当时只觉得不服不行。然而不久就听说古狗要将波动卖了，觉得有点意外，但没深究。周末小道说丰田科研要买，合约没最后定，但所有细节都已定：

《技术内幕》Toyota is closing in on a deal to buy Google's robotics company Boston Dynamics, and the 'ink is nearly dry'

 

此报道言及古狗忍疼割爱的背景：裴矩（Larry Page）和高层班子跟波动部门不但不合，简直要闹翻了。追踪了一下彭博原来的报道：
《彭博》Google Puts Boston Dynamics Up for Sale in Robotics Retreat
《钛媒体》造出逆天机器人Atlas的波士顿动力，谷歌准备把它卖掉了

原先觉得意外，因为裴矩是科学科班l出身，对领前技术最关注，常常压下巨款。波动属于烧钱的玩意儿，但技术很先进，不过机器人好的不好做，你见到的技术最尖端的大多属于自动化那种，实际上比古典力学大概高一个档次，有人工智能，但基本线性，科幻电影里的，都是模拟的。采用新的人工智能加上最新机器模拟骨骼结构，实际上很难做，机器模拟，实际上是三维空间+三维空间，看似灵活，目前还是机器思维。

 

美国最出色的大学'机器人竞赛，美国国家航空航天局（NASA）主持。难

机器人用上仿生学（Bionics），更难，最先进的是人工手脚，已经很神了。

 

这一切，都是美国在做。

像如此领先而且有占据行业阵地的技术，裴矩是不会放弃的。裴矩是烘烘，看来世界上烘烘还有其他人；裴矩厉害，有能力有资格有本钱，其他人也厉害，有不服的。

 

《技术内幕》Boston Dynamics employees were frustrated with Google's plan for a household robot

唉，俗话说一山不容二虎。裴矩既要领先，又要钱，也不好当家。

 

 

《彭博虻报》Google's Chaos Theory

原来古狗内部也是乱成一团，乱花钱，不过自己钱多，别人更差劲，也没事儿。

 

---

最先进的技术，烧钱的玩意儿，无法无天的班子，大家都敬而远之，像个烫芋头。这正合中国的口味，中国人就信这儿。只是美国怎么也不会让中国拥有这般的技术的，所以中国企业不论花多少钱也没戏。下一个买主自然是日本了，合情合理。

 

尽管难，没技术，就得自己研发，中国眼巴巴望着先进技术也百多年了，不是不想，近两年来李克强大肆宣传的“转型、升级、工业4.0”，也说得很响，热情投入也不差，然而实情却没多大进展，这是我几个月前观察的结果（中国机器人现状):

国内能静下心来踏踏实实做的，不多，深入人心的是山寨精神，浮躁、投机心太重。只是如资中筠所说，有些东西靠花钱是买不回来的。

资中筠：思想不是用钱可以砸出来的
 

美国人所说的“创业精神”（entrepreneurial spirit），中国现在也很盛了。资中筠却点明了美国为什么强大，都是“软”实力，自身强，实力就体现出来了。要做到这地步，我想提的两个方面是：产权保护和严厉执法，思想自由。

 

知识产权的关键，现在国内已经有足够的认识，山寨不是占了外国人的便宜，而是坑了自己，没人愿意做难得，长期投资的发明。目前不但立法不齐全，执法更性如儿戏，层层干预，在此环境下，大家说“创业”，在技术上有领先、跳跃性的发明不多，跟政策吃饭为主。

 

思想自由比较复杂。有言论自由，媒体自由，我想说“思想自由”。思想自由是中国环境下无可奈何的选择。言论自由、媒体自由在中国都是禁区，大家说闲话，行，但有界线。既然言论自由、媒体自由难，那么至少允许学术界的思想自由。可是中国知识分子历来是政府官僚的一个主要参与者，是利益的得益者之一，往往与统治高层达成共识，以社稷为代价。对于独立、系统之外的思想，政府压制是常规。这一来，这个社会思维单一，不思进取，日趋保守，结果。

 

