Ambiguous Images

Take a look at this picture. What do you see? When you look at it again, do you see something else?

Image courtesy Steve Stuart Williams and Tim Urban

Most people see a man, off balance, running into a snowy forest. Then after looking again, they see a dog running toward us. Some people see the dog first and have a hard time seeing the human.

What's going on is that there are two opposite streams of information processing going on in your brain. One stream is like a camera. Light enters your eye and resolves into shapes and patterns that move to the back of the brain and up through the cerebral cortex to higher level processing. 

But while this is going on, the brain is constantly generating theories of what it's seeing and delivering those theories down the pipeline, optimizing what you're actually seeing to fit its dominant conception.

All along you're reality-checking the top-down theory against the information coming up the pipeline from the eyes.

If the first top-down reading doesn't continue to fit the bottom-up facts, you start generating new interpretations.

A similar process happens with auditory processing when you hear a gunshot...or was that a firecracker?...or was someone popping a paper bag? You can feel your adrenaline surge when you think it's a gunshot, and all that changes when you realize it isn't.

Two Ways to Input an Image Generator

There are two basic ways to generate and image with a system like Open AI's Dall-E 2.

 

One way is to write a text prompt such as "Outer Space Man action figure."

Both images so far were generated that way.

You can also upload a photo (above) into Dall-E2 and let it create variations (following): 

I find the first two text-prompted images a little dull, and the remaining three variations seem both horrifying and hilarious. There are a few interesting new ideas in there, but for the most part they look lumpy, awkward, weird—sometimes funny and sometimes disturbing.

I don't think we need to worry (yet) about this technology taking over from human designers.

Open AI's Dall-E 2

Holding Your Water Cup

My water cup for sketching is a 2-ounce Nalgene cup stuck to the steel palette area by neodymium magnets embedded in 2-part epoxy putty. 

There are lots of other sketch-tech tips in the Facebook group “Sketch Easel Builders.”

Gumroad video: Sketch Easel Builders 

Creatively Collaborating with People and Machines

Here's the collab painting after I added Reddy Kilowatt and his little buddy in the upper left.

It's a group effort by me, Ten Hundred, Kiptoe, Jess Karp, and Greg (Craola) Simpkins.

Here's a link to the the new video about it on YouTube. 

After finishing my section I got access to Dall•E 2. 

OpenAI, the creators of this new computer-generated art system, describe Dall•E 2 as an "AI system that can create realistic images and art from a description in natural language."

I tried writing a prompt to see if it could generate a character that was similar to Reddy. 

For this one, the prompt was "Reddy Kilowatt is a marionette puppet from the 1950s, waving to us, Rolleiflex TLR photo."

Although I didn't use Dall•E 2 in my creative process this time around, it was interesting to try this experiment. 

The technology still has a way to go in terms of image coherence and collaborative interface, it's one of a powerful set of new digital tools that many artists are likely to adopt into their creative process.

Collab Premiere tomorrow at 11AM

Should I burn this painting? Find out what happens tomorrow, Thursday, at 11AM New York time, when I premiere a new video about a big YouTube art collaboration experiment.

I join in with YouTubers Ten Hundred, Kiptoe, Jess Karp, and Greg (Craola) Simpkins.

The video also includes a visit to the toy collection of Mel Birnkrant, AND my first experiments with the AI art program Dall•E 2.

Menzel's Skull Studies

 Skull studies, Adolph Menzel, charcoal with white gouache highlights

Warm Winter Scenes

Some of the most effective winter scenes are essentially warm in color. 

John Everett Millais - Glen Birnam

The camera doesn't usually see it that way. Cameras would probably see the fine branches as blue.

A City Fairyland Winter 1886 Copley Square Boston by Childe Hassam

Palace of the Fine Arts

Palace of the Fine Arts, a remnant of the 1915 Pan-Pacific Exhibition.

Arcane: Documentary on the Making of an Animated Series

Arcane is an animated series on Netflix based on a video game called League of Legends. The series was a big hit for Netflix and for Riot Games.

It was an artistic breakthrough too, weaving existing characters from the game into compelling story arcs, and mixing 2D and 3D animation in a way that made it look like paintings come to life.  

Nine months after release of the TV show, they have released the behind-the-scenes documentary on the story of the making of the show, and the documentary is free to watch on YouTube.

As behind-the-scenes documentaries go, this one covers all the beats: interviews, studio shots, and informal scene setting. They went the extra mile here, with some really intriguing camera angles and settings, and creative editing. There will be additional episodes of this behind-the-scenes series, and the first one ends with a good cliffhanger.

It's Hard to Classify Will T. Bradley

Will T. Bradley (1868-1962) was the highest paid American artist in the early 20th century*, but don't feel bad if you haven't heard of him.

At a time when magazines were the chief vehicle of cultural expression, Bradley was a prolific and noteworthy cover artist. 

His style featured simple shapes, smooth lines, and large areas of flat colors, which lent itself not only to covers, but also to posters.

When he wanted to, he could draw in a more ornate pen-and-ink style, influenced by Aubrey Beardsley and Alphonse Mucha.

Bradley was a type designer he created the original version of what is now known as Bradley DJR (designed by David Jonathan Ross.) And he was also an art director for Hearst films.

---

*More about Will T. Bradley at Wikipedia