Super-Intelligent Humans Are Coming

Stephen Hsu in Nautilus:

Feynman-bongosIn Daniel Keyes’ novel Flowers for Algernon, a mentally challenged adult called Charlie Gordon receives an experimental treatment to raise his IQ from 60 to somewhere in the neighborhood of 200. He is transformed from a bakery worker who is taken advantage of by his friends, to a genius with an effortless perception of the world’s hidden connections. “I’m living at a peak of clarity and beauty I never knew existed,” Charlie writes. “There is no greater joy than the burst of solution to a problem… This is beauty, love, and truth all rolled into one. This is joy.” The contrast between a super-intelligence and today’s average IQ of 100 would be greater still. The possibility of super-intelligence follows directly from the genetic basis of intelligence. Characteristics like height and cognitive ability are controlled by thousands of genes, each of small effect. A rough lower bound on the number of common genetic variants affecting each trait can be deduced from the positive or negative effect on the trait (measured in inches of height or IQ points) of already discovered gene variants, called alleles.

The Social Science Genome Association Consortium, an international collaboration involving dozens of university labs, has identified a handful of regions of human DNA that affect cognitive ability. They have shown that a handful of single-nucleotide polymorphisms in human DNA are statistically correlated with intelligence, even after correction for multiple testing of 1 million independent DNA regions, in a sample of over 100,000 individuals. If only a small number of genes controlled cognition, then each of the gene variants should have altered IQ by a large chunk—about 15 points of variation between two individuals. But the largest effect size researchers have been able to detect thus far is less than a single point of IQ. Larger effect sizes would have been much easier to detect, but have not been seen. This means that there must be at least thousands of IQ alleles to account for the actual variation seen in the general population. A more sophisticated analysis (with large error bars) yields an estimate of perhaps 10,000 in total.1 Each genetic variant slightly increases or decreases cognitive ability. Because it is determined by many small additive effects, cognitive ability is normally distributed, following the familiar bell-shaped curve, with more people in the middle than in the tails. A person with more than the average number of positive (IQ-increasing) variants will be above average in ability. The number of positive alleles above the population average required to raise the trait value by a standard deviation—that is, 15 points—is proportional to the square root of the number of variants, or about 100. In a nutshell, 100 or so additional positive variants could raise IQ by 15 points. Given that there are many thousands of potential positive variants, the implication is clear: If a human being could be engineered to have the positive version of each causal variant, they might exhibit cognitive ability which is roughly 100 standard deviations above average. This corresponds to more than 1,000 IQ points.

More here.