This is a lot like what a YouTuber by the name of Randy has been doing: he decided to create his own game engine instead of using something like Godot/Unity/Unreal and it's been quite the adventure.

His main YouTube channel: https://www.youtube.com/c/RandallThomas/videos

His secondary YouTube channel: https://www.youtube.com/c/RandytheSequel/videos

Of course, he still has nothing close to an actual game and the progress he's made in a year is visually similar to what i threw together in a month in Unity for a freelance client once. Though admittedly, he's probably learnt a bit more about how game engines work by creating his own and it's been quite the educational experience, as well as something that actually has gained him a nice following of people who support him towards more videos.

But that's still not very conductive to shipping finished games. It's very much the same as people whose games tend to balloon in scope, another thing that's as damaging to finishing a product in a reasonable amount of time as writing your own engine is.

I do suspect that not everyone has the utilitarian approach to game development that i or other people like me have: who just pick Godot/Unity/Unreal for every single project of theirs without ever caring much about the inner workings of everything. Then again, i'm the kind of person for whom, for example, adding graphical effects to their game consists of utilizing pre-made shaders instead of learning to use the shader language properly and all that.

Not that there's anything wrong with that, but those are wildly different mindsets.