Ask HN: What's your litmus test for detecting a tech bubble?

I'm looking for very falsifiable predictions. Ideally, people reading this thread a few years in the future should be able to say conclusively whether your litmus test was triggered: "If venture capital investment in tech startups exceeds X% of United States GDP by the end of 2012, that would constitute an investment bubble."

If there's a big enough response, I'll do something neat with the data in R.

I realize this question isn't easy, and there may be good reasons why it's not possible (or just not worthwhile) to give exact figures. Nonetheless, take a stab at it. For science.