Deep Learning, AI and NASDAQ?

I'll clean up, but as it's more philosophy than an academic exercise and use of technical vocabulary, it's not straight forward. Even so, here is some text: Given progress and results in deep learning systems pushed forward with ontology from human produced semantic structures on top of attention and memory guided by an invisible hand out of blue sky thinking. In many ways this is the equivalent of lensing structure and attention in human knowledge by blacking out half of a semantically coherent process. The amount of resources used to create the equivalent of a epistemological forward process taking advantage of natural inter-processing within understanding of semantics and logic. Gives concern about the feasibility of the next evolution namely feed forward ontology and epistemology in the other direction. Needing a feed forward identity process able to creatively deal with the intra-processes in-between these functions: memory and logic, attention and semantics; and a process able to navigate with ensuing identity processees. A information super highway much like the corpus callosum -- the front pole -- back to sensory motor as seen in human cognition and brain structure with analogous cognitive functions. In Norse philosophy, what it seems like -- too much faith is put into half a horse -- loki without idun. Thus, an intelligence that is a headless horse guided by a democracy and cannot be trusted, it's just annoying feed-forward logic with little to no command control over attention and memory. Overly optimistic enthusiasm coupled with polarising fear may have set in motion a need for an AI winter. Can forget building anything that can solve problems for life on earth. Perhaps more efforts adding an Idun to help with neutral and intelligent information processing is a more correct way.