> Another tip that I'd add is that you should increase the science cost, I found it easy to run out of things to research while it still took hours to expand things for the next tier, which in practice meant your factory just stops doing anything because it doesn't need to do anything.

That's not a wrong way to play, but the intent of the default balancing is that you make those expansions smaller and faster. If a new resources has 5 steps to make, you can slap down 2-5 buildings per step with absolutely minimal planning/routing and feed it into a launcher. There's no need to spend hours on an expansion until your science production needs that much. And an orbital base can go very far with one assembler per recipe.