so a curved world is inferred from mathematics between 2 similar triangular bodies in a 3d space. lining up simplified curvature with physics can be efficient. Using strong pretraining and on the fly AI training modules and on the fly compute. This research venture needs to find all the data it can to help the world but use the optimal use of avoidance and exclusion when searching for an optimal as problem complexity increases.
when calculating a body you can compute particularised rotated cubes of the body in a mixed size sampling relative to a field determining of render points required in the sampling. When viewing for example you can have a mix of load done on the workstation and cloud AI leading to mixed sized square pixels then image upscaled and frame upscaled although the AI may be doing more efficiently focused calculation to derive the view you are having of the earth 2 simulation relative to other parts of the simulation.