Genetic Enhancement v. Artificial Intelligence

Will robots and artificial intelligences take human jobs? Perhaps but the nature of humanity is not carved in stone. Genetic enhancement (GE) is within a hairsbreadth of reality.

It's true that the practical applications of AI are moving faster than GE but GE has a head start of over a billion years. Moreover, although GE is still impractical, the costs of GE are falling fast. The costs of sequencing Cost_per_genome a genome (shown at right, click to enlarge), for example, are falling far faster than even Moore's Law would predict. Sequencing takes us only part of the way towards H+ but it's an important part.

Genetic engineering already works wonders, even when used haphazardly. My own efforts at GE (I had the help of a PhD microbiologist) have produced two promising NIs. When used in a more controlled manner the results of GE will be even better ("it's still us, only the best of us.")

I used to worry that religious objections would prevent the evolution of H to H+, especially in the United States. But should courage fail us, the Chinese, the Indians, the Russians or perhaps even the Singaporeans will move humanity forward. In this case, the slippery slope works in favor of progress: from avoiding genetic disease towards making improvements will prove irresistible. You can't keep a better man down.

The contrast of GE and AI in the title is meant to remind us that AI is not the only technology relevant to debates about future jobs but the opposition of GE and AI is obviously false. AI is helping to create GE, of course, but it's deeper than that. In the not so long run it's not about computers substituting for labor or even complementing labor, it's about designing labor to complement computers (and vice-versa). Think about how quickly the phone has migrated from the desk, to the hand, to the ear, to the ear canal. The technology to enhance humanity with access to the internet is literally burying itself into our heads, call it I-fi. There is more to come.

And now for some music.


This is like saying that birds could compete with jets if only you messed with the genetics of the bird a bit.

I don't think so.

Genetic enhancement will not give us capabilities much outside the range of what some of us can already do. Not for a long time. We just don't know how to design for major new abilities biologically: the best we will be able to do will be to assemble bits and pieces in new combinations.

The first major uses of genetic enhancement will be to better compete with unenhanced humans in survival, business and killing. Not to "work wonders".

I cringe at the potential. Just look at the monsters that plastic surgery has produced. As a minor example, excessively botox-ed lips have become almost a signature of wealth: just look at any reality TV show ("The Housewives of XXX" franchise comes to mind) - the women on that show in an earlier age would have been grotesque caricatures; now it's a look that seems to be actively sought out.

NI? H+? Are you writing to be read, or just to impress yourself?

My final sentence may be the most important point. AI returns are a better payoff than GE returns in the foreseeable future. First, the payoffs from genetically engineering higher work productivity will take 55-70 years to be fully realised. It's hard to see how the voluntary economic exchange is going to happen here, and I'd like to hear people's opinions on the price for engineering an expected present value of USD 1 million of extra lifetime output. Second, assuming we still can't engineer individual motivation, it's a risky investment in a specialised economy to breed a better scientist. What if she wants to be a family doctor instead?

Genetic Ennancement.

Adds a new meaning to the phrase:

"GE, We bring good things to life"

If promoting fertility rates among professionals to a level proportionate to their representation in the population is "eugenics"

When such a policy is advocated with an explicit genetic basis, that is "positive eugenics" by definition. It's been true since at least the 1920s that educated people had fewer children than uneducated people and this was a matter of great concern for people like Francis Galton and Margaret Sanger just as it is for Lee Kwan Yew today. He has at least on one occasion remarked that Singapore will become a nation of dummies if current trends in fertility rates (high among high school-educated Tamils and Malays, low among college-educated Chinese) continue.

Before we talk about genetic enhancement we need consider why genetic diseases are still so prevalent. And it is NOT because the FDA is holding up the cures waiting for 10 year trials. We cannot reliably manipulate the DNA of individual cells. What we can do is expose many cells to vectors of DNA such as viruses then select for modified cells. If we try to modify hundreds of plants and most are not modified, and some don’t produce viable seeds but a few produce seeds with the DNA enhancement then that is fine for agricultural improvement, but it is not good odds for genetic enhancement of individuals.

...and what was Alex smoking today?


The enhancements are likely to be far easier, when they are developed, for just-fertilized eggs, or even the eggs and sperm separately.

I think that when people consider genetic enhancement, they truly are not giving serious thought to the world and technologies that are developing in parallel. Genetic enhancement, gattaca style, assumes a society that is advanced enough that the protein expressions of the genetic code are understood completely, but their benefits for humans are frozen at whatever level it was at the time of their birth. In reality, it will be an ongoing process and who in a technologically advancing society wants to be stuck with a 20 year old, probably obsolete germline? I think that with the growing biotech analysis and synthesis, aided by IBM Watson's descendants, it will become easier to plug in synthetic viruses/bacteria/ribosomes/phages/nanobots in the blood to do whatever task you wanted your genes and their expressed proteins to do. Whatever your birth genome is, any benefit that is available in the market will be yours.

I think we'll have uploads before we have on-demand genetic modifications. Such modifications require nanite manufacture & dispersion on a scale that unavoidably puts us in serious danger of a gray goo scenario.

Prakash, you forget just how much is not available at any price. How much would Steve Jobs be willing to give up in order to have his aliment fixed? What are the odds that if he poured 90% of his wealth for the last fifteen years into the relevant research that a cure would even exist today?

Even if we assume market availability there is a serious problem. The market provides 777s for private purchase. Why can't I have one?

I'm not sure if I've ever said this before, but I agree with Mike Huben.

It's just a lot cheaper and less ethically constrained to do experiments with AI, than do to experiments with genetic engineering (and even with the Human Genome sequenced, genetic engineering is still in its infancy).

GE is far more theoretically sound at this point though. We know how to genetically manipulate something, but have no idea how to produce an artificial sentience.

Personally, I think it's likely we'll see the augmentation of human capabilities through artificial implants.

Sort of Knowledgeable,

We will do both since the client base is different in the two cases, but it is far more likely that such treatments will be more successful on producing new humans minus the disease than on curing existing humans with the disease, and for exactly the reasons you outlined in your original comment. Of course, I am ignoring the abortion issue in the first case, but the truth will likely be that implantation will occur at a later stage of embryo development (or possibly by the time such genetic manipulation is even possible, not at all), thus allowing genetic sampling much earlier than, let's say 2 to 3 weeks gestation.

I am not against H+ stuff. People can do what they want. There will always be H's though. You don't apply to the government to get babies; you fuck. People like fucking and they like babies so the H's are not going away.

I can't wait for more blonde/blue eyed women with bionic bodies and pea-sized brains

Me, too, but the problem is that they won't be waiting for us.

Although NHGRI has paid my salary in the past - it's a bit unfair to compare de novo sequencing with resequencing...

Or perhaps the graph is just for resequencing of human genomes (or human-sized genomes)?

This should be better labeled - let's try to avoid distorting data...

I tend to share Alex's optimism about GE -- and then I read stories like the one in today's NYT about our utter cluelessness about what makes some people lefthanded.

The teacher is pointing at coffee cup fragments of the said: 'you must feel sorry for this cup, but no matter how you feel sorry for could not make coffee cup return to coach outlet its original shape.

Sorry: the billion year head start doesn't count for very much.

Bruce Sterling covered this, and covered it well, in the 1970s in his book Schismatrix Plus. The Shapers used genetics to advance intelligence and physical attributes, the Mechanist use computers and mechanical exo-skeletons for the the same. They lived together and didn't always get along.

That was a sweet book, the kind that plants ideas you never forget and often refer back to. Like most good science fiction.

Comments for this post are closed