In 2013, Facebook bought the WhatsApp app for $22 billion. WhatsApp was a start up with only 55 employees. Those 55 digital geniuses – certainly people of high IQ – created more economic value in four years than Peugeot's 194,000 employees in its two centuries of existence (it started with Napoleon!).
Sequencing the human genome for the first time took thirteen years (1990-2003) and 3 billion dollars. In 2019, anyone can sequence their genome in a few days for $150 by sending a saliva sample to a biotech company.
Some people believe that La Sexta gives information.
Subscribe to Actuall and thus you will never fall into temptation.
subscribe nowIn 2017, Google introduced its AlphaZero program. Matthew Sandler and Natasha Regan, authors of the book “Game Changer”, tell of him: “No one has explained to him that three pawns equal a bishop or that the king is insecure in the center. The ingenuity learned all that and much more on his own, after playing 44 million games against himself, more than a thousand per second, at first random. Nine hours were enough, thanks to a huge "neural network" (only within the reach of Google) that mimics the work of our brain. 1,500 years of human experience were replicated in a single day of work”. Currently, according to Elon Musk – owner of Tesla Motors, ranked 21 on the Forbes list of fortunes and one of the Silicon Valley gurus – AlphaZero is already capable of defeating any human or computer in any game: “You give it the rules of the game, he reads them, and he can immediately defeat the world champion.”
The comparison between WhatsApp and Peugeot is taken from Laurent Alexandre's book "La guerre des intelligences", which has caused a stir in France. Alexandre (IQ jealously guarded), a surgeon by training, defines himself as a liberal and fervently supported Emmanuel Macron's campaign (IQ 170). The play is fascinating… and terrifying (although Alexandre flaunts technophile optimism). The central idea is the inevitable arrival of an increasingly powerful Artificial Intelligence (AI), and how humanity will have to adapt – co-evolve symbiotically – to it.
Yes, for almost a century a future - sometimes idyllic, sometimes dystopian - of more or less friendly robots has been conjectured in vain. What Alexandre comes to say is that this time things are serious: the development of AI -relatively stagnant between the 60s and 90s- has acquired exponential speed since 2012, when it went from "prehistory" (traditional programs , with manually adjusted algorithms) to the era of deep learning: programs can now “learn by themselves”, in such a way that “AI is no longer programmed, but educated”, fed with tons of data (it is one of the keys to the incontestable hegemony of Silicon Valley companies: they have immense bases of big data, which we constantly expand through our use of Google, Facebook or Amazon). Thus, the result is an AI with an enormous computing capacity, but still “narrow”, specialized in specific tasks (like a deep Asperger, capable of making extremely complex calculations in infinitesimal time, but incapable of preparing a coffee). The third phase -which Alexandré foresees for 2030- will be that of "transversal and contextualizing" AI, no longer confined to specific missions, capable of crossing data from various fields. The fourth – which, according to some, will never happen – will be “strong AI”, that is, self-aware and endowed with a will of its own. Transhumanist Raymond Kurzweil, an engineer at Google, predicts its advent – the Singularity – in 2045. With self-awareness, AI would gain an understanding of the mechanisms that make it possible, and could in turn design new, even more intelligent machines. This would produce an exponential "intelligence explosion" from which transhumanists expect feats such as the solution to aging and death or the colonization of the cosmos. The "strong AI" of 2045 would be 1 billion times more powerful than all human brains combined.
Although it is still only in phase 2, AI is already revolutionizing industry, medicine, the job market... The autonomous car -announces Elon Musk- is ready: the profession of trucker or taxi driver (but also that of driver of Uber-Cabify) would have its days numbered. Many other sectors are going to be affected soon. AI programs can interpret X-rays or skin spots more reliably than the most experienced dermatologists or radiologists. More and more tasks will be automated, not only in medicine: also in engineering, aviation, industrial production... According to a study by the University of Oxford (2013), in the US, 47% of the 702 jobs analyzed have a "high risk" of being affected, and 19% a "medium risk". Nurses have a brighter future than doctors: computers will never be able to transmit human warmth to convalescents, but they will know how to make infallible diagnoses. Artists and chefs have a future.
