There is natural demand growth for start-ups due to these factors, which manifest themselves as new acquirers springing up - as an example Google wasn't a factor in the last bubble as an acquirer, and may be Facebook will be very acquisitive soon.
So my question can be rephrased as: are we outrunning that demand growth in terms of supply of start-ups? It seems to be so sometimes. And in a recession demand growth could temporarily reverse too ...
Economies of scale are based on transaction costs versus measurement costs ( http://en.wikipedia.org/wiki/Theory_of_the_firm ). Transaction costs have declined dramatically with improvements in IT infrastructure, but measurement costs have changed relatively little. In fact, measurement costs in knowledge industries are very high. Thus I expect startups will continue to become more popular.
It is true that specific technologies saturate. Browser based AJAX may run its course in a few years, but there will be new platforms to replace it.
Wealth is value added for people or companies. If you create a web-app that saves a company money, then you've created wealth. The demand that companies have for saving money is unlimited. The limits are on the supply side: Can you think of a web app that saves a company money?
On the consumer side, it's basically the same. If you can save customers time, or money, if you can entertain them, or educate them, or make their lives better in any way, then you've created wealth.
Again, the demand for a better living through web-apps is unlimited. The limits are on our side - how much can we deliver?
If you have really created wealth, step two is how to monetize it. That's where Google and Yahoo come in. But that's another issue altogether.
Economists and sociologists alike have long preached to the choir that too much of one thing in a society can be bad for the big picture. Take for example the problem with American Obeisity. The more buffet items a restaurant displays, the more a person is likely to eat while negatively effecting their health overall.
Make the parallel to these startups, social content sites are now a dime a dozen, feed aggregators are everywhere (so much so that there are even aggregators that aggregate other aggregators) and consumers find themselves overwhelmed with the choices they have to make when selecting the simplest problem they need to address. When this happens, people often opt not to do anything and would prefer a direct fix through another means than to go and experiment with what's new on the market. This is essentially why you see very few startups aimed to address corporate interests where obviously the potential to make money lies, but instead targets a much smaller sector of the market (37Signals for example).
All of that said, prices and income potential for the startup in my opinion is coming very close to reaching a stalemate, while supplies continue to grow, demand is going to drop.
Before someone jumps in and points out there has always been death of start-ups, the YC companies are quite good, and quite plenty (with even more coming ...) which is what raised the supply-demand concern.
1) The cost of starting Apple today would be much lower than it was two decades ago. However, the cost and complexity of starting a computer company that is truly groundbreaking by today's standards is just as high, if not higher.
2) The cost remains constant, but the price of startups, as set by investors, is actually getting higher. Anyone could make a Facebook in 1998 at approximately the same cost (even though servers were more expensive), but no one would have valued it at $15bn. Pricing would either go back down or the change in economic thinking (i.e., advertiseming is more important than hard cash sales) would become the standard.