Closely related is the adage that great ideas often look like bad ones. I think it's worth thinking about all these nuggets of wisdom in terms of efficient markets:
Given that there are so many entrepreneurs building startups the best sounding ideas are already taken and what's left are the bad and seemingly bad ones. The market-economy is very good at picking low-hanging fruit. If you dropped a $20 bill in Grand Central station chances are it'll be taken in minutes, while if you dropped it in a random alley somewhere, you can probably go back the next day and find it.
So for you to find a good idea that the market hasn't picked up yet it needs to be somehow hidden in plain sight -- a crumbled $20 bill that looks like a piece of garbage.
A similar concept that I've been thinking a lot about in the context of my startup is a "scaling bottleneck". For you to get big and service everybody you have to first service a demographic that is often thought of as a dead-end in terms of making money. Say kids, or a hobbyists in a certain field. I think the Facebook example applies here too: it was hard to see how this toy could scale beyond college students. Worded as a statement about efficient markets: good ideas for startups that will initially service and grow on a potentially non-monetizable demographic will not attract entrepreneurs and therefore will be available for ones that have that insight.