[layer8]

The problem I'm trying to hint at is that we are unable to specify what the distinguishing properties of "experience" are supposed to be. It always ends up being self-referential. So maybe it is just a self-referential thing, a system perceiving parts of itself (mixed in with other things it perceives). Perceiving perception as such is trivial, e.g. I can have a system monitoring its own image processing pipeline, or I could feed information about the inner intermediate states of an LLM's neural network back into the LLM as part of its token input. The question is, how is "experiencing perception" any different? I'm not sure it is. The fact that I name part of my inner perceptions "feelings" or "qualia" or "texture", is largely immaterial to the question. They are objects of perception just the same, just originating from the inside rather than from the outside. I also disagree about the color example. I don't see how the experience of what a color feels like, or any insight related to that, could not be described mathematically in detail. Maybe it takes a bazillion of synaptic states that would need to be captured in that description, but that doesn't pose any conceptual obstacle.