When I look at what AI has been doing to the emulation scene, I think it has been a net positive. You have some emulators like Yaba Sanshiro and Ymir where the devs have embraced it and game compatibility/feature development has skyrocketed. Sega Saturn has always had mediocre emulation and now it is rapidly improving.
Even for emulators where the devs don't want to accept AI code into the main codebase, people are creating PRs/forks with features that have been requested for years. You can now create a custom version of melonDS with built-in shaders, RetroAchievements, a Vulkan renderer for better Mac compatibility, and texture/sprite layer scaling. All a technical user needs to do is ask their copy of Claude/Codex to glue it all together. The end result might be a bit buggy, but emulation has always been buggy.
I can understand why maintainers are annoyed by these PRs, since they are essentially being asked to take on the maintenance of vibe-coded features. However, I think people should start seeing the PR tab on GitHub as more of an "optional feature list" where you just pick and choose whatever you want.
I believe the end result of this will be an increasing number of forks. You will have projects that are known as well-tested/battle-hardened, people will build AI-assisted forks on top of them, and some features that prove their worth will be merged back into the base project. In other cases, where a project never merges AI-created features, maybe because the devs don't want the maintenance burden or hate AI, most of the new development will likely happen in forks, and eventually one of those forks will become the version of the software most people use. From a user's point of view, the point of software isn't to be an artisanal work, the point is that it accomplishes a task.
p
When I look at what AI has been doing to the emulation scene, I think it has been a net positive. You have some emulators like Yaba Sanshiro and Ymir where the devs have embraced it and game compatibility/feature development has skyrocketed. Sega Saturn has always had mediocre emulation and now it is rapidly improving.
Even for emulators where the devs don't want to accept AI code into the main codebase, people are creating PRs/forks with features that have been requested for years. You can now create a custom version of melonDS with built-in shaders, RetroAchievements, a Vulkan renderer for better Mac compatibility, and texture/sprite layer scaling. All a technical user needs to do is ask their copy of Claude/Codex to glue it all together. The end result might be a bit buggy, but emulation has always been buggy.
I can understand why maintainers are annoyed by these PRs, since they are essentially being asked to take on the maintenance of vibe-coded features. However, I think people should start seeing the PR tab on GitHub as more of an "optional feature list" where you just pick and choose whatever you want.
I believe the end result of this will be an increasing number of forks. You will have projects that are known as well-tested/battle-hardened, people will build AI-assisted forks on top of them, and some features that prove their worth will be merged back into the base project. In other cases, where a project never merges AI-created features, maybe because the devs don't want the maintenance burden or hate AI, most of the new development will likely happen in forks, and eventually one of those forks will become the version of the software most people use. From a user's point of view, the point of software isn't to be an artisanal work, the point is that it accomplishes a task.
More options
Context Copy link