龙科多：中国科学素质基准，和美国的有什么不同

作为一份高规格文件，近日出台的《中国公民科学素质基准》却出现了“力是自然界万物运动的原因”等低级错误，以及要求公民“知阴阳五行”等诸多值得商榷的提法。有媒体指出，作为“十年磨一剑”的产物，对比研究稿、征求意见稿，可以发现《基准》是越搞越不靠谱。
媒体称《基准》编排很奇怪的根本原因，在于制定者一定要跟着2006年国务院文件《科学素质纲要》走。复旦大学物理学系的施郁则撰文推测，之所以发生一系列失误进入《基准》这样的系统性问题，可能是因为有关部门有一个误会，以为“自然辩证法”等相关专业就是指导科学普及的对口专业。施郁称，他本人不认同“哲学可以指导科学”的说法。

难道觉得奇怪吗？

 

“思想自由”，是在默认言论自由，媒体自由是个难题的前提下允许学术界拥有“思想自由”，能畅所欲言。如果知识分子不能说真话，研究有禁区，研究的结果要符合“马列毛”“邓江习”思维，那知识分子就不再是知识分子，而只是御用喉舌，不能思考，任何创新的机会都会被扼杀。说来也怪，我说没这“思想自由”，科学革命、科学创新的努力也是徒然，不是说毫无结果，但不可能领先世界。没有学术界的“思想自由”会导致整个社会在所有领域僵化、萎缩，大家只会跟风、马屁，向钱看，现在学术界的虚假腐败不但无法杜绝，还会恶化。

 

 

这个中国厉害（科技三会召开 习近平提科技事业发展三步走目标：2049成世界科技强国）

 

 

领先好几十年的技术

 

连贵州也不落后，要“造一个”出来

贵阳要建“中国数谷” 外国科技巨头也想分一杯羹，李克强打气

 

这也有点儿虚了：

 

 

在《新华社》英文版，李克强下午单独作报告提到“自由”（Chinese Premier Li Keqiang has promised more freedom and financial support），然而中文版用词不同，不知道是翻译没翻好，还是有意。

 

不过基础进展不大，使用中国历来不差。近日《金融时报》以中国机器人的兴起对全球制造业做了个观察（见下），《观察者网》给了个翻译：

《观察者网》本·布兰德：中国掀起全球机器人革命

 

这里说的“革命”，是生产手段的革命，尽管中国不拥有机器人的知识产权，但并不限制中国对整个制造过程进行改革、甚至“革命”，来大幅提高生产效率。效率高，竞争力就大，中国有希望将制造业成本日益增高，质量相对下跌而导致的制造业外流局势扭转。这也很关键。

 

这看上去是潮流，但大家现在为什么都急着做？

 

人民日报记者调查用工成本：员工挣7300元，企业要掏1.6万

“不涨工资，招不来人；工资涨上去，企业吃不消”
用工成本压力不小，但企业都表示不敢不涨工资。“焊接工种技术要求高、工作强度大，年轻人大多不愿意干，再不涨工资，更是招不着人”“就是年年涨工资，核心员工离职的情况还是很多”
一名月薪1万元的员工，扣除个人缴纳的社保及个税等，拿到手的部分是7300元，企业总计为其支付约1.6万元
“机器换人”渐成潮流。企业普遍认为工资已不可能降，关键要提高劳动生产率

目前中国制造成本不断上升，总的因素较复杂，但人工成本是个首当其冲的原因。中国真是怪，人口过多，但制造业基建（楼房交通等）人口却日趋短缺（创业的人口，那是火），是人口老化的恶性结果（大众却未因此大幅提高生活质量，经济学家悲叹为“未富先老”，也许是发展得太快了），结果中国制造业面临两个方面的挑战：低成本，如越南、南亚、东南亚，非洲，都与中国竞争；高效能、高质量，如美国，自动化程度突飞猛进，质量越来越高，能源费用日益下降。两者都造成中国竞争力下降，难以竞争，光靠人民币贬值是救不了的。所以这机器人生产方式的大幅采用，会对中国重新占据制造业有关键的作用。

