A clever game outlined below illustrates the basic idea. Players 1,2 and 3 are asked to make (simultaneous) choices to earn prizes (money for the adults and older kids, points for toys for the younger kids).
The sophisticated, rational choice becomes successively more difficult as we from from player 3 to player 1.
Player 3 is simply asked to match a shape. In the case shown, for example, player 3 earns the most by choosing the red square labelled C since it matches the shape of the blue square labelled A. Player 2 earns the most by choosing the color chosen by Player 3. Of course, Player 2 doesn’t know what color Player 3 will choose and so has to reason about Player 3’s actions.
What color do you choose? Player 1 earns the most by choosing the same letter as Player 2 but now must reason about Player 2 which involves reasoning about how Player 2 will reason about Player 3. What letter do you choose?
What do the authors find? First, for both adults and kids either they get it or they don’t.
The ones who don’t make the right choice as Player 3 but then randomly choose when playing either Player 2 or Player 1. The ones who get it, play correctly at all three levels. In other words, almost everyone who reasons correct as Player 2 (1-level reasoning) also reasons correctly as Player 1 (2-level reasoning).
Second, there is a marked increase in the ability to perform k-level thinking between ages 8 and 12 but after age 12 (fifth grade) there is shockingly little growth.
Together the first and second points suggest that k-level thinking is more of a quantum leap than an evolution in reasoning ability.
Third, most adults reason correctly in this simple game but a significant fraction do not. As the authors put it “some very young players display an innate ability to play always at equilibrium while some young adults are unable to perform two steps of dominance.”
Forced ranking is just one of many recent flashpoints that have generated scrutiny of workplace practices at China’s large tech companies. Allegations of overwork, abuse and injury have become the subject of heated nationwide debate. Last December, the deaths of two young staffers at one of Alibaba’s competitors, Pinduoduo, fuelled the fire. (One collapsed on the way home from work. The other died by suicide.)
Such working conditions compelled Chinese software developers to launch a campaign to raise international awareness of their working conditions, christened “996 ICU”. The name refers to a colloquial saying among tech workers that if you work 9am to 9pm, six days a week, as some managers demand, you’ll end up in intensive care. Over the past two years, 996 ICU has crowdsourced allegations of mistreatment from employees at more than 200 Chinese companies.
Even for the large number of engineers working under less brutal conditions, a pervasive sense of the drudgery of uncreative, repetitive tech work has led them to self-identify with manual and agricultural labourers. Many refer to themselves as “code peasants”, and the most common nickname for the tech giants as a group is “big factories”.
Netizens have latched on to a newly fashionable term: “involution” (nei juan). Anthropologist Clifford Geertz popularised the word in the 1960s in work describing unusual aspects of one rural economy in Java, Indonesia. Geertz posited that, over the course of centuries, ever-increasing amounts of labour had poured in but output remained constant. No innovation had occurred. Geertz called this involution.
In 2018 and 2019, the term became popular in China thanks to the writing of Cao Fengze, a PhD student in China’s top school of engineering at Tsinghua University. Cao’s posts chronicled intense, zero-sum competition between classmates for admission to elite universities. He argued the result of so much competition for so few spaces is that students work harder and harder without gain.
“Every effort is nothing other than an attempt to push another snowflake over the edge,” Cao wrote, alluding to the proverb “every snowflake in an avalanche is responsible”. For him, involution is catastrophe.
Cao, whose hundreds of thousands of online followers call themselves the “Cao School”, believes the root of involution is the struggle over limited resources. The metaphor of resource competition is one of the most pervasive in Chinese society. Discussions about the country’s problems tend to culminate in the assertion, “there are too many people in China”
Yet for all the glorification of innovation, a lack of creative work is common for the vast majority of technical employees. “They don’t have the time to come up with ideas or control the direction of their work,” says one ByteDance employee. “They have so much mental load from the tasks they have in the narrow scope they’re given, that they don’t have the latitude to think about anything extra.”
According to a Tencent manager, squeezing employees’ time is often an intended outcome of the management system, rather than an inefficiency. “When we consider why some products succeed, it’s not necessarily because they have better technology,” he says. “It’s because they simply have more people labouring away.” Most developers, he says, are in fact completing repetitive tasks known as “create, read, update and delete” work, such as changing the minute details of a user interface over and over.
“The growth of China’s tech giants has not come from true innovation but from labour intensity. It’s very difficult to automate certain parts of the software sector,” says Xiang Biao, a professor of social anthropology at Oxford university.
“Forced ranking is not about efficiency or fair reward, but about control. This single method destroys all solidarity between peers. It generates obedience and fear towards the person above.”
