DevLog Ep3

Stable Diffusion came into the world, and reviews from people who tried it started appearing one by one... At first, it was funny.

Images that looked like anime girls composited onto trains. Six fingers on a hand, or eating ramen in some grotesque way.

You've probably seen them too, if you're reading this.

Everyone tried it out of curiosity, posted the results, and we all laughed together.

It was fascinating, but that was about it.

"Oh, a fun toy just came out."

But It Kept Getting Better

I wasn't paying attention, and before I knew it, things had changed a lot. It started drawing proper hands with five fingers, and ramen was being eaten neatly with proper chopsticks.

Backgrounds got cleaner, lighting became natural, and clothing wrinkles turned realistic.

What I laughed at a month ago, I couldn't laugh at a month later.

I thought it was a toy,

but it wasn't.

The speed was the problem.

Technology always advances — but this was too fast.

The speed at which yesterday's impossibility became today's reality.

The Moment the Laughter Stopped

I don't remember exactly when, but there was a turning point.

One day, scrolling through my timeline, I thought "Oh wow, that's amazing!" and zoomed in to take a closer look.

Looked closer. It was AI.

"It can draw like this now..." I thought.

Honestly, I was quite shocked. Until then, I'd figured it was only good enough for a toy.

The Timeline Kept Flowing

Days passed, weeks passed. AI art started blending naturally into the timeline.

Which ones were made by AI, which ones were drawn by people. It became harder and harder to tell.

I bookmarked quite a few — sometimes without realizing they were AI, sometimes knowing but still thinking they were good.

But deep down, I kept separating AI art from human-drawn art in my mind.

Human-drawn art comes first... That's what I kept thinking.