Some speak of the “third industrial revolution”: the first, at the end of the 18th century, was that of coal; the second, around 1900, that of oil and electricity. The "raw material" of the third is immaterial: information and intelligence. Hence, the history of the 21st century will consist, according to Alexandre, of a "competition for intelligence", which has so far taken the form of a race to capture the best human brains (brain drain), capable of designing and marketing increasingly powerful AI programs. By the way, the only serious rival of the Californian GAFAM (Google-Amazon-Facebook-Apple-Microsoft) is the Chinese BATX (Baidu-Alibaba-Tencent-Xiaomi): the axis of technological power (which will also be political and economic) is it has definitively displaced the Asia-Pacific region; Europe is not in the race: it trains good scientists, but it does not know how to pay them adequately, and they go to Silicon Valley.
Later, the development of AI will force human intelligence to overcome its structural limits, if it does not want to lose control of it. It is Alexandre's central thesis. His considerations on intelligence are very interesting: the left has turned IQ (intelligence quotient) into a taboo, insisting on the fiction that the different levels of academic and socio-economic success are due to discrimination based on class, sex, race... , and that “school failure” is the result of insufficient educational investment, not of intrinsic intellectual deficiencies. The hard truth, however, is that there is a very close correlation between IQ and degree of academic-professional-economic success. And the IQ, according to studies cited by Alexandre, is very unequally distributed among humanity, and 68% is genetic (in the remaining 32% the family environment -more or less stable and instructive- weighs more than the school system). By the way, after growing between 3 and 7 points per decade between 1950 and 1990 (“Flynn effect”), the average IQ plummeted four points in the West between 1990 and 2010 (in Europe the average is 98), while it continues to grow in China, Hong-Kong or Singapore (108). In the new “knowledge economy”, IQ becomes more decisive than ever: “With new technologies, the field of possibilities has been expanded as never before in history. The intellectuals, the innovators, the start-uppers, the managers, the globalized elites move like a fish in water in this new society. […] But the possibility of taking advantage of the digital feast is only available to innovators who enjoy a high IQ. The others, the vast majority, will remain as spectators”. Aware of this, most of the GAFAM mega-millionaires already defend a universal basic income for the soon-to-be idle mass.
Alexandre does not accept the idea of the “end of work”. Because AI is not going to be unlearned – our productive system depends more and more on it: the wings of a Boeing carry hundreds of intelligent sensors; there will be more and more connected objects: "Internet of things" - if it is not going to continue to develop, it launches an ordeal/blackmail to society: if we do not want to become "the Labrador retriever of AI" (and if we want to avoid a dualization social extreme between innovators and spectators), we must: 1) in a first phase (2020-2035), revolutionize the educational system; 2) in a second (2035-2060), expand human intelligence through neural implants, embryonic selection and genetic engineering.
How to install #DirectX in #Linux using #Wine http://bit.ly/d5fPZd
— Steve Roberts Mon Jul 12 22:23:33 +0000 2010
The pedagogical revolution would consist of overcoming the millenary-handcrafted master class system, students meeting in the face-to-face classroom, teachers forced to a "Gaussian" strategy (explanation level adapted to the midpoint of the IQ bell, too difficult for the most clumsy and too boring for the brightest). Alexandre proposes a new school not necessarily face-to-face (or with role reversal: the theoretical classes would consist of MOOC videos from the best teachers in the world, in the style of the Khan Academy; he would go to school to do what is done today in home: “duties”) nor necessarily public (on the contrary, pedagogical innovation will come from the private sector, as long as the state law stops muzzling it). Personalized teaching, adapted to the student's IQ and to their emotional and characterological characteristics, discernible in their genome.