Tim, the recent Alibaba hire who spent years competing to get there, describes the company’s values as “more like ‘chicken soup’”, slang for a meaningless story told merely to soothe the listener. Another term he uses in describing the relation of older managers to younger employees is “PUA”, short for pick-up artist. He and his colleagues mostly use it to describe managers who over-promise opportunities for advancement to the point of creating a dishonest or even abusive relationship. PUA is another neologism. In China, the linguistic space for complaining about your boss has been expanding lately.
In 2019, the highly renowned Peking University revealed that, of all the graduates that year who had signed university-approved employment agreements, 17 per cent went into China Communist party work or other parts of government, up from 11 per cent four years prior. In total, counting those who had gone into state-administered institutions and state-owned enterprises, three-quarters of graduates had gone “into the system”. In the same period, the proportion of graduates going into private businesses such as tech had been halved.
In two studies, we test how easily people learn an unusual physical or social causal relation from a pattern of evidence. We track the development of this ability from early childhood through adolescence and adulthood.
In the physical domain, preschoolers, counterintuitively, perform better than school-aged children, who in turn perform better than adolescents and adults. As they grow older learners are less flexible: they are less likely to adopt an initially unfamiliar hypothesis that is consistent with new evidence. Instead, learners prefer a familiar hypothesis that is less consistent with the evidence.
In the social domain, both preschoolers and adolescents are actually the most flexible learners, adopting an unusual hypothesis more easily than either 6-y-olds or adults.
Empirical evidence suggests that children may sometimes be better, and
particularly more flexible, learners than adults. Ideas from the literature on developmental neuroscience, machine learning, and cultural learning may help to characterize and explain these developmental differences more precisely.
Among humans, younger learners are more able to learn new linguistic distinctions than older learners (17, 18) and they are better at imagining new uses for a tool (19). Younger children also remember information that is outside the focus of goaldirected attention better than adults and older children (20, 21).
In particular, younger learners are more likely to infer an initially unlikely causal hypothesis from a pattern of evidence. … children might be especially flexible learners.
… also plausible that a playful protected environment may lead to more flexible, exploratory and childlike learning, even in adulthood, and that even in childhood, stressful or resource-poor environments may lead to less flexibility and a more adult-like emphasis on exploitation.
You can and will work at a desk for seven hours straight, routinely.
the discipline of intensive and constant work.
You learn what you don’t know you are learning.
Students join forces on the problem sets, and some students benefit more than others from these weekly collective efforts. The most brilliant students will invariably work out all the problems and let other students copy, and I pretend to be annoyed when I learn that this has happened.
But I know that by making the effort to understand the solution of a truly difficult problem discovered by one of their peers, students learn more than they would by working out some less demanding exercise.
By and large, “knowing how” matters more than “knowing what.”
at MIT, “knowing how” is held in higher esteem than “knowing what” by faculty and students alike. Why?
It is my theory that “knowing how” is revered because it can be tested. One can test whether a student can apply quantum mechanics, communicate in French, or clone a gene. It is much more difficult to asses an interpretation of a poem, the negotiation of a complex technical compromise, or grasp of the social dynamics of a small, diverse working group. Where you can test, you can set a high standard of proficiency on which everyone is agreed; where you cannot test precisely, proficiency becomes something of a judgment call.
In science and engineering, you can fool very little of the time.
An education in engineering and science is an education in intellectual honesty. Students cannot avoid learning to acknowledge whether or not they have really learned. Once they have taken their first quiz, all MIT undergraduates know dearly they will pay if they fool themselves into believing they know more than is the case.
On campus, they have been accustomed to people being blunt to a fault about their own limitations-or skills-and those of others. Unfortunately, this intellectual honesty is sometimes interpreted as naivete.
You don’t have to be a genius to do creative work.
The drive for excellence and achievement that one finds everywhere at MIT has the democratic effect of placing teachers and students on the same level, where competence is appreciated irrespective of its provenance.
Students learn that some of the best ideas arise in groups of scientists and engineers working together, and the source of these ideas can seldom be pinned on specific individuals. The MIT model of scientific work is closer to the communion of artists that was found in the large shops of the Renaissance than to the image of the lonely Romantic genius.
You must measure up to a very high level of performance.
What matters most is the ambiance in which the course is taught; a gifted student will thrive in the company of other gifted students. An MIT undergraduate will be challenged by the level of proficiency that is expected of everyone at MIT, students and faculty.
The expectation of high standards is unconsciously absorbed and adopted by the students, and they carry it with them for life.
The world and your career are unpredictable, so you are better off learning subjects of permanent value.
You are never going to catch up, and neither is anyone else.
