The point about art being hard for programmers hits home. I hit the same thing when dabbling with game programmer (at a much less skilled level than OP). Difficult to stay motivated when the early drafts look like crap and you’re coding against stickman art.
I’m guessing these days there are placeholder art libraries available?
> Difficult to stay motivated when the early drafts look like crap and you’re coding against stickman art.
I think learning to see past this and be able to evaluate "Is this fun?" regardless of it looking like shit is a skill to learn like any other.
A great way to train this is to start playing random games people publish on low-stake platforms like Itch.io. Most of them lack in the art department, but even some of those have really addicting gameplay hooks, or otherwise novel gameplay elements you can notice shines through the awful art.
Hopefully after a while you'll be able to discern more between "Is this not fun because it doesn't look fun, or because it doesn't feel fun?"
In a way, I feel like you'd get better games if rudimentary placeholders were used for the whole game, and the gameplay and story were focused on, before adding the polish.
I say this because it means you're forced to focus on the gameplay and story. If they aren't compelling, graphics (usually) won't save you anyway.
At least on the surface, thats what I see. Wireframes go a long way.
This is an unpopular opinion on here, but generative AI is getting very good. I think it will soon be the way non artists create art assets for a variety of purposes. It’s not perfect yet, but it’s rapidly improving.
How so? Can it produce a sprite sheet with a believable 'walk cycle' for a 2D character? Can it produce a rigged 3D model?
As usual, it seems like we're maybe _almost_ there (if you can generate video, like with OpenAI's Sora, you could probably get a walking animation for a character, and I've seen proof-of-concepts which produce not-rigged 3D models), but it seems like AI can't do a lot of things that you would want for game development.
The one thing it _is_ really "good" (emphasis on the quotes) at is generating static 2D assets like character portraits, HUDs, item/skill icons, etc. Unity's asset store is now full of gen-AI stuff like this (lots of packs of 1,000 spell icons, which are all basically variations of a 'fireball', except maybe this time it's green, etc).
There is an entire vast ecosystem of services and a community of artists offering production quality assets of every conceivable type, often for free. Generative AI doesn't solve any problems in this space.
That is great if your idea for a game or other product only requires those already existing assets. However, what about assets that don't already exist? You would have to commission an artist to create them, which costs money that you as a bootstrapped independent developer may not have. It also takes considerably more time than generative AI does.
This is a contrived example but, what if I wanted a Walrus riding a surfboard, wearing a top hat, holding a katana in his right hand, and holding a slice of Hawaiian pizza in his left hand.
Despite the biases people have against generative AI, it will solve a LOT of problems in this space.
>That is great if your idea for a game or other product only requires those already existing assets.
Which many will. Just look at the indie games on Steam, a vast amount of them use pre-existing assets.
Jim Sterling poisoned an entire generation of gamedev's minds against it, but there's nothing wrong with doing so.
>You would have to commission an artist to create them, which costs money that you as a bootstrapped independent developer may not have.
If you want a game with quality, unique artwork, likely a style you want to build a brand around and monetize, you should be willing to spend the money on an artist to create it. Using a technology which is trained on the work and style of artists (without their permission, mind you) to extrude an art-like product just to avoid having to pay for it is gross.
>It also takes considerably more time than generative AI does.
Does it? Chances are that "Walrus riding a surfboard, wearing a top hat, holding a katana in his right hand, and holding a slice of Hawaiian pizza in his left hand" is going to be replete with errors, not have a consistent style, have bad geometry if it's a model, and need to be edited anyway. It isn't going to be what you imagined in your head, because generative AI is a mediocrity machine, and it isn't going to compete against a coherent design implemented by real artists who care about their work.
You'd be better off just buying assets or hiring an artist either way. It isn't even that expensive, artists are desperate for work now that AI is eating them alive.
Hiring artists is still a cost and potential time sink, for a game done in your free time that is really unlikely to turn a profit. Generative AI is free and takes just a few minutes.
I can actually draw pretty well but art for games is a whole other thing. Like, my ability to draw portraits doesn't really help that much haha. I mean it probably does but it's a slightly different skill and the sheer amount of assets needed is overwhelming, so it still takes ages.
I’m guessing these days there are placeholder art libraries available?