This pedagogical revolution would allow an improvement in the "human intelligence/artificial intelligence cooperation coefficient" in the period 2020-2035, and even an increase in the average IQ, perhaps up to 125 (contradiction with the previous thesis that IQ is mostly genetic) . But from then on it will be insufficient, because the AI will not stop progressing. In his drawing of the period 2035-2060, Alexandre reveals his transhumanist side: NBIC technologies will have to be fully used to multiply human potential so that everyone can reach IQ 220, which is retrospectively attributed to Leibniz or Newton. To "increase" we will use two ways. The first is the enhancement of brain capabilities through neural implants. Elon Musk is already working on it. Science fiction? We already manufacture visual prosthetics that connect to the brain or the optic nerve and solve certain forms of blindness or deafness. We already have implants that allow quadriplegics to send commands through thought alone. By 2035 we will have gone much further. The loophole through which human enhancement will find progressive social acceptance will be compassion: it will begin by proposing brain chips that help overcome Alzheimer's or Parkinson's diseases. Who, except a heartless, could oppose? And if we do not deny his brain extension to an Alzheimer's, why deny it to a child condemned by his insufficient IQ to be left out of the hypertechnological society?
Neuronal augmentation will not suffice: Alexandre also openly defends eugenics. The Chinese are already working on tracking genes related to high IQ. It will be possible in the medium term, either to carry out a massive embryonic selection that allows selecting the smartest individual, or to manipulate the genome to introduce the “gene of genius” (it is not only one, of course). Did you see "Gattaca"?
The price will be the extermination of the remaining embryos... and the abandonment of natural reproduction, replaced by in vitro fertilization. Yes, we will be facing the “baby à la carte”. We will do it? We already do, argues Alexandre: 96% of fetuses with Down syndrome are aborted. It's just about taking eugenics one step further. "By 2100, giving birth to children with an IQ of less than 160 will be as bizarre as it is today and knowingly bringing a baby with trisomy into the world."
Alexandre tries to adopt a neutral and non-evaluative approach. "I do not rejoice at the likely entry into hypereugenics," he says in a footnote. Rather, he insists that all of this is inevitable. He will push in that direction the same egalitarianism that today prohibits talking about congenital talent differences: we will demand “IQ 220 for everyone now!” The redistribution operated until now by the Social Democratic State consisted of smoothing out differences in income that were in turn a consequence of an unequal distribution of congenital IQ. With neuro-augmentation and eugenics, we will be able to correct that unfair nature that rewards some with double IQ than others, without having deserved it. Goodbye to the arbitrary genetic lottery! Welcome to mass production of humans! Finally the same...
It is true that Alexandre concludes by listing three red lines that must not be crossed if we do not want to lose our humanity:
– “The physical body”, compared to those who speak of a future transmigration of our minds to digital media that would provide us with immortality.
– “The individual spirit” planted in physical reality, in the face of those who conjecture, either our future permanent connection to a Matrix-style virtual reality, or the dissolution of the individual mind in some kind of cloud or “global hub of consciousness” ( something that has already started with the Web, a kind of universal mind in which “we are, we move and we exist”).
– Some dose of chance, so as not to fall into total design and predictability, which would make life uninteresting (however, we have seen before how it opened the door to eugenic engineering). It seems to say that although we can choose the IQ of the unborn child, we should still entrust other traits to the genetic lottery.
Those of us who do not want to reach that "brave new world" should start thinking about how to counter. We have almost everything against us: the inertia of the competition between companies and States, which gives priority to the fastest and most daring (whoever manufactures the most powerful AI first will have an advantage over the others); the old human aspiration not to die or grow old, and not to have to work (Alexandre acknowledges that the temptation to become “willing slaves of the AI”, letting it do all the productive work, will be very strong); the impotence of States, overwhelmed by a technological race endowed with its own dynamism (and furthermore, uninterested in it: our rulers are dealing with more important matters, such as Franco's tomb and the gender gap in rugby). "The law of technological platforms already weighs more than the law of Parliament."
But one trump card: how can we be sure that a future “strong AI” would be kind to us? Remember "Terminator". Elon Musk already warns that "AI is much more dangerous than nuclear weapons", and calls for strict international regulation of its development. Darwinian logic predicts that the big fish will eat the small fish. A self-aware AI would seek the elimination of the only being that could unplug it (an unplug that will not be as easy as in “2001, a space odyssey”, since it would be a ubiquitous AI, disseminated in the cloud: Google is already working on a “ red button” for immediate deactivation of systems). The only time two intelligent species lived together on Earth - Neanderthal and us - one of them did not live to tell about it.
Comments
Comments