You're right about the definition of thinking. I shouldn't have used the word "think".

I should amend my earlier statement. Replace instances of "thinking" in my statement with "computation". Qualia and survival behaviors due to computation are two circles which form a Venn diagram. Humans are in the center of it. This is what I meant to say.

Here is my argument for why we shouldn't treat computation and qualia as the same: computation can still result in survival, regardless of whether the entity actually feels anything or not.

It is reasonable to extrapolate current trends to hypothesize a robot which survives in the environment as well as any thinking human. But I would hesitate to say they feel anything. Does it feel like anything to be a robot? It's just performing boolean operations. Even rows of dominos can be arranged perform boolean operations. A long enough chain of self-righting dominos can also do sophisticated computations (very very slowly). But I wouldn't grant dominos the status of feeling, that would be preposterous. If you don't like the dominos example, just replace with a mechanical turing complete system of your choice. It seems to me then that intelligent computation (which can be used for survival) and feeling (qualia) are two different matters.

However, it is also possible that computation and qualia are never separate, even outside biological brains. In that case, pansychism would be true. But how could we know? For now, we can't, and we may never know.

Tl,dr: I think you are incorrectly assuming survival computations and ability to feel a subjective experience are both only present together, in a person or animal. I'm suggesting there is also a possibility they can exist separately. There is also a third possibility, that all computation (particle interactions) in the universe coincides with qualia, which is a form of panpsychism. Panpsychism doesn't require everything has "mind", it can just be the "mindlike aspect" of qualia.