 

神乎其神的深圳怡丰机器人的国内自动泊车机器人，技术很好，但未必是吹嘘的“全球首创“。《观察者网》转载《雷锋网》报道

 

目前中国这种产业更新是件好事，有必要，会有帮助。但除了“引进先进技术”，还得自己拥有先进技术，才能真正领导世界制造业，而最关键的，是建立思想、研究的大环境，这与习近平僵固思想xia相冲突。这是中国的另一个挑战，不知习近平在叹息为何中国外交无人才之际是否意识到他自己是导致此种局面的促进者。

 

《金融时报》Artificial Intelligence and Robotics
China’s robot revolution
Across China, factories are replacing humans with robots in a new automation-driven industrial revolution. And the effects will be felt around the world

APRIL 28, 2016 Ben Bland

The technical staff repair the robot in YingAo Kitchen Utensils Co., Ltd.

The Ying Ao sink foundry in southern China’s Guangdong province does not look like a factory of the future. The sign over the entrance is faded; inside, the floor is greasy with patches of mud, and a thick metal dust — the by-product of the stainless-steel polishing process — clogs the air. As workers haul trolleys across the factory floor, the cavernous, shed-like building reverberates with a loud clanging.

Guangdong is the growth engine of China’s manufacturing industry, generating $615bn in exports last year — more than a quarter of the country’s total. In this part of the province, the standard wage for workers is about Rmb4,000 ($600) per month. Ying Ao, which manufactures sinks destined for the kitchens of Europe and the US, has to pay double that, according to deputy manager Chen Conghan, because conditions in the factory are so unpleasant. So, four years ago, the company started buying machines to replace the ever more costly humans.

Nine robots now do the job of 140 full-time workers. Robotic arms pick up sinks from a pile, buff them until they gleam and then deposit them on a self-driving trolley that takes them to a computer-linked camera for a final quality check.

The company, which exports 1,500 sinks a day, spent more than $3m on the robots. “These machines are cheaper, more precise and more reliable than people,” says Chen. “I’ve never had a whole batch ruined by robots. I look forward to replacing more humans in future,” he adds, with a wry smile.

 

Across the manufacturing belt that hugs China’s southern coastline, thousands of factories like Chen’s are turning to automation in a government-backed, robot-driven industrial revolution the likes of which the world has never seen. Since 2013, China has bought more industrial robots each year than any other country, including high-tech manufacturing giants such as Germany, Japan and South Korea. By the end of this year, China will overtake Japan to be the world’s biggest operator of industrial robots, according to the International Federation of Robotics (IFR), an industry lobby group. The pace of disruption in China is “unique in the history of robots,” says Gudrun Litzenberger, general secretary of the IFR, which is based in Germany, home to some of the world’s leading industrial-robot makers.

China’s technological transformation still has far to go — the country has just 36 robots per 10,000 manufacturing workers, compared with 292 in Germany, 314 in Japan and 478 in South Korea. But it is already changing the face of the global manufacturing industry. In the process, it is raising broader questions: can emerging economies still hope to follow the traditional route to prosperity that the developed world has relied upon since Britain’s industrial revolution in the 18th century? Or will robots assume many of the jobs that once pulled hundreds of millions out of poverty?
…

Chen Conghan, deputy manager at Ying Ao: ‘These machines are cheaper, more precise and more reliable than people’

China’s spending splurge on industrial robots has its roots in a pressing economic problem. From the 1980s onwards, as Beijing’s Communist rulers opened up to global trade, the country’s huge, cheap workforce helped make it the world’s biggest exporter of manufactured goods. Breakneck economic growth lifted hundreds of millions of Chinese out of poverty and transformed swaths of the country, as workers migrated from the countryside to the city. But a growing middle class and an ageing population have led to rising wages, eroding China’s competitive advantage. Partly because of the one-child policy, formally phased out in 2015, China’s working-age population is expected to fall from one billion people last year to 960 million in 2030, and 800 million by 2050.

