Raw Thought: Aaron Swartz's weblog

I wrote this years ago (probably the summer after college), filled with nostalgia, in a small notebook on a plane to or from Boston. (I finally understand how people can write books long-form!) I meant to type it in as soon as I got back and post it here, but I lost it and never got the chance. While moving recently it turned up again and I thought I should seize the opportunity to type it in while I still had my hands on it. (Also, coincidentally, I had lunch today with Tim.)

Apologies for the poor writing; I've improved a bit in the subsequent years and it's rather hard to write well when you're pouring your memories into a notebook on a plane.

In 7th grade we were asked to do a project on the Great Men Who Made America, or something like that. Other kids chose civil rights leaders and politicians and even scientists. Then a big tech geek, I chose my hero, Tim Berners-Lee, inventor of the World Wide Web.

One of the first assignments was to do some research and answer a list of questions provided by the instructor (one of them being "what did your contemporaries think of you?" -- I still wince at not then knowing what "contemporaries" meant). I was able to answer most of them off the top of my head and most of the rest from TimBL (as he was universally called by all but Dave Winer)'s web page. But a few, like "Do you have any regrets?" were not really answerable form such material. So, with some trepidation, I sent him an email. His website said not to email him about school projects on the Web, but I convined myself that school projects on him didn't qualify. So I carefully composed my email and sent it out, hoping for the best.

A few days later I received a rather short-tempered reply. "Regrets?" he said. "I have a few. But then again, too few to mention." (I didn't get the reference until years later when I heard the words again over the radio -- I thought he was just being poetic.) But my favorite answer was his response to one of the theme questions I asked: "How do you think your work has shaped America?" or something like that. (I was taking the assignment a little too literally, I guess.) He shot back with a phrase that's stuck with me ever since: "I was an Englishman living in France and working in Switzerland -- it's got nothing to do with America." I proudly reported this back to my teacher, complaining about his assignment's Amerocentrism.

The next time I check Tim's web page, it said not to email him concerning school projects on the Web or on Tim Berners-Lee.

I didn't really run into TimBL again until a few years later when I was visiting Cambridge, Mass., where he worked at MIT. I was working on RSS at the time with W3C employee Dan Brickly and I felt cocky enough to try to go see TimBL's office. I got as far as the nameplate outside his door when a gruff man with a beard stopped me. "What are you doing?" he asked. "Seeing Dan Brickley," I lied. "That's not Dan's office," he said. "Go wait in the hall. I'll get Dan," said the man.

The man, I later learned, was Ralph R. Swick (or RRS), and it always struck me that he hated me, or at least looked down on me because of my age. In this encounter he at least could claim to not know who I was, but even later he would ignore me, ignore what I said, or in one particular instance, make fun of me. (I'll get to that later.)

Dan, on the other hand, was glad to see me and took me on a tour of the W3C, where he and TimBL worked. "Is Tim around?" I asked casully. "Oh no," he said, "he's off in [foreign country]." "Oh," I said. "But I can take you to meet our PR director and she can answer any questions you might have."

This seemed like an odd move to me. At the time, at least, I was a co-worker, not a journalist. Still, why not. She gave me the standard spiel about the W3C, which of course I'd already knew. (Although there was one funny moment: "The Web is just 10 years old," she said, obviously repeating a standard line. "Were you everything you are at 10 years old?" There was an awkward pause as she tried to calculate my age and realized I might be.)

Eager to show her I knew what I was talking about, I decided to make an informed criticism of the W3C. Maybe I challenged is corporate control, or the bylaws' undemocratic requirements that the Director (TimBL) approve everything, or maybe that and more. She gave me standard responses, but I kept pushing and she got increasingly agitated and eventually started shouting. I wasn't hurt, and pressed on, but against my will I started crying (although I imagine it may have looked intentional -- a cheap arguing tactic, perhaps). She offered me tissues and toned down, asking me to send her some concrete suggestions by email to consider.

As DanBri took me out, he tried to console me. "Don't worry," he said, "she scares everyone. That's what makes her so good." For whatever reason, she always seemed extra-nice to me after that.

I next ran into TimBL at a W3C conference in Cambridge, where they were having a meeting on RDF, which I was beginning to get interested in. The meeting involved a series of presentations on various topics and through it all, TimBL struck me as incredibly brilliant. He would type away at his laptop the whole presentation, apparently lost in some project and not paying attention to anything. Then, towards the end, he would quickly raise his hand and ask the one question that revealed all the flaws of the presenter's assumptions, cutting through everything they said.

