“They would feel a kind of sadness, right?”: Slay the Spire 2 dev on using slightly rubbish placeholder art over AI

Those of you that have been playing Slay the Spire 2 in its current early access form will have seen that some of the placeholder art looks, well, arguably bad. I am not necessarily making this argument! But it could be made. As it turns out, according to Mega Crit co-founder Casey Yano, this is very much the point, and has more purpose behind it than just being easier to do.

As Yano explains it in an interview with GameSpot, the rough placeholder art looks like something made in MS Paint because he “wanted to set the precedent that some of the stuff is incomplete, and I wanted to make it obvious. I think things being obviously incomplete is actually pretty important for early access.” Yano does make the point that he thinks the art time probably would have liked to have been further along with the art than they are, he still “knew that we were going to be drawing a bunch of garbage during early access.”

But even more than that, he notes that if they “use art that looks nearly complete, then people would think that that’s going to be the final art. It has to look like shit. It’s important that it looks like shit.” It’s an approach to early access you don’t always see anymore. Increasingly there seems to be this expectation that even early access games should have a particular amount of polish to them, so it’s nice to see a novel attempt at pushing back on that.

Perhaps most importantly though is Yano’s thoughts on using placeholder art like this over AI-generated art. “Anybody who has gone through the [artistic] process… I wouldn’t say they would feel cheated, but they would feel a kind of sadness, right?” he explains. “You don’t see the path that somebody took to get better and develop their own style. You just see somebody who’s like, ‘I just want something in this style.'” He does also note how there are artists out there whose job is to imitate certain styles, but “in this case, the game is in our style, so I don’t know why we’d need AI to mimic our own style.” Always a win to see a dev thoroughly disinterested in generative AI. And I’m sure Julian will feel vindicated about this one too.