Yaroslav Chupachupovich

Yaroslav Chupachupovich 

Визуальный разработчик. Концепт художник.

109subscribers

63posts

goals2
19 of 100 paid subscribers
Когда наберу 500 платных, смогу забить на фрилансы и развивать свой проект. Будет больше контента связанного с разработкой сеттинга, ворлд билдингом.
1 of 2

ENG 💬A chill wall of text about AI

So I was sitting there, working on stuff, doing some 3D, sketching things out, and all the while YouTube and Twitter kept flashing posts about how crazy AI tools in 3D have gotten lately. Maybe I’m in some kind of bubble, but I see that kind of thing a lot.
I’ve already voiced my opinion before on why AI, in concept art or art in general, won’t fundamentally change anything. (If you're curious, let me know and I’ll explain.)
But that’s not the point right now. What I want to talk about is the logic (or lack thereof) behind people dropping basic tools and diving headfirst into AI tools. They start downloading sketchy crap, registering on a bunch of weird sites (you know the kind—the "generate for points" ones). Maybe they think they’ll hit some kind of speedrun.
Look, if someone just wants to hustle a bit of cash through stock sites while the AI scene is still blooming—no problem, go ahead. But if they actually believe that this stuff is going to help them creatively or make them more in demand? What the hell, people. Think for a second.
First of all, those who jump to these tools too early are shooting themselves in the foot by skipping out on understanding how visual storytelling actually works. And second—here’s my main take. Yes, it’s just a hypothesis, but to me it feels really obvious given where things are headed:
AI tech is evolving insanely fast because of the money being thrown at it. So grounding yourself in any one approach right now is just not rational. You might blink for a couple months, and suddenly there’s a brand new tool, new methods, new architecture. Let me break that down.
Getting used to using AI tools takes time. But THE APPROACH ITSELF CHANGES LITERALLY EVERY MONTH. Okay, maybe not literally, but close. It’s not like traditional painting, or design, or even 3D, where workflows stay relevant for decades. There, it makes sense to get settled. Not with AI.
In my opinion, it’s way smarter to just spend this time building a solid base—composition, design sense, visual thinking. Because in half a year (maybe more, maybe less), these tools will be built right into your favorite software anyway, like we’re already seeing with Photoshop. Blender's getting there too, but right now it still feels clunky and hacky.
What I mean is proper UI/UX integration. Something you can use like a brush in Clip Studio, or like an edit mode in Blender.
🎨 Let's get practical for a sec
Say you're a concept artist. Like me. You've been tasked with designing a power gauntlet—grade 5—for Warhammer. Big, ornate, over-the-top.
You start with dirty sketches and design exploration. That phase usually takes about 4 hours—coffee breaks, refs, thinking time included. It’s just raw brainstorming to dig into the theme and feel things out. You show it, get some feedback, and then start refining. You bash things in, add paint, solve local design problems, add detail. That might take another 3–4 hours.
In total, you’ve spent like 5–8 hours. Pretty normal if you know what you’re doing.
Now imagine someone doing this with an AI tool—but they have, let’s say, "sausages for fingers." Okay, I’m joking. But seriously, they don’t have trained artistic skills.
First issue: they probably won’t even fully understand the brief the way an experienced artist would. And where does that experience come from? From grinding for years, moving pixels manually.
There’s a weird kind of logic here that a lot of people miss: what takes an experienced artist 30 minutes might take someone without skills half a day. Sounds dumb, but think about it—imagine someone who's never drawn trying to parse what their art director means when he talks about rhythm and material diffusion in design. Sure, maybe they can quickly generate a "version," but to get it to a cohesive place where all the parts make sense together? That’ll either take as long or even more time than if they'd just drawn it.
Second issue: revisions. Someone might say, "But ControlNet and Stable Diffusion can do edits now!"
Bro. Take your pants off before making statements like that.
Yes, AI can technically revise. But if we're comparing to an actual artist? While you're tweaking prompts and masking areas and filtering through garbage generations, a human artist just knocks out the change in 15–20 minutes with a few brushstrokes and some bash. That’s the earlier point again—experience matters.
🧠 But what if an experienced artist uses AI?
Now that’s where it gets more interesting. I can see the benefit. Like, sure, I can toss in a dirty sketch and let the AI refine some of the edges or give me new ideas. That sounds good on paper. But right now, it rarely works that cleanly.
You need the stars to align for AI to refine your sketch in a meaningful way. It only works with certain styles and very specific inputs.
🧩 Here's the real takeaway
You could speculate forever about how AI could be used. But here’s my point:
There’s no need to stress about AI right now because the pace of change is too damn high.
Big companies like Adobe, OpenAI, and whoever comes next are just going to bake this stuff right into the tools we already use. And they’ll do it properly. No janky workarounds. No "make six accounts and download 6GB of random models." Just clean, native integration.
Honestly, a lot of the frustration artists feel about AI comes down to this: it doesn’t feel like real drawing or modeling. There's this internal dissonance. Like, "I want to draw characters, not wrangle blobs using prompts."
But once AI tools become part of your creative flow—you’re on a canvas, mixing and matching ideas, painting directly on top, refining as you go, and it feels like real work again? I think people will embrace it wholeheartedly.
Even in 3D: imagine jumping into a program and not needing to plan a whole pipeline just to get your textures down. You just paint directly on the model the way you want. That’s bliss. And I think AI is going to take us there eventually.
Man, I feel like I’m bouncing between ideas. Let me summarize:
Right now? Just chill. Hold off. Hone your skills. Sketch, explore, understand how the world works. Because when AI becomes a real, practical assistant—something that removes technical barriers instead of adding new ones—you’ll be ready.
Worried that "I spent six years drawing and now some random guy is cranking out stuff with prompts"?
Nah. Sooner or later, we’ll get native AI tools with the same level of control as traditional workflows. And again, the one who wins will still be the person with better artistic instinct, stronger vision, more experience. That hasn’t changed.
And bonus: the AI wave will filter out those who were just in it to farm quick cash by spamming 1000 bricks a day. No shade. Those tasks will be automated. Real creative work will still need qualified artists.
Sure, it takes a certain level of education. But hey—you don’t become a doctor or engineer after a three-month course for 30 bucks, right? And yet a few years ago that was weirdly common in game dev because of how bloated the market got.
Alright alright. That’s it. The end. Hope it was interesting. Drop your thoughts if you feel like it.
This is very close to my own take. I took over a year off to focus on AI and it is absolutely useful "if" you already understand design theory. anatomy, etc. The inconsistency and lack of intention in all AI "apps" pushed me more into Blender - which IMO is the future tool for both 2d and 3d work - not to mention video editing
Subscription levels3

Трикстер

$3.3 per month
1. Записи трансляций. 
2. Записи фидбеков/лекций.
3. Поддержка моего проекта.

ДаркТрикстер

$19.4 per month
1. Записи трансляций. 
2. Записи фидбеков/лекций.
3. Поддержка моего проекта.
4. Бленд файлы, холсты в исходном разрешении.
Subscription Spots Are Limited

ДаркТрикстер +

$107 per month
Все что выше + индивидуальный фидбек+оверпэинт на твою работу каждый месяц. + запись, которая может быть опубликована.
Go up