Paul Graham - Essays

[ 11] Letter to Ottoline Morrell, December 1912. Quoted in:

Monk, Ray, Ludwig Wittgenstein: The Duty of Genius , Penguin, 1991, p. 75.

[ 12] A preliminary result, that all metaphysics between Aristotle and 1783 had been a waste of time, is due to I. Kant.

[ 13] Wittgenstein asserted a sort of mastery to which the inhabitants of early 20th century Cambridge seem to have been peculiarly vulnerable—perhaps partly because so many had been raised religious and then stopped believing, so had a vacant space in their heads for someone to tell them what to do (others chose Marx or Cardinal Newman), and partly because a quiet, earnest place like Cambridge in that era had no natural immunity to messianic figures, just as European politics then had no natural immunity to dictators.

[ 14] This is actually from the Ordinatio of Duns Scotus (ca. 1300), with "number" replaced by "gender." Plus ca change.

Wolter, Allan (trans), Duns Scotus: Philosophical Writings , Nelson, 1963, p. 92.

[ 15] Frankfurt, Harry, On Bullshit , Princeton University Press, 2005.

[ 16] Some introductions to philosophy now take the line that philosophy is worth studying as a process rather than for any particular truths you'll learn. The philosophers whose works they cover would be rolling in their graves at that. They hoped they were doing more than serving as examples of how to argue: they hoped they were getting results. Most were wrong, but it doesn't seem an impossible hope.

This argument seems to me like someone in 1500 looking at the lack of results achieved by alchemy and saying its value was as a process. No, they were going about it wrong. It turns out it is possible to transmute lead into gold (though not economically at current energy prices), but the route to that knowledge was to backtrack and try another approach.

There's something interesting happening right now. Startups are undergoing the same transformation that technology does when it becomes cheaper.

It's a pattern we see over and over in technology. Initially there's some device that's very expensive and made in small quantities. Then someone discovers how to make them cheaply; many more get built; and as a result they can be used in new ways.

Computers are a familiar example. When I was a kid, computers were big, expensive machines built one at a time. Now they're a commodity. Now we can stick computers in everything.

This pattern is very old. Most of the turning points in economic history are instances of it. It happened to steel in the 1850s, and to power in the 1780s. It happened to cloth manufacture in the thirteenth century, generating the wealth that later brought about the Renaissance. Agriculture itself was an instance of this pattern.

Now as well as being produced by startups, this pattern is happening to startups. It's so cheap to start web startups that orders of magnitudes more will be started. If the pattern holds true, that should cause dramatic changes.

So my first prediction about the future of web startups is pretty straightforward: there will be a lot of them. When starting a startup was expensive, you had to get the permission of investors to do it. Now the only threshold is courage.

Even that threshold is getting lower, as people watch others take the plunge and survive. In the last batch of startups we funded, we had several founders who said they'd thought of applying before, but weren't sure and got jobs instead. It was only after hearing reports of friends who'd done it that they decided to try it themselves.

Starting a startup is hard, but having a 9 to 5 job is hard too, and in some ways a worse kind of hard. In a startup you have lots of worries, but you don't have that feeling that your life is flying by like you do in a big company. Plus in a startup you could make much more money.

As word spreads that startups work, the number may grow to a point that would now seem surprising.

We now think of it as normal to have a job at a company, but this is the thinnest of historical veneers. Just two or three lifetimes ago, most people in what are now called industrialized countries lived by farming. So while it may seem surprising to propose that large numbers of people will change the way they make a living, it would be more surprising if they didn't.

When technology makes something dramatically cheaper, standardization always follows. When you make things in large volumes you tend to standardize everything that doesn't need to change.

At Y Combinator we still only have four people, so we try to standardize everything. We could hire employees, but we want to be forced to figure out how to scale investing.

We often tell startups to release a minimal version one quickly, then let the needs of the users determine what to do next. In essense, let the market design the product. We've done the same thing ourselves. We think of the techniques we're developing for dealing with large numbers of startups as like software. Sometimes it literally is software, like Hacker Newsand our application system.

One of the most important things we've been working on standardizing are investment terms. Till now investment terms have been individually negotiated. This is a problem for founders, because it makes raising money take longer and cost more in legal fees. So as well as using the same paperwork for every deal we do, we've commissioned generic angel paperwork that all the startups we fund can use for future rounds.

Some investors will still want to cook up their own deal terms. Series A rounds, where you raise a million dollars or more, will be custom deals for the forseeable future. But I think angel rounds will start to be done mostly with standardized agreements. An angel who wants to insert a bunch of complicated terms into the agreement is probably not one you want anyway.

Another thing I see starting to get standardized is acquisitions. As the volume of startups increases, big companies will start to develop standardized procedures that make acquisitions little more work than hiring someone.

Google is the leader here, as in so many areas of technology. They buy a lot of startups— more than most people realize, because they only announce a fraction of them. And being Google, they're figuring out how to do it efficiently.

One problem they've solved is how to think about acquisitions. For most companies, acquisitions still carry some stigma of inadequacy. Companies do them because they have to, but there's usually some feeling they shouldn't have to—that their own programmers should be able to build everything they need.

Google's example should cure the rest of the world of this idea. Google has by far the best programmers of any public technology company. If they don't have a problem doing acquisitions, the others should have even less problem. However many Google does, Microsoft should do ten times as many.

One reason Google doesn't have a problem with acquisitions is that they know first-hand the quality of the people they can get that way. Larry and Sergey only started Google after making the rounds of the search engines trying to sell their idea and finding no takers. They've been the guys coming in to visit the big company, so they know who might be sitting across that conference table from them.

Risk is always proportionate to reward. The way to get really big returns is to do things that seem crazy, like starting a new search engine in 1998, or turning down a billion dollar acquisition offer.