This marriage of logic with statistical, ML based, LLM is highly non-trivial (if at all possible) and if it happens it will most likely be called another major "AI revolution".

In fact the current AI mainstream considers this as a research direction to avoid at all costs (they term it "the bitter lesson"). The strategy and rallying cry is roughly translated: you can achieve gee-wheeze results here and now by ignoring these deep problems that stymied generations of AI researchers.