Anatoly Karlin • March 13, 2018
In his September 1, 2017 speech to incoming Russian schoolchildren, Putin made waves by proclaiming that whoever becomes the leader in AI will become “ruler of the world.” This provoked a variety of reactions, from Elon Musk commenting on his belief that competition for AI superiority will be the likeliest cause of World War III to discussions of the geopolitical aspects of the “control problem” at the more esoteric rationalist venues like /r/slatestarcodex. Many of the reactions were skeptical, citing Russia’s traditional weaknesses at commercializing its inventions. Nonetheless, Bloomberg columnist Leonid Bershidsky, who can hardly be called a Russia optimist, cautioned that Musk’s concerns be taken seriously, citing a range of civilian and military AI applications being developed in Russia.
But here’s another story that happened to unfold on the same day. Back in 2015, Sergey Chemezov, the head of Rostekh state technology corporation – one of Putin’s KGB chums from their time in 1980s East Germany – proudly presented the Russian President with one of his company’s latest “innovatory” offerings: A thin, double-screen, YotaPad-based tablet which was “of entirely Russian make”, meant to be used as an electronic textbook in schools. But they were actually made in Taiwan, and when the devices were distributed to some Russian schoolchildren at the start of the school year, it emerged that they took three minutes to start up, only worked with a stylus, and weighed 1.5 kilograms. According to an investigation by the online journal Znak, the device in question was actually a slightly rebranded version of the American device enTourage eDGe, an outdated and unsuccessful product from 2009 that could be bought wholesale for $20 apiece as of 2015 (you can still get it for $30 on Ebay today). Meanwhile, the official cost of the 8,000 tablets in the trial electronic textbook program was 24o million rubles, which translates to around $500 apiece. This isn’t even very impressive innovation so far as siphoning away taxpayer money into private pockets is concerned, to say nothing of technology.
So which of these stories best reflects the real state of Russian science and technology?
The one in which a technologically adept elite are seriously driving the development of things like strong AI and pondering on its world-historical consequences – or the one in which a clique of kleptocrats pay lip service to innovation while skimming off even the modest resources they bother investing into science and technology?
As per usual, I believe that the best guide aren’t anecdotes, which are the singular of “statistics,” but numbers, numbers, and more numbers in international comparison, as I did in 2006 with respect to China’s scientific/technological convergence with the United States in terms of indicators like published scientific articles published, the prevalence of industrial robots, and the number of supercomputers. I will repeat the same exercise, but with Russia.
The SJR maintains a database of scientific publications by country and subject for the past 20 years.
The Soviet Union in 1986 produced around 7.6% of the world’s scientific articles, which was a quarter of the American rate and comparable to other leading industrialized countries like the UK, Japan, West Germany, and France. In the wake of the brain drain and financial collapse in the wake of the USSR’s dissolution, this figure plummeted to below 3% by the mid-1990s and below 2% by the mid-2000s, in a drop made all the more remarkable by the absence of a “publish or perish” scientific culture in the erstwhile USSR. It was only in 2014 that Russia’s relative standing began to recover.
However, with 73,000 articles published in 2016, Russia remains far below the United States (602,000) and China (471,000), as well the bigger European countries like the UK (183,000), Germany (166,000), and France (113,000). As the 13th most scientifically productive country in the world, it is wedged in between South Korea and Brazil. This is true across the board. For instance, even in the sphere where Russia does best, in the Soviet mainstay of “Physics and Astronomy”, it is still only fourth in the world with 23,000 articles, well behind both China (79,000) and the United States (59,000).
Moreover, even the very modest overall figures conceal a yawning gap in some of the most recent and prospective spheres of modern science. Before worrying about the dangers of AI “eating us” – let alone fantasizing about “sharing this know-how with the entire world” – it would have perhaps served Putin better to first concern himself with the question of why Russia only published 552 papers in the field of AI in 2016, relative to 11,800 in China and 6,700 in the US. Another important sphere that is seeing blistering progress are the genomic sciences, some of whose applications – for instance, human germline engineering for higher IQ – will be world-transforming. Could Russia lead the world in producing “[genetically] spellchecked supermen“? With 690 published papers on Genetics to America’s 13,600 and China’s 9,600, 386 in Biotechnology to China’s 7,100 and America’s 6,400, and 350 in Bioengineering to China’s 6,600 and America’s 4,900, this question answers itself.
