AI Is Taking My Job and That’s a Good Thing

If you follow me here on the site or on my social media, then you probably know that I haven’t been shooting a lot of photos as of late or writing much.   I can make no excuses.  The truth is, the muse comes and goes, and I do not try to force it when it goes on holiday. Generally, when I have done that in the past, my art has suffered for it.  So, what is all this about AI taking my job and why would I ever think that is a good thing.  Probably should first explain, WHAT my “job” is.

My job as it were, is being a creative. Part of that creativity is also sharing my work, and hopefully helping others out with their roadblocks and creative journeys.    That first word there, “creative” is a tough one and there are many steps to it all rolled up into one simple word.  Anyone, who has tried to be creative or IS a creative knows full well the struggles of the word.  Holding up a consistent, quality stream of creativity is easier for some that it is for others. But with that said, everyone hits a roadblock at some point.  Similar to  owning a motorcycle, it’s not if you crash but when.

Since beginning my journey as a creative I have moved many times between genres.  Music, photography, woodworking, 3D modeling, graphic design, heck even oil painting at one point.  The main focus always was to stay creative. If I didn’t, I usually got into trouble.  Too much partying, bad decisions, you name it.  As long as I was working on something I was stable.  I have faced and overcome many roadblocks along the way but always managed to get by them somehow.  Sometimes by turning around and going back where I came from and starting from scratch and sometimes a simple detour solved the issue.  Now that I think of it, there was a LOT of starting over. And that’s okay. Mistakes WILL be made, and lessons and growth will happen!

As of late, the roadblocks have felt insurmountable.  Just when I felt like I had found a way around my slump in 2020 the pandemic hit.  I had now what I believe was the virus in December of 2019.  All the symptoms that have been described were present.  It was horrible, but I have not been sick since.  Couple that with the mass shootings, angry un-hinged people and well it is easy to just pull back and blame it all on the world.  Truth is, I just haven’t felt it.

None-the-less, I have still sought out inspiration.  Diving into building furniture for our house starting with this coffee table for our living room.

Along with the woodworking I have also been deep into a new 3D print business called AD3D where I specialize in mounting hardware for computer accessories.  And while this all keeps me busy and fuels my mind for creation, it does nothing for my photography at all.

Then earlier this year a story broke in the net about a new AI (Artificial Intelligence) bot that could make “art”.  That bot was named DALL-E.  Part Salvador Dali and Part Wall-E. There had been AI Art bots before. I have made quite a few pieces with a website called Artbreeder, but this was different.  You feed DALL-E a description of what you want to see, and it draws it.  The bot has been hugely popular and in the news with some astonishing imagery.  The original DALL-E has been renamed Craiyon and you can go play with it for free at https://www.craiyon.com

With the public release of this bot and AI in general there has been a lot of talk about it replacing humans.  Of course there would be.  Anything and everything is a threat to humans don’t you know.   You can go play with it and make that determination for yourself.  And that is what I want to address here today with my journey into the realm of AI Art.

Let’s start where I started with Artbreeder.  With Artbreeder you feed in an image, and then start messing with a bunch of suggestion parameters for the bot to create with.  These are in the form of sliders for things like mountains, rivers, sky, trees, etc.  Here is one of the first photos of mine that I fed Artbreeder.

I have sold many prints of this image so I really wanted to see what I could create by using it with Artbreeder.  Here are 3 images that I got back.

I am not impressed to say the least.  But this was my first time using it, so I fed it through again with more adjustments.

Again, I could not see what all the hype, and fear was about.  In everyday speak, this bot SUCKS!  Still thinking though, it maybe that I just don’t know what I am doing, I went at it again.

Okay then!  Not bad, not bad at all.  Impressionistically it was something.  Not my best work but I could see possibilities here.  Not once though did I think, this AI is going to replace me.  Or, oh no the art world, whatever shall we do?  Then again, I am not that sort of person as I usually just roll with it and adapt.  I had a lot of fun with Artbreeder, was using my images and making new works and guess what, that is creativity!  You can see a bunch of these first images on my NFT page over at https://opensea.io/collection/ancient-worlds

Fast forward to the release news of DALL-E 2, a new iteration of the original DALL-E or DALL-E Mini/Craiyon as it is now known.  Here are some comparison images from DALL-E/Craiyon and DALL-E 2. (editors note: DALL-E 2 is now known as DALL-E but is left as DALL-E 2 in the article to avoid further confusion)

I told the bot that I wanted a rusted car in an abandoned city, resulting in the above image.  I fed this same phrase into DALL-E 2.

As you can see, the more sophisticated version of DALL-E does a much better job.  And while this is clearly not a photograph, you can tell the bot you want a photo version.  Here is a “photo” of two that do not exist using the prompt, “A photograph of a person.” with DALL-E 2.

