Great work!
I wonder if it will finally be possible to explore the behavior of general nearest-neighbor rules on such surfaces. Definition of such rules would require enumeration schemes for the grid. And maybe also the rules will become more complicated, for example, if the number of adjacent cells varies.
You may also have a look at this post: https://community.wolfram.com/groups/-/m/t/2027803
It explores the behavior of totalistic cellular automata on 3D Lattices.
