木曜日, 9月 1 2022

Posted By Danny Choo in Instagram

Many folks think Artificial Intelligence (AI) will make humans jobless - but that's what they said about technologies such as computers and the Internet. While I'm sure some folks did end up losing their job, such technologies have undoubtedly created many new jobs that didn't previously exist - like computer engineers or UX developers.

Just like graphic tools such as Photoshop, Illustrator, and Zbrush, AI Art tools such as #Midjourney and #Dall-e empower the creator to create art - but just like the aforementioned graphic tools - one has to learn how to use the software first.

#AIArt tools are dubbed as a way to generate an image from a text prompt. While this is true, AI will give you its interpretation of the text - which is 90% of the time *not* the aesthetic you are looking for.

With AI Art tools, you need to learn how to craft a prompt - which is reliant on your knowledge of popular culture, science (lighting, physics), artist names, movies, games, anime - and anything on the Internet relevant to what sort of images you are trying to generate.

And then, you need to learn variables. I mainly use Midjourney due to the variety of variables available to tweak the generated results in the direction I want.

I don't feel that AI Art tools are taking away creativity - on the contrary, one has to learn creativity in "prompt crafting" to get AI to generate a certain aesthetic - you will see what I mean when you first give it a try ;-)

While AI can produce images based on your crafted prompt - it is ultimately you who decides whether the image is aesthetically pleasing.

Today's concept art is generated using Midjourney - I'm trying to generate an aesthetic for a black cybernetic character - I think a few of these may very well end up being a Smart Doll ;-)

If you are learning graphic tools such as Photoshop or Zbrush, I highly recommend you look at what AI Art tools such as Midjourney can do for your career. Google sensei will lead the way.

#smartdoll

View Original