The state of affairs in the social sciences is even worse. While Russia’s two (sic) published articles in Women’s Studies in 2016 are nothing to worry about – sooner the converse – that’s about where the happy news ends. Not only do the social sciences suffer from all the other weaknesses of Russian science, but the Soviet legacy there is, if anything, negative value added.
For instance, one sphere that I am personally highly familiar with, psychometrics – the science of measuring mental capacities and processes – was declared a “bourgeois pseudoscience” in 1936, with research in it banned up until the 1970s (though they, unlike the geneticists, seem to have at least largely escaped Stalin’s murderous gaze). Consequently, pretty much all of it had to be re-imported wholesale from the West. While there are now some very good people working on psychometrics in Russia, they have to do it on ageing computers in a creaking building, and financed almost exclusively by European grants.
Far from atypical, this is a steady pattern in the social sciences. To take another example, consider Sinology. Many of the USSR’s leading Orientalists were executed in the late 1930s on spying charges (trumped up ones, I hope it goes without saying). Today, as China expert Alexander Gabuev explained in a couple of articles in Kommersant several years ago, which I summarized in a recent article for The Unz Review (The State of Russian Sinology: Past Chequered, Present Dismal, Future Uncertain), the field of China Studies in Russia is a minnow relative both to China Studies in the West, and to Russia Studies in China. And why should it be otherwise? As of when Gabuev wrote his overviews, the average salary of a docent at the prestigious Moscow State University’s Institute of Asian and African Studies was around $500. Consequently, there is a near total lack of expertise in the country that Kremlin talking points describe as Russia’s “strategic partner.” Though one can cite any number of amazing anecdotes from Gabuev’s articles, I will limit myself to just one. During the Russian-Chinese military exercises “Maritime Cooperation 2012,” the Chinese had nearly 200 young officers with a solid knowledge of Russian at hand to provide linguistic support; the Russians could only muster three translators, and presumably, the Russian GRU intelligence service’s sole China analyst wasn’t one of them. Consequently, not only is the Russian military’s degree of China expertise incomparably lower than America’s, but it is also likely far lower than the PLA’s understanding of the Russian military.
One observes a catastrophic lack of understanding of China across the entirety of the Russian ideological spectrum, not least as regards the extent to which their own country is falling behind.
Scientific Articles: Adjusted for Quality
But if Russia’s raw research output is nothing to write home about, it diminishes to near irrelevance when adjusted for quality.
Here’s one important thing you should know about our world if it were a Civilization playthrough: The Anglo-Saxons have won the Cultural Victory. The majority of cultural output in the world happens in the English language, and this rises to at least 95% so far as science and technology are concerned. The Germans were competitive earlier in the century, before the Nazis (and American demographics) ruined everything, and the Soviet Union maintained a technical mini-civilization partly secluded from the global mainstream, but since its collapse, the Anglo system has become the only game in town.
Most of the really important scientific research gets published in a handful of high-impact factor journals. If there is a proxy for modern day scientific productivity adjusted for quality, and without the generational lag problems that you encounter with the Nobel Prizes, then it is the number of articles an institution or country manages to publish in those elite journals, which are proxied by the Nature Index.
Source: Nature Index, WFC 2016
The US absolutely dominates high-quality research, producing about a third of the world’s total, even though China has gained considerably ground, going from 9% of the global total in WFC 2012 to 14% as of today.
Despite modest improvements since 2012, Russia remains a complete minnow, accounting for less than 1% of elite global scientific research. It is worth noting that it lags China not only absolutely, but in per capita terms as well. In total, Russia produces as much elite level science as does Singapore, Belgium… and the University of Cambridge.
It is hard to imagine any plausible adjustment which would cardinally improve its position. Although it is possible that Russia’s scientific potential is somewhat underestimated by linguistic insularity and its incomplete integration with the global science scene, this is unlikely to be a major factor; since Russia is not actually a world scientific leader in any sphere but a few rather narrow areas of metallurgy and nuclear physics, much of the conversations that take place in exclusively Russian language journals will be outdated and useless. It is also likely that a significantly larger chunk of Russian scientific research relates to military applications than in most other countries, and is effectively “black.” That said, even we assume – very generously – that this underestimation is on the order of 50%, that would still mean that 146 million Russians produce fewer Science Points than the 8 million citizens of Switzerland. Even in Physics, its area of greatest relative strength, Russia barely manages to match Australia; as for the Life Sciences, it is nestled in between Czechia and Argentina.