At the end of the meeting, DanBri insisted I meet TimBL and so I went over to shake his hand. TimBL moved and talked so fast he appeared almost as a blur, bursting with energy and bouncing around the room. DanBri took my camera and TimBl slowed down for a moement to grab me around the shoulder and pose for a picture, a photo which I later used frequently. (God, I looked so young!)

I met a bunch of other cool people at the meeting -- B.K. DeLong, who introduced me to Tantek (then at Microsoft, he looked for all the world like a Borg; he later joined Technorati and hounded me incessantly to take a summer internship there (I was at dinner with him the other nigh and apparently he still thinks I should!))), libby, and the elusive mnot. But the one I remember most was Dan Connolly, or DanC as he was called.

At first, for some reason, I thought he was a kid. He wore a simple t-shirt and pulled out his guitar during breaks. His enthusiasm was even greater than TimBL's and he was clearly just as sharp, although in some ways more so, since he had a programmer's logicalness that he applied to every aspect of his life.

He worked from home in Kansas City, where he lived with his wife and kids. TO balance work and family, he worked out a contract with his life, which meant that at 5:00 he would stop whatever he was doing, no matter how important, for Family Time.

I would follow him around on the RDF chatrooms and could feel myself growing more logical just by listening to him. When I could I would sort of apprentice under him, following along and helping as he wrote programs to do various tasks. I even considered going out to visit him in Kansas City at one point.

He was an enormous influence, despite his socially conservative politics (which I always attributed to his location). He was usually quiet about them, but once he posted a petition in favor of parental abortion notification laws. At the time I was a radical on children's rights at least, and thought parents shouldn't be notified about anything. He also mentioned he was trying to figure out why his school wasn't teaching Intelligent Design (then a new phrase to me), absurdly suggesting he had to follow the money. It turns out following the money is much better at finding why people do teach Intelligent Design.

Still, I suppose everyone has flaws, and these were relatively minor. It's clear why TimBL chose DanC to be his right-hand man, handling the details for anything important.

The next time I met TimBL was at the WWW2002 conference in Hawaii. DanC invited me to wake up early for a "Semantic Web Swim" with him and TimBL but, ashamed of my body at the time, I stupidly declined -- a move I still regret.

One night DanC, some other people, and I went out to a place on the beach. I, I have to say, didn't have a sip of alcohol -- I still haven't, actually -- but everyone else got hammered. (I was worried some of them might walk into the ocean by accident and get swept away.) I left my participation ambiguous, however, when I posted a short note in an obscure place on my blog, but John Robb, then-CEO of UserLand Software (Dave Winer, founder) found it and quoted me on his blog with the comment "Aaron is 14. Someone should tell his parents." (No one ever did, as far as I know.) The next day, when I went to see UserLand employee Robert Scoble give a talk, I accidentally let my badge in the hotel room and was dragged out of the talk by security.

On Semantic Web Developer Day, I gave a short presentation about a project I was working on. Ralph Swick (the gruff fellow, you'll recall) also gave a presentation. I got in line to ask him a question, but I was last and we were out of time. "It'll be quick," I lied. "OK, they said. "Have you thought at all about the privacy implications of this?" I asked (for his software was a total privacy nightmare). "Yes," RRS responded and everybody laughed. I went back t the audience feeling like a jerk and RRS ignored me when I tried to ask him to elaborate one-on-one.

It was in the hallway one day that I first really talked to TimBL. He was breezing down the corridor when he spotted me and ran up. "Aaron!" he said. "I really appreciate all the work you['re doing -- it's great stuff. Do you think maybe you could help with a little project?]" Flustered, I said of course. He outlined a project having to do with encrypted RDF and I eagerly volunteered, amazed at my good fortune.

A few days later, at a conference dinner, I got my food and was looking for a place to sit, when TimBL came by. "Here," he said, "I think I've got a place for you" and he brought me over to the W3C table, seating me directly across from him, which was simply amazing. The table was full of good cheer and Tim commemorated the moment by passing his camera around it and asking everyone to take a photo from their point of view.

I last saw him as we were leaving when he pulled my mom aside and introduced himself. "Aaron's doing great work," he said, "but I'm a little worried. The other night at dinner all he had was rice." "Oh, he's just a fussy eater," my mom explained.

The last time my path crossed with TimBL is when I was applying for Stanford. My Dad happened to be in Cambridge at the time and insisted on asking TimBL for a letter of recommendation, by going over to TimBL's office. Apparently he had better luck than I, since I'm told TimBL agreed and I was later accepted to Stanford. I guess a letter of recommendation from the creator of the Web counts for something, even if he is an Englishman.

