I identified a useful and cogent point in your post and it was this: Whenever you receive data from any source (your brain, your eyes, a drug study, Less Wrong) you've got to be aware of how that data has already been packaged. Taking the data at face value -- for example imagining your brain is actually making a claim about the RGB values of the pixels -- can lead to problems, misconceptions, mistakes.