This analysis is backed up by the performance of individual Russian institutions and scientists.
The most productive (and elite) Russian university, Moscow State University, is in 254th place on the Nature Index, alongside the likes of Oregon State University and the University of Liverpool; fine institutions though they might well be, they do not have a reputation as academic powerhouses. Although Russians tend to complain about the low positions of their universities on international rankings – and I will admit to having once espoused such beliefs myself – it is worth noting that since Moscow State University is 93rd on the latest ARWU Shanghai Ranking and 194th on the THES ranking, it would seem that if anything, the rankings overstate Russia’s performance.
There are a grand total of three Russia-based researchers in Clarivate Analytics’ database of highly cited researchers (of whom only one, Sergey V. Morozov, has his primary affiliation there; the other two primarily work in Spain and the United States). Amazingly, this means that there are as many Russian highly cited researchers in just one American university, U.C. Berkeley – Alexey Filippenko, Igor Grigoriev, Natalia Ivanova – as there are in the whole of Russia! In fairness, Russia’s BRICs rivals Brazil and India don’t do substantially better. However, China has long left its colleagues behind; there are almost 200 highly cited researchers who have their primary affiliation in the Heavenly Kingdom, who are producing 20% of the world’s high-impact academic publications as of 2016.
Russia spends a relatively low but far from catastrophic 1.1% of its GDP on R&D, which is similar to the Mediterranean and Visegrad countries. It also used to have one of the highest concentrations of researchers in the world, with almost 8/1,000 workers employed in R&D, which was higher than the equivalent figures in all the major OECD countries except Japan. Since then, this figure has declined to 6/1,000 even as the average OECD figures went up, so here Russia, too, now keeps company with the Mediterranean and Visegrad. Even so, this was hardly a disaster – the USSR overproduced “researchers” in the same way as it overproduced “doctors” and “engineers”, many of whom would have been mere nurses or technicians in the West. So the thinning out of a good fraction of those fake “researchers” should in theory have been a good thing, assuming that the system was purging itself of dead wood. But the reality was sooner the other way round. Due to the utter lack of prospects in Russian academia, the most talented either continued to emigrate West (with the bulk of that outflow occuring in the 1990s), or went into the private sector.
Many explanations have been proposed as to why Russian science has been in an unending death spiral. Some of the more ideological works cite factors such as the lack of democracy and human rights, and its estrangement from the West – as if Yeltsin’s Russia was a fount of innovation (or democracy, for that matter), while the scientific explosion in modern day China is a mirage (not to mention countless historical counterexamples, e.g. the most scientifically dynamic country in the world prior to World War I was authoritarian Wilhelmine Germany). In Becky Ferreira’s recent profile of Russian science for VICE, one researcher is quoted as saying the following: “If people really only went to countries which do not invade other countries and respect human rights, then they would stick to countries like Andorra or Bhutan… Maybe it sounds a bit cynical, but in my observation, most people in science are driven by opportunities. Regardless of whether such an attitude is moral or not, it is clear that science should be free of any politics.”
No, the real reasons are much more banal: Money, or rather the lack thereof.
According to an exhaustive study of global academic salaries published in 2012, the average Russian academic received 2-4x less money than his equivalents in Visegrad, the Baltics, and even Kazakhstan, and an order of magnitude less than in the developed world.
Source: Paying the Professoriate by Philip G. Altbach et al. (2012).
Here is what the authors have to say about the practical consequences of this breadcrumbs-based approach to scientific funding:
In Russia, young faculty earn approximately 70 percent of the average wage in the workforce; professors’ salaries often fall 10 percent below the average wage of others in the workforce who have completed higher education. In most countries, a middle-class income generally depends on additional employment, either within the same institution, at another academic institution, or in nonacademic employment. All of this added pressure decreases the attractiveness of the academic career and will further deter the “best and brightest” from choosing academe.
Finally, it would be remiss not to mention the astounding prevalence of corruption in Russian academia. According to a Slate article by Leon Neyfakh, the Russian plagiarism detection project Dissernet has found improper borrowing in around 4% of all the dissertations defended in Russia. This doesn’t include plagiarism-free ghostwritten work: Ararat Osipian, a specialist in academic corruption, estimates that around a quarter of all dissertations written in Russia after the collapse of the Soviet Union were purchased.
