Milky Eggs

Underrated aspects of the “genetic singularity”

We now know that human traits are largely:

1. additively composed of thousands of non-pleiotropic genomic variants,
2. largely independent without noticeable tradeoffs between different traits at the margin, and
3. increasingly predictable from genetic data as models improve and biobanks get larger.

For simplicity I will define the “genetic singularity” as the point when people become freely able to explore the high-dimensional space of human genetic variation without restriction, easily selecting or editing embryos with +5 standard deviation boosts to tens or hundreds of desirable traits. In my view, this is inevitable and is perhaps 1-2 decades out.

There is much that one could ask, but here are some questions that appear relatively underexplored:

1. Is this accomplished by iterative embryo selection, generation of egg cells from somatic cells, or multiplexed CRISPR-Cas9 gene editing? Whoever bets correctly on the right technology stands to profit a lot.
2. What happens to medicine in a world where everyone is +5SD more resistant to every major human disease? The typical person will be superhumanly strong, have effectively no buildup of arterial deposits, physically and mentally age extremely slowly, etc. Certainly something will end up killing you eventually, but one can imagine that a great deal of modern biopharma development becomes effectively irrelevant. And what is it that we expect to actually kill you in the end?
3. It is not always straightforwardly obvious what trait to optimize for, e.g. there is in general a sensitivity vs. specificity problem in immune response. What are the traits where it’s not immediately apparent what direction to select in or how far to go in that direction? It is probably possible to anticipate these and to start disentangling the relevant issues.
4. If you can multiplex in 100 base pair edits, is there any real difficulty in multiplexing in 1,000 or 10,000 (or more) edits? I.e. if we can CRISPR out 150 IQ babies or 200 IQ babies, what is stopping us, if anything, from CRISPRing out 500 or 1,000 IQ babies? Here I am less familiar with the underlying technology here. With embryos you are still subject to distributional issues but direct editing seems much more unrestricted.
5. How does the physical world reconfigure itself when everyone is 100x as athletic by default? Perhaps people will naturally walk much faster! What sorts of unusual, hyper-optimized (by our standards) travel modalities are enabled when people are 100x as attentive and careful & better at multi-tasking?
6. What is enabled by understanding the fine structure of human cognitive ability? At a minimum every person will have a much better sense of their comparative advantages over others, although hypothesizing about job market efficiency feels a little parochial when (conceivably) we might invent real AGI within the next century.
7. How does life change when everyone has an eidetic memory like von Neumann? Aesthetic appreciation will be more common ー takes less time and effort to read a book! Perhaps old languages will bloom once more if memorization of tens of thousands of vocabulary words is trivial. Complexity of new art may increase dramatically, with much more “intertextuality” than we are now used to.
8. The moral dimension is also interesting ー gifted children are known for being unusually empathetic and altruistic compared to same-age peers. Unsurprising if you think about it, because empathy and altruism require you to form (relatively complex) models of other people’s mental states, reason abstractly about moral principles, and so on. If humanity separates into multimodal distributions of “native” vs. “enhanced” peoples, will we perhaps see the greatest differences as not those of pure health or intelligence but in fact greater versus lesser degrees of moral sensibility?

Exciting times ahead!