我成长于20世纪70年代，当时“职位阶梯（corporate ladder）”的概念十分风行。通常一个人的标准计划就是：上一所名牌大学，接着进入某家公司，然后得到提升，承担更大的责任。雄心勃勃只不过意味着更快的升职。 
 Lecuyer, Christophe, 《制造硅谷》 MIT出版社, 2006.
For nearly all of history the success of a society was proportionate to its ability to assemble large and disciplined organizations. Those who bet on economies of scale generally won, which meant the largest organizations were the most successful ones.
Things have already changed so much that this is hard for us to believe, but till just a few decades ago the largest organizations tended to be the most progressive. An ambitious kid graduating from college in 1960 wanted to work in the huge, gleaming offices of Ford, or General Electric, or NASA. Small meant small-time. Small in 1960 didn't mean a cool little startup. It meant uncle Sid's shoe store.
When I grew up in the 1970s, the idea of the "corporate ladder" was still very much alive. The standard plan was to try to get into a good college, from which one would be drafted into some organization and then rise to positions of gradually increasing responsibility. The more ambitious merely hoped to climb the same ladder faster. 
But in the late twentieth century something changed. It turned out that economies of scale were not the only force at work. Particularly in technology, the increase in speed one could get from smaller groups started to trump the advantages of size.
The future turned out to be different from the one we were expecting in 1970. The domed cities and flying cars we expected have failed to materialize. But fortunately so have the jumpsuits with badges indicating our specialty and rank. Instead of being dominated by a few, giant tree-structured organizations, it's now looking like the economy of the future will be a fluid network of smaller, independent units.
It's not so much that large organizations stopped working. There's no evidence that famously successful organizations like the Roman army or the British East India Company were any less afflicted by protocol and politics than organizations of the same size today. But they were competing against opponents who couldn't change the rules on the fly by discovering new technology. Now it turns out the rule "large and disciplined organizations win" needs to have a qualification appended: "at games that change slowly." No one knew till change reached a sufficient speed.
Large organizations will start to do worse now, though, because for the first time in history they're no longer getting the best people. An ambitious kid graduating from college now doesn't want to work for a big company. They want to work for the hot startup that's rapidly growing into one. If they're really ambitious, they want to start it. 
This doesn't mean big companies will disappear. To say that startups will succeed implies that big companies will exist, because startups that succeed either become big companies or are acquired by them.  But large organizations will probably never again play the leading role they did up till the last quarter of the twentieth century.
It's kind of surprising that a trend that lasted so long would ever run out. How often does it happen that a rule works for thousands of years, then switches polarity?
The millennia-long run of bigger-is-better left us with a lot of traditions that are now obsolete, but extremely deeply rooted. Which means the ambitious can now do arbitrage on them. It will be very valuable to understand precisely which ideas to keep and which can now be discarded.
The place to look is where the spread of smallness began: in the world of startups.
There have always been occasional cases, particularly in the US, of ambitious people who grew the ladder under them instead of climbing it. But till recently this was an anomalous route that tended to be followed only by outsiders. It was no coincidence that the great industrialists of the nineteenth century had so little formal education. As huge as their companies eventually became, they were all essentially mechanics and shopkeepers at first. That was a social step no one with a college education would take if they could avoid it. Till the rise of technology startups, and in particular, Internet startups, it was very unusual for educated people to start their own businesses.
The eight men who left Shockley Semiconductor to found Fairchild Semiconductor, the original Silicon Valley startup, weren't even trying to start a company at first. They were just looking for a company willing to hire them as a group. Then one of their parents introduced them to a small investment bank that offered to find funding for them to start their own, so they did. But starting a company was an alien idea to them; it was something they backed into. 
Now I would guess that practically every Stanford or Berkeley undergrad who knows how to program has at least considered the idea of starting a startup. East Coast universities are not far behind, and British universities only a little behind them. This pattern suggests that attitudes at Stanford and Berkeley are not an anomaly, but a leading indicator. This is the way the world is going.
Of course, Internet startups are still only a fraction of the world's economy. Could a trend based on them be that powerful?
I think so. There's no reason to suppose there's any limit to the amount of work that could be done in this area. Like science, wealth seems to expand fractally. Steam power was a sliver of the British economy when Watt started working on it. But his work led to more work till that sliver had expanded into something bigger than the whole economy of which it had initially been a part.
The same thing could happen with the Internet. If Internet startups offer the best opportunity for ambitious people, then a lot of ambitious people will start them, and this bit of the economy will balloon in the usual fractal way.
Even if Internet-related applications only become a tenth of the world's economy, this component will set the tone for the rest. The most dynamic part of the economy always does, in everything from salaries to standards of dress. Not just because of its prestige, but because the principles underlying the most dynamic part of the economy tend to be ones that work.
For the future, the trend to bet on seems to be networks of small, autonomous groups whose performance is measured individually. And the societies that win will be the ones with the least impedance.
As with the original industrial revolution, some societies are going to be better at this than others. Within a generation of its birth in England, the Industrial Revolution had spread to continental Europe and North America. But it didn't spread everywhere. This new way of doing things could only take root in places that were prepared for it. It could only spread to places that already had a vigorous middle class.
There is a similar social component to the transformation that began in Silicon Valley in the 1960s. Two new kinds of techniques were developed there: techniques for building integrated circuits, and techniques for building a new type of company designed to grow fast by creating new technology. The techniques for building integrated circuits spread rapidly to other countries. But the techniques for building startups didn't. Fifty years later, startups are ubiquitous in Silicon Valley and common in a handful of other US cities, but they're still an anomaly in most of the world.
Part of the reason—possibly the main reason—that startups have not spread as broadly as the Industrial Revolution did is their social disruptiveness. Though it brought many social changes, the Industrial Revolution was not fighting the principle that bigger is better. Quite the opposite: the two dovetailed beautifully. The new industrial companies adapted the customs of existing large organizations like the military and the civil service, and the resulting hybrid worked well. "Captains of industry" issued orders to "armies of workers," and everyone knew what they were supposed to do.
Startups seem to go more against the grain, socially. It's hard for them to flourish in societies that value hierarchy and stability, just as it was hard for industrialization to flourish in societies ruled by people who stole at will from the merchant class. But there were already a handful of countries past that stage when the Industrial Revolution happened. There do not seem to be that many ready this time.
 One of the bizarre consequences of this model was that the usual way to make more money was to become a manager. This is one of the things startups fix.
 There are a lot of reasons American car companies have been doing so much worse than Japanese car companies, but at least one of them is a cause for optimism: American graduates have more options.
 It's possible that companies will one day be able to grow big in revenues without growing big in people, but we are not very far along that trend yet.
 Lecuyer, Christophe, Making Silicon Valley, MIT Press, 2006.
Thanks to Trevor Blackwell, Paul Buchheit, Jessica Livingston, and Robert Morris for reading drafts of this