]

The prompts we are using average about 35-40 words of description. Once you get good at it, it often gets everything you say into the image. There is a character limit of a bit more than that, beyond which it just ignores any more words.

[TheMix]

Those are amazing.

[Jaymann]

I tried creating a scene out of a book I am writing. It was supposed to be two men fighting, one of them bald and muscular.

This in your face idea was not what I was going for, but it actually makes sense in context. The white whiskers are in the book but were not specified, but one of the images had them anyway. Incidentally, "shirtless" and "bare chested" are forbidden words (even for a male) but it was rendered that way without them.