AI in video games is a caustic enough subject that Valve requires developer disclosure if a title utilizes the generative technology. This way, people who have qualms about AI or its impact can opt out of purchasing anything that uses the genAI. One developer, however, is saving everyone from the moral quandary in the first place by just deleting their game altogether.
Hardest is a free-to-play roguelike on Steam that was released in the summer of 2025 with the tagline, “stop time, summon tsunamis, shoot with bubble guns, feed cards to mimic, collect rare negative cards!” Except for a user who says the game helped him bond with his son, Hardest mostly got a negative reception. “I assume the whole thing is AI slop,” one reviewer wrote.
You’d think flopping like this would be the end of the story, but half a year later, Rakuel, the developer, has undergone a revelation. On Jan. 10, the indie creator posted an update to Hardest announcing that he would pull the game from the platform by the end of the month.
Why doesn’t he just… I dunno, develop the AI bits? Is he just going to give up?
The announcement suggests the developer wrote all the code, but used the slop robot to generate assets. Sounds like the issue is that making art assets actually takes skill, and is something most programmer types underestimate.
I’d agree with this. I’m a dev, I can make things work, I can’t do art/graphics/assets/etc for shit. Give me 5 or 6 hours and I might be able to get you one image that is semi passable if the intended artistic style is “3 year old with crayons”
Also, it still kinda feeds the AI narrative if he recodes the AI part.
“See? He used AI to make it faster and get some money then he went back and touched it up, really helpful tool”
This challenge never stopped Chris Sawyer and he went on to develop the most influential video game for his time.
You could simplify art hella if you dont go detailed, procedural (nonai) is fun
I’m curious if a dev that carefully manages placeholders could at least garner interest from artists this way. Clair Obscur’s debacle with their Indie Award demonstrates how horrible this can turn out if they miss even one asset; but sadly, I empathize coming from a position where I devoted my studies into learning coding and writing techniques, not artistry.
My space game was cubes and cylinders colliding.
I’m not entirely sure what you’re asking, but the main difficulty here is that using AI, even just for temp assets, is a virtue signal that demonstrates bad virtues. That’s why it’s socially repulsive. It’s like inviting someone into your home and watching them stick their fingers in the soup.
It’s not that using an AI asset for exactly 5 minutes only before swapping it out, and never even committing it to your git history—it’s not that this disqualifies your work from being meaningful in other ways, it’s just that being weak on this front, morally, makes you seem like kind of a dipshit. It’s a failure to reject the siren’s song that leads sailors to their death, you know?
And for what it’s worth, I love seeing passionate work. As a proper art enjoyer, a professional liker of things, cubes and cylinders do nothing to dissuade me.
I think he means trying to make a proof of concept to get an artist to do the work.
If I ever get round to making games, I will need an art guy and a sound guy, because I know full well I cant do anything remotely good there. But I cant afford to pay an art/sound guy, so theyre not going to just do work for a random guy on the promise of a cut of the profits because what profits? I’m just a guy with a dream and a promise.
Incidentally, if people had skill, they wouldn’t use Ai?
Again, reading the announcement, it sounds like the advice the developer received in school was to use it. He’s realized now that it was a bad call.
It sounds like the bigger issue was that the game was bad.
He says he might release a new version with original assets in the future