There have also been private complaints of “ethnic capture” of certain Russian academic departments, primarily by Caucasians. To the best of my knowledge, this is an unquantified phenomenon (though it would not surprise me if this was true, since such a pattern has been confirmed in Italy, where as you go south – which is more corrupt – the incidence of identical surnames within university departments increases, indicating rising nepotism). However, consider the case of the Ingush. They produced six times fewer scientists per capita than Russians during the less corrupt Soviet period; today, their homeland is the highest unemployment, most subsidized region in Russia. And yet they somehow manage to have the highest concentration of postgrads per capita in all of Russia, around 50% more than in second-place Moscow. I will leave readers to draw their own conclusions.
As if the poverty level wages were not enough, the corruption and cronyism also cannot help but discourage the more talented and conscientious from academic careers.
The age when enthusiasts could jerry-rig their own scientific equipment are long gone. You need powerful supercomputers to simulate protein folding, climate change, and the integrity of your nuclear arsenal. You need high throughput sequencers to do serious experimental work in genetics.
But money isn’t any more forthcoming here than it is for salaries.
Twice a year, the Top 500 website compiles a list of the world’s five hundred most powerful supercomputers. Since 2010, China has exploded out of the margins to overtake the United States – as of November 2017, it had 202 top supercomputers to America’s 143, and that included the world’s most powerful supercomputer, the Sunway TaihuLight, which runs on entirely Chinese processors.
Table: Country Share of Top 500 supercomputers in November 2017
Russia’s performance is… rather underwhelming – its measly 0.6% global share of the world’s top 500 supercomputers is equivalent to Switzerland, and lower than that of Sweden, Ireland, and Saudi Arabia.
Nor are the trends encouraging. While there was an uptick in Russia’s numbers of top 500 supercomputers to around 2% of the world total around 2010-2011, those figures have been dwindling ever since.
High Throughput sequencers
This is a very typical picture: A modest cluster in Moscow, while the rest of North Eurasia is a scientific desert.
Russia’s performance in patent applications isn’t too bad by global standards – comparable in per capita terms to the UK and France, much higher than in the BRICS minus China (and it’s not exactly a secret that many East Asian patents are of a spurious nature).
|Patent applications (2015)|
But you can’t realize ideas without money, and despite growing by leaps and bounds in the past decade, the Russian venture capital industry remains tiny from a global perspective.
In 2016, VC funding in Russia (€295 million) was at the level of Ireland (€367 million) and Finland (€324 million) in absolute terms, though a bit above sluggish and overly bureaucratic Italy (€162 million).
And this is relative to Europe, a continent that grossly underperforms relative to its wealth and demographics. According to another source, the old continent had just $14.4 billion worth of VC activity in 2015, relative to $72.3 billion in the United States, $49.2 billion in China, and $8.0 billion in India.
In per capita terms, this means that VC funding in Russia it is at just around 5% of the Chinese level and 1% of the American level.
This expresses itself across the entire range of the hi-tech sphere, but we will just focus on one of the most important and “hip” applications.
Artificial Intelligence Startups
Let’s go back to artificial intelligence, the brains behind the coming wave of automation. How does Russia stack up?
It accounts for 13 of Europe’s estimated 409 AI startups as of mid-2017…
… or just 0.7% of the world’s 1951 total.
The US enjoys near total dominance in this sphere – with more than a thousand AI startups, it accounts for more than half of the world total. China is assuredly moving into second place position, hurtling past Japan and the major European countries.
Meanwhile, Russia is once again in the company of countries like Sweden, Finland, and Switzerland, who have less than 10% of its population.
According to a just released report by CB Insights, in 2017 China leapfrogged past the US to dominate global equity funding to AI startups. They are fast becoming the only two relevant countries in this sphere, with countries that are not China or the US accounting for a mere 13% of the global total.
For all the lunacies of the Soviet economic system, their planners did at least appreciate the importance of robotics and their role in enhancing productivity in manufacturing.