MIT students often complain of being overworked, and they are right. When I look at the schedules of courses my advisees propose at the beginning of each term, I wonder how they can contemplate that much work. My workload was nothing like that when I was an undergraduate.
There is some satisfaction, however, for a faculty member in encountering a recent graduate who marvels at the light work load they carry in medical school or law school relative to the grueling schedule they had to maintain during their four years at MIT.
The future belongs to the computer-literate-squared.
The undergraduate curriculum in computer science at MIT is probably the most progressive and advanced such curriculum anywhere. Rather, the students learn that side by side with required courses there is another, hidden curriculum consisting of new ideas just coming into use, new techniques and that spread like wildfire, opening up unsuspected applications that will eventually be adopted into the official curriculum.
Keeping up with this hidden curriculum is what will enable a computer scientist to stay ahead in the field. Those who do not become computer scientists to the second degree risk turning into programmers who will only implement the ideas of others.
Mathematics is still the queen of the sciences.
Having tried in lessons one through nine to take an unbiased look at the big MIT picture, I’d like to conclude with a plug for my own field, mathematics.
When an undergraduate asks me whether he or she should major in mathematics rather than in another field that I will simply call X, my answer is the following: “If you major in mathematics, you can switch to X anytime you want to, but not the other way around.”
Alumni who return to visit invariably complain of not having taken enough math courses while they were undergraduates. It is a fact, confirmed by the history of science since Galileo and Newton, that the more theoretical and removed from immediate applications a scientific topic appears to be, the more likely it is to eventually find the most striking practical applications.
Consider number theory, which only 20 years ago was believed to be the most useless chapter of mathematics and is today the core of computer security. The efficient factorization of integers into prime numbers, a topic of seemingly breathtaking obscurity, is now cultivated with equal passion by software desigers and code breakers.
I am often asked why there are so few applied mathematicians in the department at MIT. The reason is that all of MIT is one huge applied mathematics department; you can find applied mathematicians in practicially every department at MIT except mathematics.
Five-digit compensation sums for egg donors with specific traits, like extensive or brand-name educational backgrounds, are fairly established practice, with some compensation values edging up and above $100,000.
Sperm donors, who often donate continuously over several months, are usually compensated by-donation and can earn $1,000 or more per month.
The lucky ones who make it through the selection process can donate at the sperm bank up to three times each week. Those who hit this maximum make an average of $1,000 to $1,400 a month. Their “terms” expire after about a year to limit the dispersal of their genetic material, and they receive a flat fee for each donation, with a slight scale in compensation based on the volume of the sample.
In his book Turing’s Vision, Chris Bernhardt deftly shows how Turing dashed one of Hilbert’s great ambitions with a masterful proof – in the course of which he inadvertently invented the modern computer.
The Entscheidungsproblem was part of Hilbert’s work to show that the basic axioms of mathematics are logically consistent. To that end, Hilbert sought an algorithm – a computational procedure – that would indicate whether a given mathematical statement could be proved from those axioms alone. Turing decisively showed that there was no such algorithm.
Turing had to first establish a working definition for the term algorithm – to define what it means to compute. Turing looked at human “computers” – people who made computations. The task involves writing symbols on paper, he noted. “The behaviour of the computer at any moment is determined by the symbols… he is observing and his ‘state of mind’.”
Breaking down apparently complex cogitation into simple arithmetical procedures, Turing made computation explicit and eliminated the human element. “Turing’s fresh insight was to define algorithms in terms of theoretical computing machines,” writes Bernhardt. “Anything that can be computed can be computed by a Turing machine.”
That’s why the machines were central to Turing’s paper. To show there were algorithms that Turing machines would run indefinitely and inconclusively was a way of showing Hilbert was mistaken. Turing proved “that there were questions that were beyond the power of algorithms to answer”.
as crucial as the theoretical machines were to Turing’s proof, they turned out to have even more impact in their own right, providing a conceptual model for modern computers. The influence was direct, informing John von Neumann’s pioneering 1945 design for electronic computers, and the room-sized machines that applied his architecture. Like Turing’s machines, the computers used ones and zeroes to encode programs and data.
Turing had to define what an algorithm was; he explained this by breaking complex calculations down into simpler parts. He also defined the very concept of computation, using the concept of ‘Universal Machines’, machines that, he proved, can compute anything that is computable. In 1936, this was all still as theoretical as it gets.
From Chapter 1: “Gödel had completely destroyed Hilbert’s program as it stood in 1920. Nevertheless, there was still the Entscheidungsproblem.”
The Entscheidungsproblem is the halting problem – whether the computer program will finish running, i.e., halt, or continue running forever. “Turing would show that there were questions that were beyond the power of algorithms to answer. He would construct a proof … showing that there was no mechanical set of rules for the solutions of all mathematical problems and consequently that our activities as mathematicians would never come to an end.”
