I agree. I also don't think one can get very far in that respect without first understanding the computational principles and mechanisms by which the brain can perform ordinary tasks that are extremely difficult for computers (e.g. vision). A purely computational account of how the brain manages to do what it does can allow for useful abstractions that can be applied to other kinds of "implementations" of intelligence. Without looking at the brain, I think it would be incredibly difficult to tease out form first principles the fundamentals of what intelligence really is and how it arises.