Everyday Utilitarianism: Who Gets the TV First?

I've often thought it would be fun to write a book on "everyday utilitarianism" -- how to apply mathematical formalizations of utilitarianism and game theory to help you solve everday life dilemmas, like who should get to use the television first or whether you should go out with that guy.

The basic idea would be that each chapter would revolve around a particular mathematical principle and demonstrate it using a concrete example from everyday life. Since I'll probably never get around to writing such a book, I figured I'd just write up such examples on my blog when I encountered them and maybe someone else would take the idea and run with it.

It's 8pm, and you settle down in front of the television to watch American Idol. Unfortunately, at the very same time your roommate is also settling down in front of the television to play one of his video games. Quickly, the two of you get into a tiff about who will get to use the television first. You both would prefer using the television first rather than second, yet, since American Idol is a live show, watching it now is a rather different experience from watching it later, while the video game will remain the same all night. How can you prove mathematically to your roommate that you should get to use the television first?

Let U(TV_0 = A), which we'll write AT0, represent the number of utiles (essentially, a measure of enjoyment) you get from watching the TV first, while BT0 represents the number of utiles your roommate gets from watching the TV first. (AT1 and BT1 represent the utiles from watching it second.) Obviously our goal is to maximize the total number of utiles (i.e. enjoyment) in the world, by picking the solution that leads to our greatest number.

First we write down what we know. Obviously you both would prefer to watch the show first, rather than second:

But since Amereican Idol is live, we can also say that the benefit you get from watching it first is bigger than the benefit your roommate gets from playing his game first. In other words:

Finally, we want to find out which is bigger: you going first and him second, or you going second and him first. Let >< represent "which is bigger?"

By this time in the proof, however, your roommate should have wandered off, leaving you to watch American Idol in peace. Unfortunately, not having seen your proof, he thinks you're just a selfish ass as opposed to trying hard to do what's best for the whole world.

How To Launch Software

37signals recommends that software developers pursue what they call the Hollywood Launch. They don't give any argument for this method, except perhaps the title (as if Hollywood was a business you should try to imitate?) -- I guess the idea is that you're supposed to do it since 37signals says to.

The basic idea behind the Hollywood Launch is simple: you release a few hints about your product to build buzz, slowly revealing more and more until the big day, when you throw open the doors and people flood your site, sent there by all the blog coverage and email alerts.

This may work well for Hollywood -- if your movie is a big hit at the box-office on opening weekend, then the movie theaters are more likely to keep showing it in the weeks to come and you get credit for being "one of the weekend's biggest films". But for software developers, it's moronic. Your software isn't being released in theaters, it's available over the Web. You don't have to worry about the theater no longer showing after week one; you can keep pushing it for years, growing your userbase.

Instead what happens when software developers try the Hollywood Launch, and I've seen this many times, is that users indeed do flood to your site on launch day but...

Tomorrow, hardly any of those users come back. Your traffic graphs look like the sharpest mountain you've ever seen: a huge climb up and then, almost immediately, a similarly-sized crash back down.

So what do you do then? Well, you do what you should have done all along: you grow the site.

I'll call this technique the Gmail Launch, since it's based on what Gmail did. Gmail is probably one of the biggest Web 2.0 success stories, so there's an argument in its favor right there. Here's how it works:

The result will be a graph that just keeps accelerating and climbing up. That's the graph that everyone loves to see: solid growth, not a one-day wonder. Good luck.

Since 37signals quotes from people who followed their advice, I thought I might as well do the same. mojombo:

The Predator State: A Summary

James K. Galbraith's The Predator State is undoubtedly one of the most important books on the economics of our era. Galbraith sets himself the task, not only of exposing the discredited economic orthodoxies of our generation, but also documenting the economy as it really exists, and setting an agenda for the future. It is a book that desperately need to be listened to. And, even better than all that, it's a fun read. Go out and buy it immediately.

That said, here is a brief, abbreviated summary of the book, to better pull out its themes and spread its message. It is of necessity less clear and less well-argued than the book itself, which I you should actually read if you want to argue, but it should give the gist of things.

1: The Reaganites swept into power on the arguments of economic conservatives: lower taxes, tight money, and an assault on all opponents of market forces (government, regulation, unions). Their views were tried and failed completely. They have no remaining defenders in academia and only slogans and cronies outside of it. There is no longer any vision on the right; the left should leave its defensive crouch and start proposing something new.

2: Friedman and friends said that markets would lead to democracy -- that "economic freedom" begets political freedom. But economic freedom isn't what it sounds like; it's not freedom from economic want but instead, as Friedman put it, "the freedom to choose" or, in other words, "the freedom to shop". But control over production is as unfree as in the Soviet Union, with advertising for propaganda, R&D for planning, and Wall Street analysts for government inspectors. "Lines form, under capitalism, every day."

