Paul Graham - Essays

What killed them? Four things, I think, all of them occurring simultaneously in the mid 2000s.

The most obvious is Google. There can only be one big man in town, and they're clearly it. Google is the most dangerous company now by far, in both the good and bad senses of the word. Microsoft can at best limp along afterward.

When did Google take the lead? There will be a tendency to push it back to their IPO in August 2004, but they weren't setting the terms of the debate then. I'd say they took the lead in 2005. Gmail was one of the things that put them over the edge. Gmail showed they could do more than search.

Gmail also showed how much you could do with web-based software, if you took advantage of what later came to be called "Ajax." And that was the second cause of Microsoft's death: everyone can see the desktop is over. It now seems inevitable that applications will live on the web—not just email, but everything, right up to Photoshop. Even Microsoft sees that now.

Ironically, Microsoft unintentionally helped create Ajax. The x in Ajax is from the XMLHttpRequest object, which lets the browser communicate with the server in the background while displaying a page. (Originally the only way to communicate with the server was to ask for a new page.) XMLHttpRequest was created by Microsoft in the late 90s because they needed it for Outlook. What they didn't realize was that it would be useful to a lot of other people too—in fact, to anyone who wanted to make web apps work like desktop ones.

The other critical component of Ajax is Javascript, the programming language that runs in the browser. Microsoft saw the danger of Javascript and tried to keep it broken for as long as they could. [1] But eventually the open source world won, by producing Javascript libraries that grew over the brokenness of Explorer the way a tree grows over barbed wire.

The third cause of Microsoft's death was broadband Internet. Anyone who cares can have fast Internet access now. And the bigger the pipe to the server, the less you need the desktop.

The last nail in the coffin came, of all places, from Apple. Thanks to OS X, Apple has come back from the dead in a way that is extremely rare in technology. [2] Their victory is so complete that I'm now surprised when I come across a computer running Windows. Nearly all the people we fund at Y Combinator use Apple laptops. It was the same in the audience at startup school. All the computer people use Macs or Linux now. Windows is for grandmas, like Macs used to be in the 90s. So not only does the desktop no longer matter, no one who cares about computers uses Microsoft's anyway.

And of course Apple has Microsoft on the run in music too, with TV and phones on the way.

I'm glad Microsoft is dead. They were like Nero or Commodus—evil in the way only inherited power can make you. Because remember, the Microsoft monopoly didn't begin with Microsoft. They got it from IBM. The software business was overhung by a monopoly from about the mid-1950s to about 2005. For practically its whole existence, that is. One of the reasons "Web 2.0" has such an air of euphoria about it is the feeling, conscious or not, that this era of monopoly may finally be over.

Of course, as a hacker I can't help thinking about how something broken could be fixed. Is there some way Microsoft could come back? In principle, yes. To see how, envision two things: (a) the amount of cash Microsoft now has on hand, and (b) Larry and Sergey making the rounds of all the search engines ten years ago trying to sell the idea for Google for a million dollars, and being turned down by everyone.

The surprising fact is, brilliant hackers—dangerously brilliant hackers—can be had very cheaply, by the standards of a company as rich as Microsoft. So if they wanted to be a contender again, this is how they could do it:

Buy all the good "Web 2.0" startups. They could get substantially all of them for less than they'd have to pay for Facebook.

Put them all in a building in Silicon Valley, surrounded by lead shielding to protect them from any contact with Redmond.

I feel safe suggesting this, because they'd never do it. Microsoft's biggest weakness is that they still don't realize how much they suck. They still think they can write software in house. Maybe they can, by the standards of the desktop world. But that world ended a few years ago.

I already know what the reaction to this essay will be. Half the readers will say that Microsoft is still an enormously profitable company, and that I should be more careful about drawing conclusions based on what a few people think in our insular little "Web 2.0" bubble. The other half, the younger half, will complain that this is old news.

[1] It doesn't take a conscious effort to make software incompatible. All you have to do is not work too hard at fixing bugs—which, if you're a big company, you produce in copious quantities. The situation is exactly analogous to the writing of bogus literary theorists. Most don't try to be obscure; they just don't make an effort to be clear. It wouldn't pay.

[2] In part because Steve Jobs got pushed out by John Sculley in a way that's rare among technology companies. If Apple's board hadn't made that blunder, they wouldn't have had to bounce back.

There are two different ways people judge you. Sometimes judging you correctly is the end goal. But there's a second much more common type of judgement where it isn't. We tend to regard all judgements of us as the first type. We'd probably be happier if we realized which are and which aren't.

The first type of judgement, the type where judging you is the end goal, include court cases, grades in classes, and most competitions. Such judgements can of course be mistaken, but because the goal is to judge you correctly, there's usually some kind of appeals process. If you feel you've been misjudged, you can protest that you've been treated unfairly.

Nearly all the judgements made on children are of this type, so we get into the habit early in life of thinking that all judgements are.

But in fact there is a second much larger class of judgements where judging you is only a means to something else. These include college admissions, hiring and investment decisions, and of course the judgements made in dating. This kind of judgement is not really about you.

Put yourself in the position of someone selecting players for a national team. Suppose for the sake of simplicity that this is a game with no positions, and that you have to select 20 players. There will be a few stars who clearly should make the team, and many players who clearly shouldn't. The only place your judgement makes a difference is in the borderline cases. Suppose you screw up and underestimate the 20th best player, causing him not to make the team, and his place to be taken by the 21st best. You've still picked a good team. If the players have the usual distribution of ability, the 21st best player will be only slightly worse than the 20th best. Probably the difference between them will be less than the measurement error.

The 20th best player may feel he has been misjudged. But your goal here wasn't to provide a service estimating people's ability. It was to pick a team, and if the difference between the 20th and 21st best players is less than the measurement error, you've still done that optimally.

It's a false analogy even to use the word unfair to describe this kind of misjudgement. It's not aimed at producing a correct estimate of any given individual, but at selecting a reasonably optimal set.

One thing that leads us astray here is that the selector seems to be in a position of power. That makes him seem like a judge. If you regard someone judging you as a customer instead of a judge, the expectation of fairness goes away. The author of a good novel wouldn't complain that readers were unfair for preferring a potboiler with a racy cover. Stupid, perhaps, but not unfair.