So the AI
So you have strings determined with up to 7 360 degree dimensional rotation vertices all determining the string at the end intercept with curve modularity manipulation along the string path for each intercept. So full 7 dimensional curve and dynamic manipulation of a line to make a string.
along the string is also focus determination information for a dynamic inference along strings. So the focusing along the strings are like neural activity and the inference between strings is like thought. So when the AI thinks it develops a lose search on the strings and strings together a thought. how it is all self stitching and how can search itself and can be linked to evolving functionality would take quite a lot of work but it is an efficient reduced analogy of neural intelligence.
So what should this AI focus it's time on.
up to 7 dimensional conic forms.
Archetypal reduced curve function discovery which is easier to compute but can yield other usually harder to get curved functionality.
Multidimensional Bitwise optimisation and functionality.
Modularity as in how to optimally extend modular form and its part to play in efficient multi fractal and algorithmic form.
Developing a more functional model of quantum space time by correcting and extending the broken attempt model by use of constructor theory.
Engineering and matching quantum algorithms with quantum nano computational structures
Engineering more specific and some more reduced sibling AI models for a huge diverse range of tasks.
Developing a good approximation of the visible universe as a model.