3: Supply-siders argued that a) saving is a public good because it leads to investment, b) America does not save enough compared to other countries, c) saving would be unleashed by lowering taxes on it, d) the resulting investment would spur an economic boom. Every piece of this is wrong: a) in an efficient market, all the benefits from investment are captured by the investor; thus investment cannot be a public good unless markets are inefficient, in which case the government should step in more, b) the correct amount of saving is a policy decision, there's no reason to believe other countries have it right (the Soviet Union had a 40% level of saving right up to its collapse), c) rich people save most of their money anyway (it's impossible to consume that much) and changes in interest rates dwarf changes in tax rates; furthermore, real investment is encouraged by high personal taxes, since this forces people to keep their money in corporations, d) personal saving is less than 1% of GDP; almost all investment comes from corporations or overseas.

4: Milton Friedman claimed that high inflation (it was 10% in the 1970s) was just the result of printing too much money. Reagan's Fed adopted this belief, sending the US and many foreign countries into deep recession. Eventually, the policy was completely abandoned and high inflation has not been seen since. Serious inflation isn't caused by printing money, but by wage-price spirals -- the price of oil shot up, causing rising prices to cover oil costs, causing workers to demand higher wages to pay those prices, causing prices to rise even higher, and so on. Today, most prices are set by overseas manufacturers and labor unions are so weak that workers can't demand wage increases. Inflation is dead.

5: Democrats (and some Republicans) repeatedly insist that we need to balance the budget or face fiscal collapse. But the budget is ruled by a simple equation: the total amount the government owes + the total amount the public owes = the total amount we owe to foreign countries. This is simple logic: whatever is not owed within the country must be owed to another country. But the international economy depends on other countries keeping large reserves of dollars (see 14), meaning our trade deficit must be high. As long as this is so, we must either have the government run large deficits or ask people to do so. The budget deficit was closed in the late 1990s because citizens picked up the slack with high credit card spending and home equity loans, inevitably leading to a slump. Balancing the budget is for suckers; Democrats should spend the money on public goods instead, promoting economic growth and thus raising tax revenue.

6: The argument for free trade comes from Ricardo's "comparative advantage" -- a clever textbook exercise, but irrelevant to the real world since it assumes constant costs. In reality, either you produce manufactured goods, in which your costs go down as you make more, or you sell off commodities, in which case your costs go up as you make more. With the former, it takes time for local industry to build up the advantage (requiring protectionism). With the latter, you end up like Mongolia, which opened up its animal husbandry market, swelling herd sizes, turning grass into permanent desert, and killing off the entire market. With no other exports, such a country is in big trouble. Ricardo was wrong: diversification, not specialization, is the way to develop -- and how every successful country has. Unfortunately, we've forced this broken system on most of the world. (China has escaped, letting state-supported banks fund money-losing new companies until they grow large enough to succeed as exporters. In the mean time, they dump their products on local Chinese, allowing them to have a very high standard of living at very low wages.)

7: There is no trade-off between equality and efficiency. Instead, equality leads to efficiency. Denmark is one of the most equal countries in Europe, and as a result one of the wealthiest. The rest are on a continuum down to unequal and inefficient. Full employment and high wages require companies to make the most of the employees they have, increasing efficiency. Raising the minimum wage doesn't raise unemployment, it lowers it -- unemployment and inequality have risen and fallen together since 1920. Higher wages lead to more jobtaking and less quitting. The remaining increase in inequality was caused by stock market giveaways to dot-commers and Bush giveaways to government contractors -- which is why it was limited to Silicon Valley and the Potomac, respectively.

8: The US is not a free market. Of GDP, 17% is health care (where experts, not consumers decide how to spend), 16% is housing (subsidized by quasi-public mortgage firms and tax deductions), 15% is federal welfare, 14% is local welfare, 4.5% is military spending, 3% is higher education (paid for mostly by government or conspicuous philanthropy¹ and consumed for status and not value). Together, 70% of US GDP is planned; it's just that our facade of a free market makes us less efficient at planning than other countries (especially in health care).

