I think you're missing the point. Of course prolog can reason better than humans, that's what it was designed to do (deterministically). The point is not to s

I think you're missing the point. Of course prolog can reason better than humans, that's what it was designed to do (deterministically). The point is not to solve reasoning. The question is, can LLMs reason? LLMs were not designed to reason, reasoning in an LLM is emergent. That should be interesting. It should also be exciting because the domain over which LLMs can reason is much more unbounded than the domain over which prolog can reason (tokens and relationships you've already supplied it)