I think a lot of programmers love the idea of being game developers but don't really have much of an idea for what to make, so they get diverted into developing

I think a lot of programmers love the idea of being game developers but don't really have much of an idea for what to make, so they get diverted into developing a game engine or even tools to support a game engine instead. People can spend years on that. And that's a perfectly fine hobby. But it doesn't usually result in actually making a game. My advice to people going down that path is: consider contributing to ScummVM, an open source roguelike, or a ROM hacking project for a game you already like. Scratch your technical itch on something that other people will actually play. It feels great, and it gives you a better path toward eventually building something that's totally your own. Trying to build your own game engine and game on top of it is like taking a plane to Nepal and just starting to walk up Everest with the intention of figuring it out as you go.