Thoughts on Trying AI Art for a Year and Then Quitting
I've been into AI art (mainly LoRA training and prompt engineering) since last year, but I've almost stopped these days. Here are some things I realized along the way.
1. Fun at first, but then it becomes studying
Setting up prompt compositions, organizing LoRA training data, tuning hyperparameters... Once you go beyond just a hobby, it's just machine learning. I spent more time with code than actually having fun drawing.
2. The community atmosphere is getting more hostile
At first, it felt like everyone was researching together, but over time, controversies like "theft" and "prompt stealing" increased. It's one thing for AI art to be attacked by opponents, but seeing people in the same camp fight each other really made me feel disillusioned.
3. It's more luck than skill
Even the same prompt and same seed produce different results, and on good days, you don't know what worked. I couldn't tell whether a good image came out because I did well or just because the model happened to work, so the fun faded.
4. Still, I learned a lot
Getting a working understanding of GANs, VAEs, and diffusion concepts became a big asset for me as a developer. I also realized firsthand that using AI tools can definitely speed up your workflow.
Conclusion: I'm not quitting because AI art is boring. I just started thinking I want to actually learn to draw, so I signed up for a life drawing class lol. Hand-drawn art is still the hardest.