In recent years, China’s central planners have been promoting automation as a way to fill the labour gap. They have promised generous subsidies — to be doled out by local governments — to smooth the way for Chinese companies both to use and build robots. In 2014, President Xi Jinping called for a “robot revolution” that would transform first China, and then the world. “Our country will be the biggest market for robots,” he said in a speech to the Chinese Academy of Sciences, “but can our technology and manufacturing capacity cope with the competition? Not only do we need to upgrade our robots, we also need to capture markets in many places.”

The march of the machines, not just in China but around the world, has been accelerated by sharp falls in the price of industrial robots and a steady increase in their capabilities. Boston Consulting Group, a management consultancy, predicts that the price of industrial robots and their enabling software will drop by 20 per cent over the next decade, while their performance will improve by 5 per cent each year.

《金融时报》报道视频

Liu Hui, an entrepreneur in his forties, is making the most of China’s robot boom. In 2001, when he opened his first factory in Foshan, an industrial city of seven million people in Guangdong, he started out making knock-offs of electric fans. As his business grew, he moved to bona fide manufacturing, producing components for Chinese home appliance brands. Then, in 2012, spotting an opportunity in a growing market, he jumped into the emerging world of robotics. Liu now imports robotic arms from suppliers such as the Swedish-Swiss conglomerate ABB, and sells them on to Chinese manufacturers, helping them integrate the machines into their production lines. It is a highly specialised business. Most of his customers are component-makers who supply motors and other parts to large Chinese home-appliance brands such as Midea and Galanz, which produce air conditioners, refrigerators and more.

E-Deodar: Robot arms are programmed to complete repetitive tasks

Business has expanded so quickly in the past year that Liu does not have enough space in his factory for all the machinery he is assembling. He has to store parts designed to support a $23,000 ABB robot under a makeshift lean-to outside. “Things are changing rapidly,” he says. “The cost of labour is rising every year, and young people don’t want to work on the production line like their parents did, so we need machines to replace them.”

The stereotypical image of China’s factories can still be found in many places: tens of thousands of people in long lines hunched over sewing machines or slotting components into a printed circuit board. But that mode of manufacturing is starting to be replaced by a more mixed picture: partially automated production lines, with human workers interspersed at a few key points.

 

Meanwhile, China is developing its own robot makers. In September last year, Ningbo Techmation, a Shanghai-listed producer of machinery for the plastics industry, launched a subsidiary, E-Deodar, making robots that are 20-30 per cent cheaper than those produced by international companies such as ABB, Germany’s Kuka or Japan’s Kawasaki. The E-Deodar factory in Foshan, with its café, chill-out zone and open-plan production line, looks more like the offices of a Silicon Valley tech start-up than a Chinese industrial workhorse. “Our global rivals are very good at making robots but their costs are higher and they are not so good at understanding the needs of local customers,” says Zhang Honglei, the company’s 35-year-old, spiky-haired technical director.

The cost of labour is rising and young people don’t want to work on the production line like their parents did
Liu Hui, Chinese entrepreneur

 

This year, Zhang plans to produce 350 distinctively green-coloured robots, which are designed for use in plastic factories and sell for between $14,000 and $18,000 each; in three years’ time he hopes to produce 3,000 a year. “We have to move fast because automation is a scale business,” he says. “The bigger the better.”
Chinese manufacturers, which bought 66,000 of the 240,000 industrial robots sold globally last year, still largely prefer to buy international brands, according to Litzenberger of the IFR. But she expects that to change, particularly in the wake of the Beijing government throwing its full support behind the domestic robot industry in recent years. “They are developing very fast,” she says.

Zhang Peng, vice-director of the economy and technology bureau, Shunde, Foshan

At an imposing, colonnade-fronted government building — known locally as the “White House” — in the Shunde district of Foshan, officials are trying to put President Xi’s call for a robot revolution into practice. The province of Guangdong has vowed to invest $8bn between 2015 and 2017 on automation. Zhang Peng, vice-director of Shunde’s economy and technology bureau, recently had his office in the building reduced in size, in line with the Communist party’s call for bureaucratic austerity. But the budget for industrial automation was unaffected. Zhang says robots are vital to overcome labour shortages and help Chinese companies make better quality, more competitive products. Unusually straight-talking for a Chinese official, he warns: “If manufacturing companies don’t improve, they won’t be able to survive.”