Source: International Federation of Robotics – World Robotics 2005
At the time of its collapse, the USSR had an operational stock of around 60,000 multipurpose industrial robots. In practice, this is a very inflated figure – a large percentage were simple, even hand-operated tools that would not have been counted as industrial robots anywhere in the capitalist world. Still, the Soviet level of industrial robotization in the 1980s was at least broadly comparable to the developed world, and several orders of magnitude higher than in a China just emerging out of its Maoist slumber.
Until the early 2000s, the publicly available databases generally didn’t even include the numbers of industrial robots in Chinese factories, so small and insignificant were their quantities. But from the late 2000s, the robotization of Chinese industry began to explode. As of 2016, it accounted for about 30% of the world industrial robots market, overtook Japan to become the country with the world’s largest operational stock of multipurpose industrial robots, and leveled with the United Kingdom in robot density.
Conversely, it has since become hard to even find any specific data for Russia… According to the World Robotics 2013 – Industrial Robots report, Russia had an operational stock of around 1,771 multipurpose industrial robots as of 2012.
Source: World Robotics 2013 – Industrial Robots
Source: World Robotics 2013 – Industrial Robots (2011 data)
Russia’s (total!) figures are slightly higher than in Slovenia, but lower than in Slovakia. In per capita terms, the rate of robotization per worker in Russia in Russia hovers between that of India and Iran, and is far behind middle-income industrial countries like Turkey, Brazil, and Mexico, to say nothing of a China fast gallivanting its way up to the levels of its super-automated East Asian peers.
Source: International Federation of Robotics – Feb 2018 press release on robot density (2016 data)
The state of affairs today isn’t any better. A 2016 report from the Russian robotics association NAURR presents two different datasets about the rate of introduction of new robots onto the Russian market in recent years.
Sales of robots in Russia, 2005-2014
Graph: World Robotics 2015
Sales of robots in Russia, 2011-2014
Although they diverge somewhat in their assessments, the underlying picture is clear – only around 500 industrial robots are introduced into Russian industry per year as of 2014, accounting for a dismal 0.25% of the global total. This is about thrice less even than Brazil’s 1,300, and two orders of magnitude lower than in China, where 57,000 were sold in the same year. It is likewise highly unlikely that Russia saw any improvements since 2014, considering that this was when it fell into a two year recession.
According to the NAURR report, the top five countries for scientific publications about robotics are the United States, followed by China, Japan, Germany, and South Korea. While figures for Russia aren’t given, it is probably safe to say that it is about as irrelevant here as it is in AI.
It would also be worthwhile to briefly survey the machine tool industry – a sector of special interest not only because of this its inherent technological sophistication, but also because of its strategic importance as the only part of the industrial economy that actually reproduces itself and makes everything else possible.
Source: Gardener Research – World Machine Tool Survey 2016
As you might expect, the lists of countries that dominate industrial robots and machine tools production – Japan, Korea, the Germanic lands, Italy, and increasingly, China – are highly similar. Russia is not an exception, accounting for just 0.6% of world machine tool production.
As with elite level science and robots, China has left Russia in the dust not only in absolute, but even per capita, terms.
share of machine tool production 1913-1995 (Brown – USA; Black –
Germany; Green – Britain; Red – Russia; Purple – Japan; Yellow – China)
The Russian Federation also massively lags even the late USSR. As an autarkic military-industrial empire, the USSR understood the necessity of being able to make the machines that make all the other machines, bequeathing the Russian Federation with 2.8 million machine tools in 1992 upon its dissolution. Since then, that machine tool stock has inexorably depreciated, and as of 2013 constituted just 760,000 pieces, with the average age almost doubling from 12 years to 21 years.
Since the end of the USSR, it has become clear that a chasm has opened up in in terms of scientific and technological output between Russia and the developed West.
This video juxtaposing the lumbering Robot Fyodor versus the agile Atlas built by Boston Dynamics seems like a good metaphor for what is perhaps the single biggest failure of Putinism in the past 18 years.
In comparison, any successes or failures in the Middle Eastern military adventures that pundits and commenters obsess over are basically irrelevant.
This is not to say that things are unremittingly bleak and getting worse.
The government has a strategic goal to get five of its universities into the global top 100 by 2020, to which end it has lavished significantly greater funding on its 21 most prospective universities. Consequently, academic salaries have greatly improved since 2013, at least in the elite institutions. They still don’t compare to the caviar feasts served up to Western professors, but at least they now constitute solid hunks of bread instead of the measly crumbs that were served up before.
