AI art

Are moustaches coming back in style?

I kind of hope not. I’ve seen a few lately and I’m getting strong 70s vibes. This is not something I have been craving, I should note.

Also, I asked Adobe Firefly to give me “A man with a large moustache standing on a sunny sidewalk, holding a cat in his arms, laughing; horizontal orientation” and this is what it produced:

Pros:

  • There is a moustache
  • There is a cat
  • There is a sidewalk
  • It is sunny

Cons:

  • He is not really laughing
  • Is that a large moustache? I say no.
  • There is nothing horizontal about this image. Maybe I should have specified “landscape.”
  • That cat is terrifying
  • The hair is also kind of terrifying

How can 2019 already be five years ago?

Time is weird. With a cat.

I mean, really. Time is weird, and it just gets weirder.

I still remember when I was really young, and my family was at an event (I want to say a car race or something involving vehicles) and I asked my mom how much longer it would be (because I was apparently not being sufficiently entertained) and she said about ten minutes. And I had no idea how long ten minutes was. I inferred from her tone that it wasn’t very long, so that’s how I started to learn how to tell time–tone of voice.

The time for this post to come to an end is now.

Bliss-ish

Perhaps the most famous desktop wallpaper ever, if only through ubiquity (and it’s also pretty!), is Bliss, the image of a green hill against a pleasant blue sky that was the default background on the Windows XP desktop.

This:

I asked DiffusionBee to make an image reminiscent of Bliss, and it came up with this:

It’s all right, and you can definitely see the family resemblance, but it’s a little too candyland for my taste. I could rework the prompt, but my curiosity has been sated. In the sage words of Homer Simpson, “Eh, close enough.”

EDIT: Ok, I tried once more, using the Image to Image option, and it came up with something a lot closer, yet slightly different, mainly in that it removes the road and the background mountains tucked in the corner. It also seems to think no one would ever cut the grass.

Then I did the exact same thing, but added “detailed” to the prompt and it did indeed add some of the detail back, while making the shadows more dramatic:

OK, now that’s close enough.

Hardworking Sasquatch

I asked Adobe Firefly to give me a Sasquatch working in an office on Adobe Photoshop, using the photography filter. It gave me this:

Sasquatch isn’t real (or is it??) but AI still having trouble with fingers.

I mean, would you tell him lunch break is over?

My most important use for AI-generated art

To amuse me. If it makes me laugh (or smile), I’m good.

Yesterday, a discussion in Discord lead to speculation over whether the Sasquatch had arrived in Australia via dugout canoe1This started when someone in my gaming group, who is going on a trip to Australia soon, noted “my research has indicated that there’s a Bigfoot like cryptid called the Yowie that lives in Australia“. I asked Adobe Firefly to produce illustrated proof of this, and behold:

Also, I just generally love the way Firefly depicts the Sasquatch

This is all I really need AI to do when it comes to art. Sasquatches in canoes.

Bigfoot eating pancakes

Adobe has gotten into the text AI business and claims their system avoids the whole “using artistic works without the artist’s permission” thing–which is great! They’ll probably attach a hefty subscription price to the service once it’s out of beta, which is…vintage Adobe.

For now, it’s free, and I’m in the beta, so here is Bigfoot eating pancakes on a sunny outdoor deck.