Slow Motion Video of a Speed Solve of a Rubik's Cube

- pick a way to solve it that requires fewer moves

- make fewer mistakes while working on the approach you picked

- make your moves faster (may require better accuracy in making close to 90 degree turns, or just speeding up your moves)

- spend less time pausing to think about future moves

- when possible (which is rarely) make more moves in parallel. A good cube allows some pipelining, though, in the sense that you can start turning a face before you completely finish turning one orthogonal to it.

To get better, pick the area you think you can improve on the easiest, train it, and iterate.

Once you find your cube keeps falling apart, get a better cube.

And complexity keeps going up, of course. It may go up slower, but I somewhat doubt it. My hunch (based on nothing) would be that god's number goes up slower than the number of cube positions. If so, finding the optimal solution gets harder the larger the cube.

I know nothing about this except what a Rubik's cube looks like, but at first cut I'd say your comment is wrong but on the right track.

Rubik's cubes don't depend on volume at all, but the ratio of increasing the surface area is still an inverse square, so you'd be right if you focused on the ratio between area to linear dimension rather than the ratio of volume to area...