…
Government support for the integration of ever cheaper and more efficient industrial robots is good news for factory owners in China, who are facing a weak global economy and a slowdown in domestic demand. But the benefits of the robot revolution will not be shared equally across the world. Developing countries from India to Indonesia and Egypt to Ethiopia have long hoped to follow the example of China, as well as Japan, South Korea and Taiwan before them: stimulating job creation and economic growth by moving agricultural workers into low-cost factories to make goods for export. Yet the rise of automation means that industrialisation is likely to generate significantly fewer jobs for the next generation of emerging economies. “Today’s low-income countries will not have the same possibility of achieving rapid growth by shifting workers from farms to higher-paying factory jobs,” researchers from the US investment bank Citi and the University of Oxford concluded in a recent report, The Future Is Not What It Used to Be, on the impact of technological change.

They argue that China’s rising labour costs are a “silver lining” for the country because they are driving technological advancement, in much the same way that an increase in wages in 18th-century Britain provided impetus to the world’s first industrial revolution. At the same time, according to Johanna Chua, an economist at Citi in Hong Kong, industrial laggards in parts of Asia and Africa face a “race against the machines” as they struggle to create sufficient manufacturing jobs before they are wiped out by the gathering robot army in China and beyond.

 

Tom Lembong, Indonesia’s 45-year-old trade minister, and a leading voice for liberalisation and reform within the government of Southeast Asia’s biggest economy, is aware of the risks. “Many people don’t realise we’re seeing a quantum leap in robotics,” he says. “It’s a huge concern and we need to acknowledge the looming threat of this new industrial revolution. But as a political and business elite, we’re still stuck on debates about industrialisation that were settled in the 20th and even 19th centuries.”

Countries such as Indonesia are already suffering from something that the Harvard economist Dani Rodrik has dubbed “premature de-industrialisation”. This describes a trend where emerging economies see their manufacturing sector begin to shrink long before the countries have reached income levels comparable to the developed world. Despite rapid economic growth over the past 15 years, Indonesia saw its manufacturing industry’s share of the economy peak in 2002. Analysts believe this is partly because of a failure to invest in infrastructure, and the country’s uncompetitive trade and investment policy, and partly due to globalisation.

Rodrik believes the country will never be able to grow at the kind of rapid rate experienced by China or South Korea. “Traditionally, manufacturing required very few skills and employed a lot of people,” he says. “Because of automation, the skills required have increased significantly and many fewer people are employed to run factories. What do you do with these extra workers? They won’t turn into IT entrepreneurs or entertainers; and, if they become restaurant workers, they will be paid much less than in a factory.”

Five factories a year have left the industrial park on the Indonesian island of Batam © Muhammad Fadli

The spread of robots makes it much harder for developing countries to get on the “escalator” of economic growth, he argues. That is bad news for the estimated two million young people who enter the workforce every year in Indonesia, a nation of 255 million, where 40 per cent live on $3 a day or less. Mahami Jaya Lumbanraja, a 22-year-old job-seeker on the Indonesian industrial island of Batam, is feeling the effects of the premature de-industrialisation phenomenon. For seven months he has been looking for a factory job in Batam, which sits just 20 miles from prosperous Singapore, but he has had no luck. Wearing faded jeans, a grey hoodie and an endearing smile, Lumbanraja says that although he has one year of experience working for Shimano, the Japanese manufacturer of bicycle gears and fishing tackle, he is not experienced enough to secure anything more than an entry-level position, and that there are many more job hunters than openings. “I can survive on the little money I get from busking and helping friends with construction work but I must get a proper factory job to save enough money so I can set up my own small shop later,” he says. Wages in Batam — around $230 per month — are double what Lumbanraja could earn in his home city of Medan, on the island of Sumatra. So he feels he must stay until he finds work.