9: In the 1970s, American industry (particularly steel and cars) was being challenged and weakened by Japan. Reagan's assault on inflation (see 4) dealt them a death blow, sending their foreign and domestic markets into deep recession, driving up the value of the dollar (making their exports more expensive than their competitors'), and raising interest rates. In the 1980s the technical staff left for Silicon Valley, and 1990s financial fraud killed off what remained. When new startup founders paid themselves exorbitant salaries from VC money other CEOs rushed to keep up, making them all wealthy enough to become a separate class. They used their new power to prey on the corporations that they ran.²

10: Previously, regulation kept the predators in check -- unions, NGOs, and progressive businesses pushed government standards to kill regressive competitors. But newly-wealthy predator CEOs had the Republicans take over and gut regulation. The result is the Predator State, where every new law is a corporate giveaway. Prescription drug benefits for Big Pharma; NCLB to defund and deskill schools (building support for vouchers); and Social Security reform to give workers' paychecks to Wall Street. (Democrats have so far prevented the latter, but corporate-funded think tanks now aim to take them down from inside.) The programs allow further predation; privatizing college loans has led loan companies to bribe student loan officers. It's not that Republican government fails at tasks like stopping Katrina; it's that such tasks of governance are not its goal -- opening up New Orleans for Halliburton contracts is.

11: The great liberal economic agenda is "making markets work" -- small fixes for market failures. The canonical example is job training to fight unemployment. But job training does not create new jobs, economic growth does; the tech boom was the last time we saw a real decrease in unemployment. Similarly, some Dems propose universal preschool since experiments find kids with free preschool grow up to get better-paying jobs. But those preschools did not create jobs, they just gave their students an advantage in getting them. Universal preschool would give everyone that advantage, leaving no net impact. And creating markets in unmarketable goods (health care, energy, the climate) is doomed to failure. In these industries markets will not work; planning is required.

12: Planning is alleged to have been disproven by the Soviet Union's fall. But it is unavoidable. The market, even when it does work, fails to take into account the wishes of the poor and the needs of the future, since neither can buy things today. New Orleans fell not because of a lack of foresight (it was predicted by the local paper) or technology (the Army knew how to build strong levees) but because we lacked a plan -- nobody in power bothered to do anything about it. Similarly, climate change will melt Antarctica and drown New York, Boston, South Florida, Houston, the Bay Area, London, the Netherlands, Bangladesh, and Shanghai. Stopping it requires a plan; an enormous one ranging from elementary school classes to government-funded research centers to a WWII-level restructuring of the economy.

13: Deregulation can have three effects: 1) increasing competition and lowering wages and prices, 2) speeding technological change and increasing quality, 3) creating monopolies and raising prices. Trucking deregulation did 1, airline deregulation did 1 and 2, but telecom, banking, and energy deregulation did 3. Charles Keating donated to the government, leading VP George H. W. Bush's task force to deregulate his industry and allow the Savings and Loan Scandal. Ken Lay was Bush's largest contributor, leading VP Dick Cheney's task force to deregulate his industry and allow the Enron energy scandal.

The solution is to lower CEO pay, raise the minimum wage, and set wage standards in between. Some liberals claim trade is the problem and the solution is to set environmental and labor standards on other countries. These are unenforceable and will be ineffective (companies moving overseas already build clean factories since that's most efficient and no significant exports are made using child or prison labor). Instead, we should set wage standards at home, like Scandinavia, forcing companies to increase productivity and pay fair wages. Wage standards should also apply to undocumented workers; illegal immigration is caused by employers who send recruiters to Mexico for compliant and low-paid workers. Applying wage standards to all will end these abusive practices.

14: Any country that can pay for its imports entirely with exports can organize its internal economy (its people and resources) however it likes. Countries that do not balance their trade depend instead on global capital markets and must play by their rules. But the US is a special case: after World War II (1944) it set up the Bretton Woods system of international exchange, pegging all currencies to the dollar and backing the dollar with gold reserves. But during Vietnam's deficits (1971), Nixon broke the system, devaluing US currency and wreaking havoc on the rest of the world. Reagan's tight money policies (1981) caused so much instability that other countries were forced to build up reserves of US Treasury Bonds in exchange for military, economic, and export security. US bubbles and the Soviet Union's fall make this system less secure than before, but as long as it remains the US can do whatever it likes economically. And it might as well, since economic success will strengthen the system and the policies proposed here will lead to economic success.

Conspicuous philanthropy is like conspicuous consumption, a way for the rich to flaunt their wealth, only far more effective -- you can outdo your neighbors simply by adding another zero to the check, the buildings with your name on them live on after you die, and the government gives you a tax deduction. ↩
See the classic Thorstein Veblen, Theory of the Leisure Class for more on predation. ↩

Utilitarian Equilibriums

Utilitarians believe that people should work to maximize total happiness across the population. They believe that the only reason to do something or not to do something is because it will make people happier or unhappier respectively. Thus, whether something is good or bad depends to some extent on people's preferences -- whether it makes them happy or sad.

