Okay, notice I'm saying "great business", not "startup." There are millions of "startups" working on cool technical challenges, of which almost none will become great businesses.
When all those things you described were happening, almost nobody cared = no business. When they turned to meet needs of a much broader audience then the few technical people who were interested in such a narrow range of issues, they took off as businesses. In my mind, "startup" does not equal a business. Business means lots of people pay you money for selling them your thing. So to a limited extent, for example, Woz had a "business" selling computers through ads to computer enthusiasts in the back fo Scientific American or whatever. But that was nowhere near as successful a business as when Jobs visited PARC and realized the research they were doing on GUIs based on the human mind could be significant for a lot of people.
The Yahoo folks may have started with a list specific to their interests, and students may have found it cool, but that was not a business. It became a business when they broadened to general idea of an organized list to the needs of a larger audience.
I've come across quite a few programmers over the last year who turned me down because they said they were busy working on their own startups, all of which were based on solving a problem only they seemed to care about. They incorporated, had a "company", etc.
But that doesn't equal business. Business means people pay you money. For that to happen you have to broaden your considerations to what those people want, not what you care to/can build. I won't, nor will anybody pay any of your startups a penny because somebody wrote really cool code that everyone on Hacker News agrees is awesome. But I'll pay you money if your code solves my needs. You don't care to solve my needs? Fine. You may have a technically awesome thing. I won't pay you for it, though. And neither will anyone else.