Lumbanraja is one of about 700 Indonesians in their late teens and early twenties who visit the community centre at the Batamindo industrial park every day looking for work. In February, 3,000 people applied in person for just 80 positions at a Japanese-owned wiring factory there, a gathering so large that executives initially feared it was a labour protest.

Mahami Jaya Lumbanraja is one of 700 Indonesians who visit Batamindo every day looking for work © Muhammad Fadli

Batamindo is a joint venture between Singaporean and Indonesian investors that was backed by Presidents Lee Kuan Yew and Suharto — the two nations’ respective rulers — when it opened in 1990. Intended as the showpiece of Indonesia’s industrialisation strategy, it has become a symbol of everything that is wrong with it. In recent times, an average of five factories a year have left the industrial park for other countries and the number of people employed there has dropped to just 46,000, from a peak of 80,000 in 2000. That is despite the fact that wages today are between one-third and one-half of the level paid in China’s Guangdong province.

Lembong, a Harvard graduate who ran his own Singapore-based private equity firm before he was appointed trade minister in August, says the government is determined to tackle the twin problems at the heart of Indonesia’s economic malaise: weak infrastructure and over-regulation.
But some argue that reform will come too late. During its period of rapid industrialisation, China invested in the modern highways, railways and ports necessary to support its manufacturing sector. In contrast, the physical infrastructure in Batam and much of Indonesia has “not changed much since the 1970s,” says Mook Sooi Wah, general manager of Batamindo.

Indonesia actually had a slightly higher “robot density” than China when the latest figures were collated by the International Federation of Robotics in 2014, although the situation is likely to have changed dramatically since then given the pace of Beijing’s automation push. This anomaly was largely the result of China’s manufacturing workforce being so much bigger than that of Indonesia, which still has no government plan or backing for industrial automation.

Many people don’t realise we’re seeing a quantum leap in robotics
Tom Lembong, Indonesia’s trade minister

 

Indonesia’s regulatory process is as fusty as its infrastructure. Recently, legitimate shipments from a paper factory were held by customs at the Batam port because of a rule meant to stop the export of illegally sourced wood. These problems leave even Batam’s boosters exasperated.

Stefan Roll, a German manufacturing veteran who worked in China during its industrial take-off in the 1990s, enjoys living and working in Indonesia. But he worries that the country is missing its “golden opportunity” to become efficient enough to compete on a global scale. “When you are dealing with multinationals, time is money,” Roll says as he shows off his new factory in Batam, which assembles coffee machines for Nestlé. “But you can only do just-in-time manufacturing if you have good roads and infrastructure.”

While few doubt the depth of the challenges facing developing countries, not everyone sees the dilemma in such bleak terms. With wages in countries such as Indonesia and India much lower than in China and their populations still relatively young, some analysts believe they can attract more labour-intensive industries, such as garment-making, where widespread automation is not yet suitable.

 

“As China moves up the industrial chain, it’s actually freeing up a lot of opportunities for Southeast Asia and India,” says Anderson Chow, a robotics industry analyst at the investment bank HSBC, in Hong Kong.

Hal Sirkin, an expert on manufacturing at Boston Consulting Group, says that from the perspective of an economy such as India’s, it does not make sense to automate now because it would drive up the price of goods — “when they have one billion people who can make things cheaply”. He is among the tech optimists who believe that, in the medium term, automation will also create new business niches for emerging economies, mitigating the damage from the jobs that will be eradicated.

 

“We think you’re going to see more localisation rather than more scale,” says Sirkin. “I can put up a plant, change the software and manufacture all sorts of things, not in the hundreds of millions but runs of five million or 10 million units.”

But Carl Frey, an expert on employment and technology at the University of Oxford, warns that without better education and more skills, developing countries will struggle to take advantage of advancements in manufacturing.
“Technology is becoming increasingly skill-based,” he says. “Many of these countries don’t have a skilled workforce so they’re not very good at adopting these technologies.”

