Essay Abstract
Artificial intelligence can already learn. Algorithmic foundations and structure to describe how interactions produce 'aims and 'intentions' is identified by exposing an unused momenta hiding before our eyes in a spinning sphere (OAM). Modelling this simplest of mechanisms as a fermion gives a classic derivation of the Dirac twin stacked inverse complementary pairs in quantum mechanics, exactly as anticipated by John Bell. The recursion of neural theories is reduced to fractals, however a computer the size of a brain may be needed to run the algorithm to get good approximations. Ironically limited primeval evolution of neural mechanisms can explain why it's own workings remain a mystery. Judgements using 'default' pattern matching rather than more complex rational analysis will tend to constantly re-embed older doctrine and reject anything new so hamper advancement of understanding. We identify that conscious 'self evolution' is required, using a non-linear 'layered' architecture already proven in logic, some human brains and in 'deep thinking' photonic AI. How such evolution may be achieved is informed by the new classic quantum mechanism, allowing a small probability of any DNA key switching on replication. No decision on existence of any cosmic architect can be reached.
Author Bio
Born 1951. Studied multiple Sciences then paralleled research with Philosophy and Architecture degrees. UKC, UCA and Westminster. Perpetual student! Royal Astronomical Society Fellow in Observational Cosmology. Worked in Energy, Renewables & Lead Consultant on major Pharmaceutical, Petrochemical, Energy and Defence projects. Visiting student Mentor at Kent University and UCA. U.K. representative yachtsman, Royal Y.C. Flag Officer. Rugby player & club chairman. Now semi retired but continuing full time research, mainly on unification and TOE's.