Of course, computers are a terrible analogy to a conscious biological system. One is predictable, and scale-segregated (we separate noise from signal to fit the computer's operations to abstract human definitions with logic gates) and it waits for instructions to do anythingAppeals to mental events are not only not required, they are superfluous to the transistor's function and by extension, to the computer's function.
The other is spontaneous and irregular. It's behavior follows exponentially diverging trajectories (i.e. it's chaotic) when compared to a minimally perturbed clone. It wasn't designed, but emerged from nature, in the wake of several different uncorrelated perturbations. It requires several parallel redundancies to be built throughout the system for it to persist in the first place.
If the system is to correlate particularly relevant information (through synchronicity, for instance as per the Varela paper) than we can reason why cognition may have a functional component (though we can agree that cognition is not important to immediate survival, it's function is geared towards long-term survival).
We can't know the computer had a phenomenal experience because we can understand everything about what it does by understanding the circuitry and the physical states and inputs.
For a computer, I agree. But with biological systems, particularly humans, we have the special treat of having the experience of consciousness and we have developed language to communicate about it. From birth, we can read each others facial expressions and body language (and even that of other mammals). This is only possible because there is a consistent relationship between the kind of stress on an organism and the muscle groups associated with them.
The muscle groups are correlated by interneurons that take central pattern generators (CPGs) and inputs (that either interact with the CPG or the motor pool itself, or booth). The CPG is something that developed over an evolutionary history, while the inputs are representative of the current moment for the organism. The interneurons allow input patterns to be associated with meaningful outputs.
Now, with all the knowledge of functional anatomy and mere "circuitry" (circuitry is, of course, and oversimplification) I can affect mental states in predictable ways by making physical alterations. The more I know about the receptor diversity of a particular physical circuit (and given the appropriate drugs) the more precisely I can target only the receptor variations that participate in a particular functional effect I want you to experience.
Furthermore, I can target particular experiences you don't want to feel and remove them from your experiences without removing the kinds of experiences you'd like to remain?
Which is why this is incorrect:
Knowing how and why every observable molecule in the brain does what it does says nothing about our subjective experience and never will because explaining interactions are the wrong kind of explanations to look for when explaining subjective phenomena.