West Hunter worries about potential problems in a scenario where genetic engineering vastly increases human intelligence:
It risks the death or domestication of the human race.
Now genetic engineering certainly exists, and human intelligence is largely genetically determined, but it does not follow that a disaster scenario is remotely plausible. The first objection is from the purely engineering side. Suppose Steve Hsu’s estimates of the genetic architecture of intelligence are correct and the number of sites that determine g is on the order of 10,000, with about 1,000 sites per standard deviation. Editing a few thousand sites would be trivial in a computer file, but biological tools don’t work like that. There is always a bit of hit or miss because of thermodynamic constraints, and the effort required increases exponentially with the hit/miss ratio. Even though gene-editing tools are improving from the older versions used by those Chinese biologists who made waves last year trying to fix beta-thalassemia in human embryos with CRISPR, they are probably a decade from practical application, and even then involving only a few edits. Similar engineering difficulties afflict de novo genome synthesis. Living cells have the immense advantage of being able to limit themselves to copying an intact genome and are extremely efficient at this, so the engineering difficulties are not in principle insuperable, but they are nevertheless very great. In one respect, genome proofreading or editing for highly polygenic traits is easier on technical requirements compared to therapeutic applications, because you don’t particularly care about any single one of the thousands of sites you’re targeting. This means you don’t need a hit rate that close to 100%. On the other hand, even a fraction of a percent failure rate per site – whether from off-site damage, incorrect DNA repair or the sheer volume of foreign proteins and RNA dumped into the germ cell – will mean that an overall success rate per embryo will be very low, and the lower the higher traits you shoot for. This will severely limit the potential growth rate of any edited population, even if you add natural increase, which brings me to the other group of objections.
Consider two scenarios. The first, and in view of the above much more likely, one is when the edited population is just 2-3 SD above average. If no deliberate effort is made to make this population “ethnically” distinct, and keeping in mind the slow rate of growth, they would just add to their nations’ relative strength. This scenario does not present any unusual features; it’s just more of the same, except a little more so. If such an effort is made, or they somehow evolve a self-consciousness as a distinct group, the first scenario will become a weak version of the second one.
In the second scenario, which is closer to what West Hunter seems to have in mind, the edited population is so much more intelligent (+10 SD?) as to be radically different. I don’t believe such people could successfully conceal themselves in the general population. It is very difficult for an intelligent person to systematically pretend to be dumb. If there are few of them, they will not make much difference overall, no more than geniuses do now (are there many Von Neumanns among the multibillionaires and Davos men? I don’t think so.) If there are more, and they decide to meddle in normie affairs, they will be a market-dominant minority and we know how that story usually ends. Increased intelligence isn’t much help against machetes and Kalashnikovs. So they’d have to secede, but then the reasons against any secession would apply (until they produce force fields or something to protect themselves, a prospect which I don’t regard as very likely from general physical considerations). If they secede internally, they would be vulnerable to mob action; if they secede externally, they’d still be very vulnerable because it doesn’t take a Von Neumann to drop an ICBM or dump a couple of New Panamax ships’ worth of starving Third World people somewhere, and we must not forget that it takes a very large number of people to keep a modern technological society going. Ideally, they would hope to fall off the radar, unobtrusively trade in necessary items and hire as few servants as possible. They could hardly hope to be useful enough to prevent destruction at the hands of competing normie factions. Trade doesn’t stop wars. On their own side, dealing with normies would be a dreary chore even if they liked them. Imagine having to entertain a kitten for a whole day. You wouldn’t hate the cute fluffy thing or wish to kill it, but you would want to leave it a ball of thread to play with and go do something more interesting. Their interests would be unlikely to intersect much with normies; they couldn’t teach them any more than we can teach chimpanzees, couldn’t use them in any but a menial capacity. This is not a recipe for a healthy relationship. They would leave*.
* the summary is bad and the English translation atrocious, but Strugackis describe essentially this scenario, with a fictional “third impulse system” in place of genetic editing. “Initiating” it gives unspecified superpowers in addition to super-intelligence, so ludens don’t need trade or menials.