Computer modeling of reasoning tasks is far easier than computer modeling of other cognitive tasks, such as the processing of sense images.
Conclusion (So)
It appears that we understand the analytical capabilities of our minds much better than we understand our senses.
Evidence
Computers can defeat chess champions, but cannot see. Computer modeling of reasoning tasks is way easier than computer modeling of sense images.
Evaluate
The language we're trying to prove in the Conclusion is .
We weren't given a rule by which to judge whether we understand one thing better than another, so the correct answer will have to supply us with a rule for why we would understand one thing better than another.
The reference to "analytical capabilities" in the conclusion seems intended as a callback to "chess" and "reasoning tasks".
The reference to "our senses" seems intended as a callback to "seeing" and "processing of sense images".
Why would our author be thinking we understand that first threesome better than we do that second threesome?
There are two comparative advantages mentioned: - computer modeling of A is far easier than that of B - computers can beat humans at A but can't accomplish B.
Goal
So they've provided us two different launchpads in the Evidence, that can potentially get us to the landing zone in the Conclusion.
If .... Then ... computer modeling we understand of A is much easier A much better than computer than we modeling of B understand B or If .... Then ... computers can we understand beat humans at A A much better but computers than we can't even do B understand B