But this might leads to some odd conclusions in the case of what might be called "perverse preferences". For example, some members of the Bush Administration say they get very sad when they see others eat ice cream in public. Yet many people like to eat ice cream in public. Should we stop them from doing so just because it makes others sad? Moreover, should they decide to stop doing so if they're utilitarians?

Or, to take a more realistic example, many people like having sex with other people of the same gender, while others insist that such sex makes them sad. Should a utilitarian gay person remain quiet about how they're having gay sex because speaking up will upset homophobes?

A naive utilitarian might say yes, while intuition says the right answer is no. I think the actual right answer is no, and here's the reason:

Distaste for gay sex is a perverse preference, in that it's a preference about other people's behavior that doesn't directly affect you. And, empirically, we find that such preferences don't usually withstand frequent exposure to the behavior itself.

In other words, the more public people are about the gay sex they're having, the less hearing about gay sex tends to make people feel bad. Thus, being public about the gay sex you're having is the optimal behavior because it:

That is, being public about the behavior helps transition from a society with lots of closeted gays and lots of homophobes to a society with lots of proud gays and few homophobes. There's more overall happiness in the second society so it's right to do things that bring us closer to it.

The Percentage Fallacy

There's one bit of irrationality that seems like it ought to be in behavioral economics introduction but mysteriously isn't. For lack of a better term, let's call it the percentage fallacy. The idea is simple:

I'm sure all of you have done something similar -- maybe the issue wasn't having to return something, but spending more time looking for a cheaper model, or fiddling with coupons and rebates, or buying something of inferior quality. But the basic point is consistent: we'll do things to save 50% that we'd never do to save 1%.

At first this almost seems rational -- of course we're going to do more to save more money! But you aren't saving more money. With both the blender and the laptop, you have the chance to save $20. Either way, you're going to have another twenty in your pocket, which you can spend on exactly the same things later on. Yet we behave differently depending on whether we got that twenty by skimping on a small purchase or skimping on a big one. Rationally, if driving back to the store isn't worth $20 when you're buying a laptop, it isn't worth $20 when you're buying a blender.

On the other hand, don't those small savings tend to add up after a while? If you start blowing $20 every time you buy a trinket, you're soon going to be out of disposable income. Meanwhile, spending several thousand dollars is much rarer, so isn't it OK to slack off a bit on such occasions?

If we work to save 50% on everything, big or small, that's the equivalent of saving 50% of our money altogether. Whereas if we only try to save fixed amounts on every purchase, how much we save is dependent on how many things we buy.

Capital and its Complements: A Summary

Adam Smith explained that in all countries with "security of property and tolerable administration of justice" citizens would spend all their money (capital), either on consumption or investment, causing the country's economy to grow. After some contention, later economic studies tended to bare this out: a shortage of capital wasn't always the bottleneck, but when it was, removing it could lead to extraordinarily rapid growth.

The problem for poor countries is that, because of high mortality rates (which require more children to have some survive) and low educational levels (which mean those children can find productive employment quickly), they have high population growth and thus low capital-to-labor ratios. Worse, trade allows you to spend your money buying manufactured goods from overseas, for which you have only your very cheap labor to provide in return. The result is that it requires an enormous amount of domestic investment to improve capital-to-labor ratios.

And so rich country economists made "the neoliberal bet" on behalf of poor countries: they hoped that loosening restrictions on international capital flows would send capital rushing in to poor countries and build their economies, the same way that Great Britain's massive investment in a young United States (in 1913 Britain's foreign assets equaled 60% of its domestic capital stock) built up that country.

But what ended up happening was exactly the opposite. Yes, NAFTA led US companies to invest the $20 to $30 billion a year on manufacturing in Mexico that its boosters predicted, but that investment was more than outweighed by the $30 to $40 billion a year fleeing the country from Mexico's wealthy wanting to invest it in the United States. Why? In part because the US was more politically stable, and thus a safer investment climate. And in part because the US treats its own workers so poorly -- with productivity rising 35% since 2000 while real wages remain flat -- it provides an excellent investment opportunity.

But meanwhile, all this investment in the US was dwarfed by the Chinese acquisition of our debt (and thus the political risk it represents). China needed to do this, since US purchase of their exports is the only thing funding the manufacturing-led industrialization of a massive portion of their economy; there would be massive dislocation if that funding dried up.

