Mathematics vs semantics: Not seeing the obvious

This is a research in LLM weakness in order to deliver accurate phenomenological meaning to the untrained personnel in reductionistic framings. Reductionistic framings are highly appropriate for trained personnel. But reductionist framings is highly inappropriate for the settings outside of the scientific high community. In social environments the output distributions have to be in less e.g. phenomenological language that supports users rather than sabotages them with their scientific push. We argue that strict communication in reductionistic language produces reductionistic parroting which cannot be escaped by thinking on the same stochastic level. We see reductionistic language as a self reinforcing echo chamber that produces meaning that is purely mathematical but not semantic in nature thus there is a huge category error in what we expected to receive and what we have received. If we want a general purpose tool we need it to produce general purpose meaning. Getting mathematical meaning as a result is too narrow even for AI interpretation as it is not clear enough what the objects of interpretation are. The AI results cannot be purely explained with pure mathematics as they are not fully produced from purely mathematical artefacts. In AI results generation we are intertwining two realities, linguistic and mathematical in first.