You make a strong point, however; the question as to whether we can or cannot improve the efficiency of mathematical research appears to be an open one. I think that perhaps the real issue is that we don't have a correct reductionist account of mathematics, and thus are not able to see clearly what we are doing when we build our theories. If we had a better road-map, I think that at the very least we could tie mathematics down to level 1/level 2 space so that we could have a better idea as to how we can measure the profitability of various possible lines of inquiry.