Bubbles are scary when they fill one criteria - investment in the bottom of the pyramid is feeding the top, making the whole thing look more profitable than it is.
For example, in the tech bubble, the new companies were buying Sun servers, Oracle databases, and Yahoo! ads. This made Sun, Oracle, and Yahoo very profitable, which encouraged more tech investment. But that snake can't eat itself forever, and when investment slowed it all went bust.
Likewise, housing bubbles affect (effect?) so much of the economy they can't help but feed off themselves.
And in a classic Ponzi scheme, the mechanism is the same - the bottom feeds the top, so people think that there are outsize gains to be made. Those gains evaporate once new entrants slow.
Now, there's some new startups like Heroku who are "selling picks in a gold rush". But there's plenty of others like Groupon and AirBnB who are selling to the wider economy. The successful internet start-ups don't usually feed off new entrants, so I don't think a really scary bubble can form.
That said, if a bunch of random airheads (who are these 60+ new incubators? top tier VC? bottom tier VC? government? universities? there will be winners and losers here) think they can attract the same talent as Paul Graham, and vet business and technology plans as successfully, they may stand to lose their shirts.