…

The Shangpin Home Collection factory where the use of robots to cut and drill wooden planks improved productivity by 40 per cent

China itself is not immune from the negative consequences of automation. More than 40 per cent of its 1.4 billion population still live in the countryside, many in poverty, having benefited only marginally from the urban economic miracle.
But the government is betting that the benefits of promoting cutting-edge manufacturing will outweigh the damage from the potential jobs lost. The industrial strategy announced by Beijing last year — known as Made in China 2025 — is designed not only to improve the technological capability of its factories but also to support the development of Chinese brands internationally.
Chow, the HSBC analyst, says that as Chinese companies try to increase their exports to alleviate the impact of the domestic slowdown, they are likely to focus more on the quality of their products: “Quite often part of that development is a better production process, involving robotics.”

Every year, the amount of time it takes for a company’s investment in a robot to pay off — known as the “payback period” — is narrowing sharply, making it more attractive for small Chinese companies and workshops to invest in automation. The payback period for a welding robot in the Chinese automotive industry, for instance, dropped from 5.3 years to 1.7 years between 2010 and 2015, according to calculations by analysts at Citi. By 2017, the payback period is forecast to shrink to just 1.3 years.

Li Gan, general manager of Shangpin Home Collection, Foshan

Automation is not just about putting cheaper and more efficient robot arms on the production line. Li Gan, the general manager of Shangpin Home Collection, which makes and sells customised home furniture, says the greater opportunity is to integrate robots on the factory floor with real-time data from customers and automated logistics systems.

Thanks to the use of robots, the factory that Shangpin opened in Foshan in 2014 was 40 per cent more productive than its previous plant, even though it employs 20 per cent fewer people. Later this year, it will start up the machines at its newest and biggest production base, where it hopes to improve productivity fourfold with just double the number of staff, by using more robots to move supplies around the factory floor and help pack outbound shipping containers.

Drilling wooden slats for the company’s diverse range of beds, wardrobes and other bespoke furniture used to be a painstaking and sometimes dangerous process. Now a worker simply picks up each piece of wood, scans a barcode and puts the wood on a conveyor belt that takes it to the robot arm. The finished product returns on another belt. The process in between is strikingly complicated: Shangpin had to design a device to make sure that each slat would be aligned in the right way for it to be grasped by the robot arm, and the drilling specifications for the slats have to be pre-programmed and recorded in a barcode, because the robots do not yet have any artificial intelligence capability. Li Gan points out that human oversight and decision-making is still crucial. “Automation is just a technical process but what is more important is our thinking about how best to do this,” he says. “Every time we change something, we ask: is it more effective to do this using humans or robots?”

Every time we change something, we ask: is it more effective to do this using humans or robots?
Li Gan, general manager of Shangpin Home Collection, Foshan

Boston Consulting Group forecasts that the percentage of tasks handled by advanced robots will rise from 8 per cent today to 26 per cent by the end of the decade, driven by China, Germany, Japan, South Korea and the US, which together will account for 80 per cent of robot purchases. Sirkin at BCG says that the rapid expansion of automation could be compared to the difference between the “human learning curve” and Moore’s Law, which posited that computing power could double every 18 months to two years. “Even if you’re very good, humans can only double their productivity at best every 10 years,” he says. In contrast, researchers can push robots to double their productivity every four years, he estimates. “Compounded over time, that makes a big difference.”

As China and other industrial leaders build more and better robots, the tasks they can take on will expand. Butchery, for example, was long considered the sort of skill that machines would struggle to develop, because of the need for careful hand-eye co-ordination and the manipulation of non-uniform slabs of meat. But Sirkin has watched robots cut the fat off meat much more efficiently than humans, thanks to the use of cheaper and more responsive sensors. “It’s becoming economically feasible to use machines to do this because you save another 3 or 4 per cent of the meat — and that’s worth a lot on a production line, where you can move quickly.

“There are things that humans can do better than robots,” he adds. “But they are getting less and less.”