There’s no very obvious reasons why Russia can’t succeed more at science. The average IQ relative to British norms is around 97, which might fall significantly short of Germanic and Anglo-Saxon (native!) averages, but isn’t really out of place relative to Mediterranean or East-Central European standards. Moreover, there are signs that Russia continues to enjoy a Flynn Effect, and besides, surely any minor disadvantage with respect to raw IQs is cancelled out by Russia’s traditionally very strong performance in international programming and mathematics contests.
Meanwhile, as regards industry, it is worth pointing out that Russia does consume around 2.7% of the world’s machine tools – it is, after all, the world’s eighth (or so) manufacturing power, not the gas station masquerading as a country of John McCain’s imagination. Infrastructure – roads, rail, airports – has genuinely gotten much better in the past decade, and with post-Soviet inflation finally tamed, Russia looks set for fairly vigorous growth.
But the problems holding Russia back are severe, and possibly intractable.
There remain strong financial and ultimately institutional barriers to unlocking Russia’s scientific potential. Putin and his clique seem to prefer lavishing resources on expensive status-signalling sporting events and white elephants as opposed to serious science and supercomputers. The former burnishes his prestige amongst simple people and provides endless opportunities to siphon away money to his Ozero chums – the latest lunatic project is to built a bridge for $10 billion to Sakhalin and its 500,000 people (a contract won by Arkady Rotenberg – who else?), which is about what the federal government spends on the Ministry of Education in a year – while the latter will only cause political trouble.
Ending corruption within academia would likewise seem a quixotic endeavor. While one can say much more on this topic, consider that PhD’s are no less a status symbol for the Russian elites than Mercedes cars and English boarding schools for their children. High-flyers found to have plagiarized their doctoral dissertations include no less than one in every nine members of the State Duma, and for that matter, Vladimir Putin himself. Waiting for these people to solve the problem of academic fraud is about as realistic as expecting them to solve corruption, or training foxes to guard hen houses. Nor is it possible to imagine a serious response to ethnic nepotism in academia in the land of Article 282, where you can be prosecuted just for arguing that the Caucasian republics should get fewer federal subsidies.
Finally, the absurdly low levels of robotization in industry raise serious questions about Russia’s political economy and its economic future. Why are Russian businesses loth to make serious moves towards automation in industry, even though Russia is, despite everything, a reasonably high IQ and well educated country? Is it because these require big capital investments that they are not willing to risk because of what they perceive as Russia’s environment of legal nihilism? It is correlated with Russian elites being the most apatride of any major civilization?
The importance of finding good answers and good solutions to these questions will only increase in the coming years and decades, as industry moves towards greater and greater automation. It seems likely that the countries that will be most successful at this will also be those who are succeeding at robotization today. Will Russia fall into a low-income trap where low wages preclude automation, and low automation preclude greater productivity and wages? At any rate, it doesn’t seem to be the case that anyone in Russia is seriously thinking about this, at least beyond empty electoral slogans – even as Putin runs for his fourth and hopefully final term, his promise to create 25 million hi-tech jobs during the 2012 Presidential elections has been conveniently forgotten.
Now this is not to say that the problem is with the Putin regime and that its removal will improve things. The pro-Western liberal elites are at least as rapacious as the kremlins, no less authoritarian in spirit, and far less patriotic to boot. Although this post was primarily about Russia, feel free to go back through the hyperlinks and study the case of the Ukraine, where liberal “lustrators” have repeatedly won; it is almost Sub-Saharan Africa so far as advanced science, native hi-tech (as opposed to offshored work), and any sort of capital-intensive manufacturing that wasn’t bequeathed to it by the USSR is concerned. Even the Visegrad and Baltic nations don’t have much to write home about. While most of them – especially, Czechia, Estonia, and Poland – do substantially better than Russia on most of these metrics, they still hugely lag the developed West and have been left behind in the dust by the Chinese juggernaut.
I don’t propose any great over-arching solution to these problems. “More money for RAN, less money for the Rotenbergs” might be a nice slogan, but as they say, the devil is in the details.
However, a solid start would be to look at the statistics and acknowledge that a very big problem exists, which, unresolved, will continue to degrade Russia’s economic, industrial, and eventually military competitiveness.