you're cherry picking an example that succeeded. Let's say DARPA's investment in ESP doesn't turn out in a few decades. Few would argue that any previous investment into ESP was not worth it. (disclaimer: I'm being paid on a stupid darpa grant now)