The view from Silicon Valley, and its infamous advertising sock puppet, are the poster children of the bust. People in Silicon Valley still snicker about the $400m failure whose demise infamously landed its chief backer on a magazine cover with the legend: "World's Worst Venture Capitalist?"

But the other day, I remembered what a marvel had been. It had gone from an idea to a 400-person, fully functioning company in less than a year. Logistics alone were a daunting task, especially in the overblown markets for people, space and technological infrastructure. In a way that makes sense only in the rarefied, topsy-turvy world of venture capital, where failure is the norm, did what it was supposed to do. It discovered that animal fanciers didn't think petfood-by-mail was a good idea. So went belly up.

But the idea had merit. Sales of pet items are huge in the US. Giant discount stores and mail-order services slug it out to attract the custom of tens of millions of pet-owners. A website that gave free advice from experts (instead of the clueless, underpaid clerks in most discount stores), a wide selection, low prices and doorstep delivery really wasn't that bad an idea. But it failed. Miserably. Spectacularly. Why?

My theory is that it is nearly impossible to build a winner from scratch all in one go. Evolution or, more precisely, the Darwinian theory of evolution provides the clue. When a new niche opens up in a natural system, a large number of critters may attempt to occupy it. Over time, however, it a single species will nearly always come to dominate the niche, with perhaps a small number of rivals that manage to hang on. It can be demonstrated that the winner isn't necessarily the best equipped; but it is the one that has endured and adapted to all the twists and turns the world has launched at it over thousands or millions of years, some of which have nothing to do with its core abilities.

Thus, small, scurrying mammals ultimately trumped huge and formidable dinosaurs. Successful start-ups have likewise scraped and clawed their way from inception to real businesses. They often succeed in ways not envisioned by their founders. Intel once made memory chips and almost crashed; Hewlett-Packard started out (with $583 in capital) making audio oscillators (it sold eight, to Disney, for the film Fantasia) long before practical computers were invented.

It seems that some principle says you have to put in the hard time. Real companies have done the excruciating and gritty work of adapting themselves to what real people will actually pay for. And that takes time, patience and an unrelenting effort.

If nature could have built an ultimate winner in one shot, it would have: it's the nature of nature to be efficient. But where nature cannot go, it seems, neither can nature's spawn, including venture capitalists. So adieu,, dodo of dot.coms. We'll wait to see which companies are really the mammals among the feet of today's dinosaurs.