"Recognition of these facts came slowly." First, Larry Summers said it was our unsustainable current account deficit. (That was the 1990s; today that deficit is four times as large.) Later, economists thought it must have been our large budget deficits. Then they began thinking it was the run-up in housing prices. But that, it is now clear to most economists, was the result of a bubble. And yet the flow of capital to the US continues. But, perhaps even more frighteningly, it could stop at any moment.

Last Goodbyes

It's minutes to midnight and I'm hurriedly packing. Early tomorrow morning I catch a flight to Boston and start my new life. I haven't really gotten much of a chance to pack until now, because I've spent the past few days in a rush of meetings, getting in my last goodbyes for everyone I know in San Francisco.

It's been great seeing everyone, but like most locals, they're all puzzled as to why I'm leaving. I've been struggling to explain why. When I say the weather, everyone just laughs. When I say San Francisco is too loud, they start arguing. When I say it's the people, they tell me to find a better group of friends.

And the thing is, they're right. It's none of these. I've been spectacularly unable to articulate it, but the real answer is simpler and more prosaic. And now, after great thought and struggle, I realize the answer is simply this: Cambridge is the only place that's ever felt like home. It's that simple. And when you put it that way, it's clear why I have to go.

So goodbye Stanford, goodbye Palo Alto; goodbye south bay, goodbye peninsula; goodbye Change Congress, goodbye Creative Commons; goodbye Mission, goodbye SOMA; goodbye friends, goodbye loved ones; goodbye San Francisco, home to everyone I've ever loved. You'll always have my heart.

Scenes

"God, I'm so sick of this stuff. Can't we just go home?" she wines. "Jesus," I say, "would it kill you to go one more place?" It's been a long hot day in strange, busy New York City, and we're not exactly at our best. In fact, the combination of heat and exhaustion has turned our love bitter, brought on the darkness and recriminations. Its at moments like these, the dark depths of a relationship, that you wonder how things could ever work. As we walk down the steps we hear a subway car approach. We accelerate, running to catch it. Its doors open. We're moving faster now, pushing our way through the bustle of Manhattanites to make it. The bell sounds and I jump inside and hear the doors whoosh closed behind me. I spin around only to see her trapped on the other side of the glass. I put my hand up to it, but the train accelerates and she's left standing there, just another face in the crowd.

"Hey, want to see the game? Want a ticket to the Giants game?" I do not, in fact, want to see the game -- this or any other game. I hate sports. Yet the scalpers, apparently unaware of this, insist on trying to sell me one. That's what I get for walking near the ballpark, I guess. As I curse my choice of scenery, a cop pulls up. He lowers his window and leans out toward the scalper. The scalper hands him a ticket and the cop speeds off. "But he didn't pay!" a man in a suit walking by complains. "Cops get a special deal," explain the scalper. The man in the suit laughs and marvels at the scene.

It's weird being back at Stanford in the summer. Everything's so empty, nobody's around. Well, not nobody -- there seems to be some action near the main quad. There are drum kits spread around and golf carts and purple uniforms lying about. But most of all, there are people -- a bunch of students just standing around awkwardly. I'm about to ask one of them what's going on when a bell rings and a voice shouts "Background!" Suddenly all the students snap to attention, begin walking in perfect lines with bookbags slung over their shoulder, bicycles ridden in perfect formation. These aren't students at all, I realize with a lurch -- they're extras. It's disconcerting. A police guard is at the side, keeping kids from running over the camera crew. I ask her what they're filming. "Disney's High School Musical," she says quickly, trying to keep a student from cycling over the director's cart.

Moving On

In November 2006, I moved to San Francisco because I had to: my company got acquired and us moving out was a condition of the agreement. It was the first time I'd ever actually lived in San Francisco, as opposed to just visiting, and I quickly realized that although it was a fun place to visit, I couldn't stand living here.

Even after all this time, I can't really put my finger on what it is I don't like -- in fact, I suspect it's probably harder for me now to explain it than it was when I first came here. The first thing that comes to mind is how loud the city is. I want a place where I can live quietly and focus on my work; but San Francisco is filled with distractions. There are always crews tearing up the street, trains that are delayed, buses that have broken down, homeless people begging, friends having parties, and so on. It's impossible to concentrate and without my concentration, I feel less like me.

The other big problem is that San Francisco is fairly shallow. When I go to coffee shops or restaurants I can't avoid people talking about load balancers or databases. The conversations are boring and obsessed with technical trivia, or worse, business antics. I don't see people reading books -- even at the library, all the people are in line for the computer terminals or the DVD rack -- and people at parties seem uninterested in intellectual conversation.

