I was with you up to this point--mostly--but this doesn't follow at all! This only holds if you have an ontological precommitment to an inscrutable universe.

(Note for reference: this is something I've also thought about a lot, and read about a fair bit, so you're unlikely to immediately raise any points that I've failed to consider. Since you also claim to be in a similar situation, we'll probably have to go around a few times before we get to anything new.)

It is true that what we know already necessarily comes prior to that which we figure out later. And our direct sensation is something we have access to extremely early on. But part of ordinary mental development of children is to use those priors to build superior "secondary" models that fix deficits in primary sensation. For instance, in Piaget's (not always very well-grounded) model of child development (see e.g. https://www.verywellmind.com/piagets-stages-of-cognitive-development-2795457), the "preoperational" stage is where children do not understand that if you pour a liquid between a wide glass and a tall glass, that there is the same amount of liquid. Direct primary knowledge tells them that the tall glass has more!

You're essentially arguing for us to stay stuck at preoperational thinking. Maybe sensorimotor thinking!

Just like with our apprehension of physical phenomena, where we understand that we can understand more deeply than as-things-first-seem (and indeed, we usually can train ourselves to adopt the superior outlook so instinctively that we forget that it wasn't really as-it-seemed), we can also understand our values and motives this way. We recognize, as I mentioned in my previous reply, that the happiness button is not a good thing, even though happiness is usually a good thing. There is no predefined limit to how deeply we can inspect these things--we simply can do it, and take it as far as the evidence leads, as with everything else.

Of course, you can't come up with a system that denies that we feel things. That is ridiculously counterfactual--it's like denying that glass is hard because you can't really see it that well. But you absolutely can question whether it's foundational to what we should do (i.e. it is the ultimate source of "shouldness"), or whether we can explain it some other way (e.g. historical adaptive value + random chance) which gives us additional options.

What I argued before is that we are now actually quite sure that it is explained some other way, and this has nontrivial implications for the foundations of morality and goal-directed behavior. Obviously we need to hold our conclusions tentatively, as with everything, but this is looking like a pretty sure bet.

You also mention that consciousness is presumably physically efficacious. Indeed! But that we have conscious perception of goodness and badness says nothing about it being a property of any part of the universe save our own minds. And over the past four hundred years we've amassed a very impressive body of evidence consistent with that being exactly the case (and basically nothing contrary): it's in minds, not more broadly intrinsic.

So this doesn't change our analysis at all. Yes, it's in us. But we no more have to slavishly obey it without question than must we slavishly conclude that there is more water when it is in the tall glass, and pouring it to the wide glass makes there be less water.