In any case, I would be careful contrasting naturalism to supernaturalism and then saying there's no plausible evidence for the latter. Non-naturalists do not necessarily believe in supernatural events with no natural explanations--the question is whether those natural explanations can be fully grounded in physical reality or entail a non-physical "absolute" reality. I would put any radical Platonism squarely in the non-naturalist camp.
That's why I suggested it as a precautionary principle, rather than a fundamental principle; i.e. within a chosen model or account, "Everything should be made as simple as possible, but not simpler"
You guys are arguing from different axiomatic presuppositions. It's not surprising that your arguments may be flying past each other if we don't specify and clarify the differences between initial assumptions. I don't want to be speaking for either of you, but it helpful to juxtapose the initial assumptions (axioms) and see where the disconnect resides.
Silmarien: Whatever we perceive as reality is first structured by our conscious mind. Therefore that structure is primary to the secondary "material reality out there", which may not exist, at least not in the way that we perceive it. Therefore one can't jump to the next assumption (material reality exists as we perceive it) and then circularly claim that matter is fundamental, and consciousness is the product of the causal interactions between matter. Likewise, it's very presumptuous to then claim that only material exists, and build false dichotomies about lack of evidence for immaterial.
In short, consciousness informs us of reality as a structure, it doesn't necessarily tells us anything beyond our conscious experience. Matter (as we perceive it) may not even exist, or likely doesn't exist. What we observe about "fundamental" from science is that it's not at all like we perceive it. Invoking supernatural as a dialectic of natural can't be made unless we describe the "nature" of natural beyond the structure that's given to natural by our conscious perception.
Therefore, the interaction problem is a non-problem. Supernatural vs natural is a potentially false dichotomy, driven by axiomatic assumption that material is natural. (Projecting) Describing reality as a collection of particles which are responsible for observable attributes of matter... is not "As simple as possible", since there are actually more direct means to simply describe ratios and attributes, which are more useful than claims about models that are being reified by each subsequent generation of physicists.
FrumiousBandersnatch: (Also, forgive for approximation) We don't really have to understand the "true nature of reality" in order to systematically approach what we do perceive as our reality. What we don't perceive as reality seems to be of secondary importance in terms of how our own being is structured, nether it seems to be ontologically consequential. It's the age-old "if the tree falls in the woods..." concept. How would you know that it fell unless you can experience and describe it? After all, we don't first observe the nature of "unseen ultimate reality". We observe reality as we observe it, and then we distill that observation into systematic relationships between observable.
Likewise, we are "we", and have to deal with other minds that we communicate with and compare our perception. The problem is then developing a system that allows us to separate false, or less likely claims from more likely or true. There seems to be no better way than existing methodology of scientific naturalism. It doesn't mean anything that hasn't been consistently observed by this methodological framework doesn't exist. It simply means that unexaminable concepts are not of much value in context of that framework.
Scientific models are merely causal models and not ontological. These are helpful to describe reality by drawing on "schematic analogies" of our level of perception that we can relate to. These are useful as systematic description. Thus, axiomatic assumptions of materialism are useful when we aligning the language of the models to our assumptions about reality.
My View: I don't think that it's the "either - or" type of dichotomy, and that we can speak of these contextually, depending on the questions that we are trying to answer.
Platonism as per Silmarien, and methodological naturalism as per Bandersnatch are answering two different questions and attempting to solve different problems.
If we think of a computer RPG, we can describe the rules of the game as the players play these in the game, or we can describe these in context of perceived author's intent, and imagined concept of the ultimate reality outside of the game. The source code could be hidden from both.
I'm not convinced that the latter is accessible to us beyond faith-based claims one would have to make bets on. I'm not sure that it's relevant. I actually view Christianity not as means to describing ontology, but rather a means to communicate human reality via historically-observed archetypes. It tells us nothing about ontology apart from claims of what God is "most like" when we look in our world. God is an ideal in context of our reality.
So, ultimately we are not pointing "upward" when we are pointing to what God is. We are pointing at human concepts and then say "It's like that, but much much better". Therefore, I'm sure that it's useful to either negate or unify these views. Let science be science, and religion be religion. Both are useful in context of their usefulness. Both idealism and realism are likewise useful in context of each ontological focus.
In terms of this thread, I'm not even sure how you can make either claim about consciousness without describing what consciousness is and how it works.
In terms of the AI, and the whole concept of the Turing test... is that if no one can tell a difference, then what would be the difference?