And so I'm moving back to Cambridge, Massachusetts -- Harvard Square in particular, the one place I've ever been to that brings a special delight to my eyes, that warms my heart just to see. Surrounded by Harvard and MIT and Tufts and BC and BU and on and on it's a city of thinking and of books, of quiet contemplation and peaceful concentration. And it has actual weather, with real snow and seasons and everything, not this time-stands-still sun that San Francisco insists upon.

But I'm also sad to leave my responsibilities in San Francisco. One of which I'd particularly like your help with. I've been honored and overjoyed to help Lawrence Lessig get his Change Congress project off the ground. If you haven't heard, he's trying to build a national movement to get the corruption out of Congress; to pass public financing of public elections, earmark reform, and other pressing concerns.

But they need a full-time day-to-day tech organizer. Someone who knows how to blog and who the bloggers are and can keep them in touch with the community. Someone who knows enough about technology to know the tools that can be built and should be. And someone with enough drive and talent to make sure those things get built. It's a dreamy job and I hope there's someone out there who will take it from me. A more formal write-up is on the Change Congress blog.

Is Undercover Over?

How to Promote Startups

When people talk about how government can promote startups, there seems to be a fairly standard consensus: we need more economic inequality. Lower income and capital gains taxes provide more incentive to work, looser labor laws make it easier to fire non-performers, and large private wealth funds provide investment capital.

But having been through a startup myself, I think there's much more you can do in the other direction: decreasing economic inequality. People love starting companies. You get to be your own boss, work on something you love, do something new and exciting, and get lots of attention. As Daniel Brook points out in The Trap, 28% of Americans have considered starting their own business. And yet only 7% actually do.

What holds them back? The lack of a social safety net. A friend of mine, a brilliant young technologist who's been featured everywhere from PBS to Salon, stayed in academia and the corporate world while all of her friends were starting companies and getting rich. Why? Because she couldn't afford to lose her health insurance. Between skyrocketing prices and preexisting condition exclusions, it's almost impossible for anyone who isn't in perfect health to quit their job. (I only managed because I was on a government plan.)

Anyone with children is also straight out. Startup founders tend to be quite young, in no small part because no one can afford to support a family on a startup founder's salary. But if we had universal child care, that would be much less of an issue. Parents would be free to pursue their dreams, knowing that their children were taken care of. And universal higher education could let parents spend their savings on getting a business started, instead of their children's tuition. Plus, it'd give many more kids the training and confidence they needed to start a company.

And those large private wealth funds that result from growing inequality? A real problem for startup founders is that they're too large. It used to be that you could borrow a couple thousand dollars from friends and neighbors to get your business off the ground. Nowadays, they're too busy trying to make ends meet to be able to afford anything like that. Meanwhile, those large wealth funds I mentioned are now so big they can only afford to invest in multi-million dollar chunks -- much more than the average founder needs, or can even justify. And the large investments come with large amounts of scrutiny, further narrowing the recipient pool.

But imagine if the government provided a basic minimum income, like Richard Nixon once proposed. Instead of having to save up (increasingly difficult in a world in which the only way to survive is on credit card debt) or borrow money to stay afloat, you could live off the government-provided income as you got things started. Suddenly having to quit your job would no longer be such a huge leap -- there'd be a real social safety net to catch you. (Not to mention if those labor laws some people want to loosen required your old job to take you back if things didn't work out.)

Of course, there is some truth to the standard proposals. Some startup founders are encouraged by dreams of financial security, and high taxes can make that dream more elusive. And complex labor regulations can make it difficult to get new companies off the ground. But it's not an issue of whether we should have taxes or labor laws -- it's an issue of how they're targeted.

Estate taxes on inherited fortunes would have basically no impact on startup founders, but could go a long way to funding a social safety net. And since most startups are acquired as stock, income taxes are basically irrelevant -- it's really capital gains tax that gets applied. There's no reason the government couldn't apply a lower capital gains tax to startups that get acquired than they do to the shares of publicly-traded companies that large investors trade.

The same is true for labor laws: preventing large companies from firing people at random can provide some much-needed stability to their lives, especially if they're saving up money in the hopes of going into business themselves. But there's no reason such laws also have to be applied to small startups, where the company is more likely to go out of business than to fire you.

Look at social democratic Europe, where these policy prescriptions have been tried. While there's much less of a culture of entrepreneurship and only 15% of Europeans think about starting their own company, nearly all (14.7%) of them actually go ahead and do it.

The fact is, if governments really want to promote startups and the economic innovation they bring, they shouldn't listen to the standard refrain of cut taxes and deregulate. They need to start rebuilding the social safety net, so that their citizens know that if they go out on a limb and try something risky, someone will be there to catch them if things don't work out.