“Turing machines are theoretical models of our modern computers. Everything that can be computed on a computer can be computed by a Turing machine, so Turing’s paper is not just of historical interest; it tells us about what can and cannot be computed by any computer. It tells us that there are limitations to computation, and that there are simple questions that at first glance look straightforward, but are beyond the power of any computer to answer correctly … As Marvin Minsky writes: The sheer simplicity of the theory’s foundation and extraordinarily short path from this foundation to its logical and surprising conclusions give the theory a mathematical beauty that alone guarantees it a permanent place in computer theory.”
why has the trade war transformed into a tech war?
The simple reason is that China could overtake the US as a major economic power by 2030. The US has been unconsciously fueling China’s ascension as a rising superpower by supplying high-tech semiconductor chips to Chinese companies. But recent actions by the Trump administration have limited the flow of chips to China, to slow their development in artificial intelligence (AI) and global domination.
A new report from the Global AI Index, first reported by South China Morning Post, indicates that China could overtake the US in AI by 2025 to 2030.
The index specifies that based on talent, infrastructure, operating environment, research, development, government strategy, and commercial ventures, China will likely dominate the US in the AI space in the next decade.
The tech war between both countries, to get more specific, has also blossomed into a global AI race, the report said.
By 2030, Washington is forecasted to have earmarked $35 billion for AI development, with the Chinese government allocating at least $22 billion over the same period.
China’s leadership, including President Xi Jinping, has specified that AI will be essential for its global military force and economic power competition against the US.
China’s State Council issued the New Generation Artificial Intelligence Development Plan (AIDP) back in 2017, stating that China’s AI strategy will allow it to become a global superpower.
In a recent speech, Xi said that China must “ensure that our country marches in the front ranks where it comes to theoretical research in this important area of AI, and occupies the high ground in critical and AI core technologies.”
The US is the undisputed leader in artificial intelligence (AI) development while China is the fastest-growing country set to overtake the US in five to 10 years on its current trajectory, according to The Global AI Index published this week by Tortoise Intelligence.
The index, which ranks 54 countries based on their AI capabilities, measured seven key indicators over 12 months: talent, infrastructure, operating environment, research, development, government strategy and commercial ventures.
The US was ahead on the majority of key metrics by a significant margin. It received a score of 100, almost twice as high as second-placed China with 58.3, due to the quality of its research, talent and private funding. The UK, Canada and Germany ranked 3rd, 4th and 5th respectively.
the mean score on the three parts of the test — reading, math, and science — for U.S. Asians was 549, which would make them the third highest scoring place in the world, behind only the utopian city-state of Singapore and four rich cities in mainland China. (Scores are on an SAT-like 200 to 800 scale with 500 supposed to be the rich, or OECD, country mean, although the OECD mean was 488.)
At 521, U.S. whites outscored all countries founded by whites (light blue bars) except Estonia. American whites edged Japan and South Korea by one point, which isn’t shabby.
This year, Pisa focused on a key skill–handling abstract concepts and discerning facts from opinions in what they read–because, thanks to ready access to information, reading is more about building knowledge, thinking critically and making well-founded judgments than extracting information.
An average of 8.7 percent of the world’s youngsters scored better than 84 percent in this critical skill, compared to twenty-six percent of Singaporean children, twenty-two percent of Chinese, 13.5 percent in the US, 11.5 percent in Britain and 14.3 per cent in Finland.
Around 16.5% of students in Beijing, Shanghai, Jiangsu and Zhejiang and 13.8% of students in Singapore scored at Level 6 in mathematics, the highest level of proficiency that PISA describes. These students are capable of advanced mathematical thinking and reasoning. On average across OECD countries, only 2.4% of students scored at this level.
Over the last 50 years in developed countries, evidence has accumulated that only about 10% of school achievement can be attributed to schools and teachers while the remaining 90% is due to characteristics associated with students.
Teachers account for from 1% to 7% of total variance at every level of education. For students, intelligence accounts for much of the 90% of variance associated with learning gains.
The largest effect of schooling in the developing world is 40% of variance, and that includes “schooling” where children attend school inconsistently, and staff likewise.
I further argue that the majority of the variance in educational outcomes is associated with students, probably as much as 90% in developed economies.
A substantial portion of this 90%, somewhere between 50% and 80% is due to differences in general cognitive ability or intelligence. Most importantly, as long as educational research fails to focus on students’ characteristics we will never understand education or be able to improve it.
:format(webp)/cdn.vox-cdn.com/uploads/chorus_image/image/63873987/GettyImages_964772232.0.jpg)
Delightful & Distinctive COLRS 