Hi there! I attended the Siggraph GA talk and I see that this forum is empty, so I hope that ignorant questions are welcome!

I’m interested in geometric algebra for ray tracing, so I thought I’d start a topic for it. As a first prompt: One of the problems for ray tracing is intersection with sharp triangles. When these triangles aren’t axis-aligned, the bounding box for them is much larger than they are, so raycasts will encounters many false positives.

If we use geometric algebra, is there some neat trick we can pull to avoid this? The purpose of AABBs is to define regions of space that can be nested hierarchically that are easy to test for intersection. Since calculating the intersection of two things is done differently in GA, can we use non axis aligned bounding boxes?