Question: This guy thinks the Cayley-Dickson Construction cycles back at 1024D? I don't actually use Maple, is he just encountering a precision error?

This guy thinks the Cayley-Dickson Construction cycles back at 1024D? I don't actually use Maple, is he just encountering a precision error?

https://www.mapleprimes.com/posts/124913-Visualization-Of-The-CayleyDickson

quote in question

"

I found no new mathematics after 1024 because higher hypercomplex numbers greater than 1024 are cyclic (they repeat all over again).  I do not offer dimensions higher than 256D to the publc Maple Application Center because the mathematics is very slow and time cosuming past 64D.  However, I did keep the code up to 256D in the public Maple Application Center.

"

I received a program for constructing Cayley-Dickson tables for my own use, but it's actually written in bc (arbitrary precision). I'm assuming Maple supports this to some extent but maybe he's encountering a precision error that looks like it's repeating? I can't verify yet because he's right about the one thing, even a 256 table took DAYS to compute. The snapshots are intriguing too but I'm assuming even a second-long video at low resolution would take forever as well.

But I still have the bc code and constructing Cayley-Dickson tables is only a few lines of codes/conditions, it's one of those 'easy for a computer, impossible for a human' kind of deals. I don't see enough complexity in the code where it would suddenly start cycling. It's kind of important to me because the implications of the Cayley-Dickson Construction going on forever are more exciting in my opinion.

By the way I'm actually personally using the bc generated tables for making music sequences/MIDI. It's not the usual multiplication and such though because that even when normalized would "stick" so to speak? It's a bit of a secret, sorry