[krapp]

99.99...(some arbitrary significant digit)% of "game ideas" are perfectly capable of being represented in Unity or most modern game engines. For the rest, it's easier to change your concept rather than put off shipping for several years to build a competitive engine just to scratch your specific programmer itches.

Plenty of people do start their own custom engines. The problem is, a general purpose game engine of any relevant quality is comparable in complexity to a modern web browser, most game developers don't have the time or talent to handle that, and once they get past a gee-whiz renderer and maybe loading textures, they wind up in the weeds and never ship anything again.

Don't get me wrong - I've done it. I've got the corpses of several "game engines" including the ersatz one I'm building around a roguelike that I haven't and may never finish, but it isn't a satisfactory general solution for most people.

[popcorncowboy]

The old "there are two types of game devs, those who make games and those who use games to write engines/toolsets". Those solo game devs on Youtube building their magnum opus are inevitably the latter. It's a different kind of fun, but yeah, don't expect to ship anything.

[thaanpaa]

Yes, a one-size-fits-all game engine will undoubtedly be extremely complicated. However, depending on your project, you will most likely only require a subset of its features.

One could also argue that if people base their creative decisions on the game engine, that's not always a good thing.

[ffgjgf1]

> For the rest, it's easier to change your concept rather than put off shipping for several

Plenty of games still have their own custom engines.

Also you have games like Rimworld which pretty much only use Unity for rendering and cross platform support.