Audio Blog

〰️

Audio Blog 〰️

Just a warning that this month’s blog is going to be really weird and rambly. It will be a heavy op-ed on AI as well a deep dive into my personal life as of late. I’ll try my best to weave the two subjects into something interesting so that it doesn’t come off in any way to be self serving or boring as a whole. And of course, we’ll end the blog with some lewd stuff cause it’s me after all and a clear sign a bot didn’t write everything!

In the last couple of months, I’ve done my best to dodge the topic of AI on every website and news post I’ve come across. It’s not that I have no interest in the subject, but the conversation has not been really in depth on the material except sensational headlines and short audio clips. This is worsened by the fact that the people who speak loudest on the subject have absolutely no clue how it works. The above video from Mentour Now! is a rare level headed view on AI “in general” and what it really is in its current state and implementation. In my interest circles (namely pop culture and video games), AI has been “downgraded” to nothing more than a tool used to make “copy cat” art. It also has the time to write strange articles mimicking objectively horrible mainstream video game aggregator pages. It’s a shitty narrow view of what AI really is, but it’s the rhetoric of Discord and social media; hence my avoidance on the subject as much as possible until now.

I implore you to listen to the audio portion of the blog to fully get all the nuance I’m about to drop about AI as it stands in the art world. In this written portion, I’m going to grossly oversimplify my points as much as possible. AI art generation as it stands is doing its best to use pre-established inputs to make an amalgamation of outputs that suit a criteria base. In essence, AI art at the moment is like a human cosplayer. AI is doing its best to resemble a particular character / subject with whatever conditions its restrained to and resources (inputs) available. Human cosplayers are no different in the same context, as they work within the boundaries of what they have available to them (resources and talent) in order to mimic someone (or something) from a fictitious environment. Basically nothing is created from nothing, but something is created from something else. I hope this extremely limited example makes sense and that AI isn’t stealing anyone’s art; it’s human beings telling it to mimic (or steal) something else because it’s already there (the concept). The above two pictures are what people consider to be generated “normally”, as the Gardevoir drawn by Faejunkie is a heavily anthromorphized Pokemon wearing skimpy armor. The cosplayer (source unknown since it’s public domain on pixabay) is an actual photo taken of an actual human being.

The picture above is recent and is entirely generated by an AI. The AI was told to draw an “Otaku Circle Princess”, a concept so foreign and meme like that any non-Japanese English speaker in the world would have absolutely no reference about. Neither did the AI that was asked to draw it too. So using what information it could gather about Otakus, Circles, Princesses, and similar group photos of actual people; it did its best. Naturally you have a bunch of errors right off the bat that is typical of AI drawings. The lighting (and cast shadows) are all wrong or off. The lighting intensity is different across the individuals in the scene as well as the background. Background shapes and figures are obviously off (check out the car in the backdrop). And of course the additional digits on any hands in the photo are pretty evident. But all of that really isn’t the point I’m trying to convey. What I am trying to convey is that the AI is doing its best to draw something “we” like and is familiar to other reference points. I’ll list what the AI did that you probably didn’t know

Everyone is drawn to be blemish free
Everyone is relatively attractive and youthful
All standing positions are modeled after real photos of people standing for a picture
The background is a 1:1 copy of a photo that exists, but is not in itself a real place
There are thousands of photos of 6 people standing like this; AI just copied the positions
Otaku is a Japanese term. AI uses models that mostly look Korean because K-pop is popular
Otaku personality is not taken into consideration for the fashion choices used for clothing
AI always uses metrics to make its decisions on what people like rather that what is accurate
Dozens if not hundreds of people critiqued previous attempts to do the very same thing
After enough people giving feedback, AI will eventually “accurately” draw an OCP by the standards given by people
AI is only as accurate is what is people say it is. If everyone said the color of a strawberry is neon purple, it would only draw strawberries as neon purple

At the end of the day, AI is extremely dependent on the feedback we give it. And the most important thing to note is that many of us who appreciate art (in whatever form it comes in) are not ourselves artists. So there’s a clear distinction to be made that the end product is often times “good enough” for those who don’t have a critical eye or have reference points to draw upon. What I’m getting at is that we like what we like, and there are numerous things out there we didn’t realize the source material to (nor care of) its roots. AI itself is no different, in it cares not what it draws for accuracy. It draws because people want it to draw something they like. And it will only be as creative as the people who it references.

This next portion of the blog is completely skippable if you don’t care, as it’s a personal update on my overall health as of late. I highly advise you all listen to the accompanying separate audio blog in regards to this as I explain in great detail what I’ve been going through the last couple of months. For the written portion of course, I’ll be as brief and drama free as possible. My eyesight is failing me, particularly in my right eye in which I have been diagnosed with Retinal Vein Occlusion (RVO). My right eye has non-diabetic macular edema and has a high chance of being permanent for the rest of my life. Treatment (that I’m on) is really meant to stop any further degradation and “stabilize” what good vision I have. The picture above is a mini-example of what my vision looks like in my right eye. The center panel is my current state, while the right panel is what may happen if I didn’t pursue treatment as I did.

I want to make clear I do not want my condition to be some sort of excuse for pity or attention. My condition is really just an example I want to use to help motivate others (like yourself) not to procrastinate in life. The above video is a good academic presentation about the harm of putting things off. I can’t stress enough that life can change at any time with any warning to anyone. It is why I try my best to avoid drama whenever possible on my blogs because the little time I have every day can be afforded to things I enjoy. I can’t fathom the notion that anyone would dedicate any of their time to negativity and wallowing in it, when the clock is ticking down for all of us the moment we’re born. So I’m encouraging you from this point forward for the rest of your days; start living life and doing the things you’ve been putting off. Take it from someone who has 90% of their vision left now, you never know what’s going to happen tomorrow for good or ill. So the present is no better time to get things done.

That ends this month’s rather strange (but hopefully interesting) update! The artist of the month is Cho!Cho! and the lovely Eternity from Last Origin. Also a reminder that June marks the beginning of the warmer months here in the northern hemisphere; so get out there and start enjoying some sunlight!