Paul Graham, You Can Open a Satellite Office

The other day, Paul Graham posted an especially weak argument for relaxing American immigration laws to make room for more elite programmers to be imported to startups in California.

Graham, unaccustomed to actually being opposed by anyone, opens his essay with a falsely dichotomous straw man:

American technology companies want the government to make immigration easier because they say they can’t find enough programmers in the US. Anti-immigration people say that instead of letting foreigners take these jobs, we should train more Americans to be programmers. Who’s right?

The technology companies are right. What the anti-immigration people don’t understand is that there is a huge variation in ability between competent programmers and exceptional ones, and while you can train people to be competent, you can’t train them to be exceptional. Exceptional programmers have an aptitude for and interest in programming that is not merely the product of training.

My suggestion would actually be to pay those American programmers more, because they are underpaid relative to their productivity. They need more generous stock options packages, higher salaries, and greater compensation for cost of living increases, especially in California. Investors need to be willing to give up more of their capital gains to technical talent.

This is part of what used to make Silicon Valley more competitive compared to alternative employers, but perhaps one of the greatest fundamental shifts in that industry since the last economic cycle is the over-compensation of investors relative to employees, and the shift in importance from the public stock markets to private gangs of insider investors.

It is convenient for California investors to blame Washington intransigence for under-performance, and certainly Washington is to blame for much, but part of the broader problem may be that the modern Silicon Valley-internationalist model is not a terribly good one.

Further, almost no elite programmers can be said to be ‘trained:’ John Carmack was a juvenile delinquent and mental patient. It’s just bad rhetoric to claim that the choice faced by American employers is either to change US immigration law or to magically train geniuses in a way that they have never been trained before.

And since good people like good colleagues, that means the best programmers could collect in just a few hubs. Maybe mostly in one hub.

What if most of the great programmers collected in one hub, and it wasn’t here? That scenario may seem unlikely now, but it won’t be if things change as much in the next 50 years as they did in the last 50.

We have the potential to ensure that the US remains a technology superpower just by letting in a few thousand great programmers a year. What a colossal mistake it would be to let that opportunity slip. It could easily be the defining mistake this generation of American politicians later become famous for. And unlike other potential mistakes on that scale, it costs nothing to fix.

Maybe so, but going back to the Carmack example, he made his own hub by forming a company around himself, and had no particular need to yammer with other geniuses over drinks, because people looking for his advice simply corresponded with the genius at a distance.

Mr. Graham might respond by saying that he’s speaking of merely ‘great’ programmers rather than geniuses, but that is only a difference of degree. While the genius comes up with an advance that is likely to impact multiple companies over a period of decades, you can at least build an entire department around the merely ‘great.’ But they must be selected in much the same manner.

Graham is demonstrating a belief in the Richard Florida theory of economic growth: that hip, happening urban centers chock-full of queers and artists are what generate economic dynamism.

Computer scientists like Jeff Dean do not result from a Richard Florida process. This ‘hub’ theory is ridiculous in light of the small number of geniuses there are, and their life histories. Geniuses are born and then possibly discovered: they can’t be manufactured.

The earlier version of Silicon Valley believed that much of genius was genetically innate. This would indicate that, to ensure an innovative future, you need to be better at producing smart babies, rather than importing them from foreign countries.

So, the real solution is to encourage artificial selection for intelligence, which would mean discouraging the most intelligent women away from the work force and towards the nursery. After all, from one woman, you can only ever extract one life-time of white collar work, but if you put her to having children, you can get her to produce the equivalent of multiple life-times of productive work through the miracle of reproduction.

Graham and his contemporaries will never do this, because they know what happened to William Shockley, himself an immigrant from England, for arguing for it.

My suggestion to Graham would be to encourage portfolio companies to open satellite offices outside of California, and better yet, to move out of California to wherever the talent is. This is not even out of the question for other elite American industries: investment banks in particular are known for setting up extensive satellite headquarters in foreign countries to better serve clients there, and to facilitate foreign recruitment.

This method gives the home country the benefits of foreign trade, without the downsides of importing a foreign population or needing to change domestic immigration laws. It would be a win-win for everyone involved except for lazy California investors, who, almost uniquely among the investing class, expect companies that they put money into to be within commuting distance of their insipid suburban estates.

Somehow, the far larger and more successful investment communities outside the Bay Area make do with first class travel around the world — maybe it’s time for Sand Hill Road to spring for gold club airline memberships like the other grown-ups and learn how to sleep on a trans-Pacific flight.