Whenever the COST trend of something is being considered, it must be balanced with the RETURN trend. It might cost more to double a transistors capabilities (a

Whenever the COST trend of something is being considered, it must be balanced with the RETURN trend. It might cost more to double a transistors capabilities (as a combination of size, speed, manufacturing reliability, etc.) but as transistors not only improve current applications, but take on whole new areas of usefulness, the returns on investment may actually go up. I am not saying they do or do not. But COSTS in isolation of RETURNS can't be used to conclude anything about the effectiveness of research efforts.