Pursuant to the discussions at the last big thread, I am making a quick post with my assessments of how realistic various “transhumanist” spheres of technology are.
Has been, is, and will continue to happen – and will affect lower-IQ occupations sooner. Since it takes IQ to design and maintain the robots, this will further privilege cognitively gifted nations (see Our Biorealistic Future). Inequality will also soar. Eventually, some kind of universal basic income will be probably be necessary to avoid a cyberpunk dystopia. Well, we are sort of drifting there anyway, but at least a dystopia without the widespread poverty.
Radical IQ Augmentation
This involves several methods:
- Embryo selection for IQ: Razib Khan/Charles Murray estimate: 3-5 years (in 2016). Of course regulatory requirements will probably take a long-ass time to get sorted out (e.g., the infamous FDA), so add another 5-10 years. Though this will probably be much quicker in “offshore” style areas with fewer regulations. Calculations of possible IQ gains from a 2014 paper by Carl Shulman and Nick Bostrom (current practice is to select from 10 embryos). See gwern on practical costs/benefits for individuals.
- Embryo editing for IQ (e.g. via GWAS/CRISPR): For instance, this Twitter user (a prominent psychometrician IRL) points out the method is unreliable and produces a lot of errors – big problem since there’s hundreds of genes affecting intelligence! OTOH, CRISPR is improving fast. Mike Johnson, who has studied the trends although doesn’t work in the field, in 2015 predicted that a “dedicated billionaire with scant regard for legalistic regulations could start genetically “spellchecking” their offspring within 5-7 years.”
Using widespread embryo selection for intelligence, a nation where this is widely implemented can go from the level of West Virginia to Massachusetts, or Southern Italy to Northern Italy, within 1-2 generations – nice, but not immediately transformational. This is not the case with CRISPR/GWAS-enabled IQ augmentation; if this starts around 2030, we can have the first augmented generations growing up by 2050, and beginning to transform society from 2060. Due to the “smart fraction” effect, this will have a substantial impact even if 1-2% of the population reproduces does this. Obviously, barring other game-changing scenarios such as superintelligence, the first countries to implement this will gain an ultra-competitive advantage.
Even though this isn’t often discussed in transhumanist circles, this is the topic that personally excites me the most, because it is the one technological sphere that is both already visible on the horizon, and will have transformational ramifications.
Incidentally, the commenter Alexander Turok has recently finished (and advertised on the comments to this blog) a book that looks at the ramifications of genetic selection for IQ along the lines of Hanson’s Age of Em.
According to a paper by Sandberg, median estimate time for the emergence of ems (emulated minds) is 2059; the age of ems will very likely lead to superintelligence within a couple of years anyway. Median prediction for the emergence of “high level machine intelligence” according to various groups of AI experts clusters around 2040-50. This remains in line with Ray Kurzweil’s classic 2045 prediction in The Singularity Is Near.
With respect to ems, my main concern is that of the “unconscious zombie”. The planetary (and possibly galactic) extinction of consciousness would appear to represent an epochal loss in value – indeed, one indistinguishable from full-scale extinction. I would second Mike Johnson in his belief that it would be very much advisable to solve the consciousness problem before allowing mind uploads to go ahead.
The dangers of superintelligence are well-known. Once we start to approach those technological milestones, it may be prudent to create a UNATCO-style global organization for AI control until, to borrow from Trump, we “can figure out what the hell is going on.”
So basically Deus Ex-style neural augments. I (or rather, Bostrom) explained why meaningful brain-machine interfaces will be very hard to implement.
However, if it was to be decided by a global singleton that machine superintelligence is too risky (or if the problem proves too hard in general), then one can certainly see these technologies getting developed in coming centuries, perhaps during the Age of Malthusian Industrialism.
Radical Life Extension
I think this will be a lot harder than Radical IQ Augmentation for a very simple reason.
The natural “range” of human intelligence spans about 7 standard deviations to either side of the Greenwich mean of 100 that prevails in the developed world, of which perhaps 5 can be realized via GWAS plus gene editing. Even today, differences of a few points in average IQ between various regions and countries can already have striking effects on socio-economic development; now imagine that some regions start converging to mean IQs of 175, and you already have a transformation deeper than anything since the Industrial Revolution, if not the appearance of agriculture.
In contrast, the oldest humans only live to 125 years or so. A society where the average person lives to 110 will not be radically different from a society where he or she lives to 85 (the longest-lived societies today). Much deeper, perhaps trans-species genetic tinkering (e.g. drawing from whales or naked mole rats) will be required, and/or a much deeper understanding of ageing pathways or at least how to keep them repaired, as in Aubrey de Grey’s SENS program. Aubrey believes we will achieve mouse rejuvenation by the early-mid 2020s, but as he himself pointed out in his book Ending Aging, mouse models are not obviously extensible to human ones.
In a 2017 Reddit AMA, breaking a long tradition of not giving any quantitative predictions, Aubrey de Grey estimated that a 25 year old man has an 80% chance of reaching “longevity escape velocity.” So perhaps call this around 2050?
I don’t see this panning out for economic reasons; maintaining a base on the open ocean or Antarctica is trivial relative to a Mars base or a Venusian cloud city, to say nothing of the stupendous challenges involved in interstellar exploration.
If we really want to make self-sustaining space colonies at least theoretically feasible – that is, to satisfy Musk’s vision of a second home away from home to increase the chances of humanity’s survival if some unprecedented disaster was to wipe out life on Earth – then we need radical measures. First, we need to send the atomophobes to concentration camps, as Thorfinnsson energetically recommends. Then we need to start work on nuclear pulse propulsion – the only feasible method of sending huge masses of material into space with technology that has been available since the 1960s.
But I don’t see this happening, and frankly I don’t see anything really exciting happening in space on timescales shorter than centuries. The only potentially profitable business enterprise seems to be in mining asteroids for minerals that are extremely rare on our own planet.
Seasteading – sorry to disappoint the libertarians, but I don’t see the economics of this ever working out.
Artificial Wombs should be feasible in a decade. Probably not going to be widely implemented for social/legal reasons.
Cryptocurrency will continue growing in prominence, but I do not see them ever having transformational effects. Reasons why.
Nuclear Fusion is always 20-50 years away. I assume this will remain the case.
Nanotechnology, in Drexler’s sense of self-assembling nanites, remains a pipe dream so far as I know (happily, same goes for the “grey goo” extinction scenario).
I support virtually all of these initiatives, though with some strong reservations on ems/superintelligence.
First, because they’re really cool.
Second, because if they can happen, they will happen anyway, and national obscurantists (typically leftists or religious fundamentalists) dragging back progress will only doom their own countries to irrelevancy.
And it is then the ultra-competitive countries that did follow through on them that will set the rules anyway: “Victory needs no explanation, defeat allows none.”
Third, because if there is are no great breakthroughs, the Frito Effect will carry the world into a dysgenic miasma of technological stagnation that will last for many centuries, until the new Malthusian conditions recreate elite intelligences. There will be a lot of suffering for no discernible gain.
PS. Before the usual sovoks come crawling out of the woodwork telling me to crawl back to Silicon Valley, I would like to point out that transhumanism was of course invented in the Russian Empire.