Even though these “people” don’t exist in the real world, they could easily fool you into thinking they are actual photographs. This is the reason why access to DALL-E 2 has been kept from the public until now.  Even in the current public beta (as of 7/30/2022), certain phrases and subjects are not allowed in the prompt. Also, while initially free with 50 “credits” and 15 extra credits each month after, you do have to purchase more credits to use the bot if you run through your 50. And you will.  They charge $15US per 115 credits.  Each time you enter a prompt, you get 4 images to choose from and then you can make variations from there. You can also erase areas and have the bot re-fill those areas and upload your own logos and photos as well.

While this is not a review, DALL-E 2 is a fun toy. And you can get some funny, interesting results by using key words and phrases.  But, I have found another bot that I think is the true artist in the AI world and I would like to share that with you!

Enter MidJourney.  A bot that is centered around the popular communications software Discord.  I started using MidJourney on July 10th when a friend of mine got me a free invite.  With the beta invite we got a limited amount of free credit. I have since upgraded to the $30 per month subscription which nets you  15 hours of what they call Fast mode.  You are able to switch between Relaxed and Fast mode.  The fast mode costing you subscription time while relaxed mode restricts files size and takes longer to produce results.   Midjourney is a less trained AI than DALL-E 2 by about half of the iterations.   It can still make some descent results.  Here is a MidJourney image created with the “A photograph of a person” prompt.

You can probably see right off the bat that MidJourney is a little different.  It has stylized the photo to look more vintage. Added a film look to it. Even added an aged look with dust and specs.   And while MidJourney can do “photographs” of “people” it’s strengths clearly lie elsewhere.  In talking with my friend Walter Arnold we have come to agree that DALL-E 2 is a realist and MidJourney is an ARTIST!  Sure you can create some “unreal” images with DALL-E 2. Like say, a Chicken with a cats head.  Or a hedgehog playing video games.  And while that is fun and all, MidJourney really puts feeling into the results you get back.

Over the few weeks now I have been testing it, the AI has gone through changes and become better and better.  I would dare say that it is probably my favorite creative inspiration right now.  Let’s focus a bit on my MidJourney, journey and then wrap this up.

I started out amazed and confused while playing with MidJourney at first.  There are no buttons, no switches, no drawing, etc.  You tell it what you want.  My first bit of messing around was of course, ZOMBIES!  Having it create zombies at first was a novel idea but I had to make something different.  So I called upon my youth and my love for the 80’s!  With MidJourney you use the prompt /Imagine.  When you type this in you are greeted with a prompt that says….PROMPT.  Here is where you type in what you want it to make.  So I typed in:

synthwave zombie wearing headphones listening to music

A definite shout-out to “Bub” the Zombie in the George A. Romero movie Day of the Dead.  Here are the results.

This was art!  After a few days of learning the language of the bot and being able to control the subject matter better I started to get into a creative groove. YES! I was creative once again.  I am now creating images with subject matter that I am known for, Abandonscapes.  With MidJourney I can control the aspect ratio, the style, the weight of the items in the composition, everything. Or even better yet, I can tell the AI to go nuts and impress me.  Not only does this allow me to create but it inspires me as well.   I have created my first series of what I call Computer Aided Digital Art Works or CADAWS for short.  Here are a couple of examples from the collection (and yes, you can purchase prints).

If you want to see the whole gallery it can be found here: https://portfolio.theexplorographer.com/Gallery/CADAWS/AI-Assisted-Abandonscapes/

The Wrap

In it’s restrictive form, I found DALL-E 2 fairly harmless.  Graphic designers need not fear as the AI cannot spell even if you try to force it.  I tried to create several logos for businesses and it failed miserably as does MidJourney as well.  Users of DALL-E 2 cannot use real faces with the generator, so photographers need not worry about losing their jobs. And as far as I know it can’t not re-imagine real places rendering them nearly unidentifiable.

There is always fear in something new especially if it is directly in your neck of the woods.  Probably the biggest problem I see is the influx of new art that will be flooding well, everywhere.  In an endless sea of creatives, getting “discovered” will be harder than ever. No more difficult though than any other industry or profession.  It does give many perspectives on the size of the world and what you are up against.  I choose to just go with it, use the tools, have fun, die happy.

These AI bots have inspired me to get back to creating. They have my friends as well!  My long time friend, Walter Arnold has created an website dedicated to AI Assisted art based on Song Lyrics called LucidLyricsArt.com He also has an Instagram page for it as well.  Instagram, that’s still a thing right?  Too soon?  Anyway, I encourage you to check these bots out.  They are pretty cool and inspiring to hang around!  Till next time, stay creative!! Beep Boop Beep

~AD

Artificial Intelligence Art Bots to check out:
DALL-E Mini/Craiyon     –     DALL-E 2     –     MidJourney     –     ArtBreeder