NewYorker | In 1993, when Nick Patterson mailed
Robert Mercer a job offer from Renaissance, Mercer threw it in the
trash: he’d never heard of the hedge fund. At the time, Mercer was part
of a team pioneering the use of computers to translate languages. I.B.M.
considered the project a bit of a luxury, and didn’t see its potential,
though the work laid the foundation for Google Translate and Apple’s
Siri. But Mercer and his main partner, Peter Brown, found the project
exciting, and had the satisfaction of showing up experts in the field,
who had dismissed their statistical approach to translating languages as
impractical. Instead of trying to teach a computer linguistic rules,
Mercer and Brown downloaded enormous quantities of dual-language
documents—including Canadian parliamentary records—and created code that
analyzed the data and detected patterns, enabling predictions of
probable translations. According to a former I.B.M. colleague, Mercer
was obsessive, and at one point took six months off to type into a
computer every entry in a Spanish-English dictionary. Sebastian Mallaby,
in his 2010 book on the hedge-fund industry, “More Money Than God,”
reports that Mercer’s boss at I.B.M. once jokingly called him an
“automaton.”
In 2014, Mercer
accepted a lifetime-achievement award from the Association for
Computational Linguistics. In a speech at the ceremony, Mercer, who grew
up in New Mexico, said that he had a “jaundiced view” of government.
While in college, he had worked on a military base in Albuquerque, and
he had showed his superiors how to run certain computer programs a
hundred times faster; instead of saving time and money, the bureaucrats
ran a hundred times more equations. He concluded that the goal of
government officials was “not so much to get answers as to consume the
computer budget.” Mercer’s colleagues say that he views the government
as arrogant and inefficient, and believes that individuals need to be
self-sufficient, and should not receive aid from the state. Yet, when
I.B.M. failed to offer adequate support for Mercer and Brown’s
translation project, they secured additional funding from DARPA,
the secretive Pentagon program. Despite Mercer’s disdain for “big
government,” this funding was essential to his early success.
Meanwhile,
Patterson kept asking Mercer and Brown to join Renaissance. He thought
that their technique of extracting patterns from huge amounts of data
could be applied to the pile of numbers generated daily by the global
trade in stocks, bonds, commodities, and currencies. The patterns could
generate predictive financial models that would give traders a decisive
edge.
