The engine doesn't matter too much.
The problems and how you solve them are usually sort of the same, in terms of 2d and 3d. Meaning:
If you're doing a 2d platformer or a 3d platformer, you will still need to thing about colliding with the floor and the solution will be similar too. For art, it's the same. if you start from scratch, learning 2d art and learning 3d art is both new to you. But you will have animations in both and engines usually just use some "playanimation" function. So 2d vs. 3d is completely up to you and what you want to make.
I can recommend pygame and panda3d if you like python. They're not new engines and they're "code only". Other engines give you an editor, but I prefer the "code only" approach, because there are no hidden settings that you have to search for in the GUI, it's just code.
But the other engines like godot and unity have a ton of learning material too, so that's probably easier to get started with.
We can't replace it fully.
We can replace it with cars. We can replace it with trains as well, but electrified track is more expensive than just plopping a diesel engine there and filling her up. Track for that is just steel+concrete and rocks and stuff.
We can not replace it with air planes, helicopters, rockets. At all. We could reduce air travel and stuff like fighter jets.
We can also not replace it for cargo ships. And that's pretty bad news. Luckily ships are crazy efficient, so the actual CO2 and other pollution per ton and kilometer is very very low. If you get a delivery, that delivery comes in a fossil fuel truck to your doorstep, that truck will emit more CO2 than the ship will, going either from china to Rotterdam or the US westcoast. And also global transportation is probably more than necessary.
Anyway, the big problem we can solve are cars and planes.
There are also a bunch of chemical and industrial processes that need coal. Fertilizer and steel are two big ones.