Kai-Fu Lee is the Chairman and CEO of Sinovation Ventures that manages a 2 billion dollar investment fund and develops the next generation of Chinese high-tech companies. He is the former President of Google China and founder of Microsoft Research Asia that trained many of the AI leaders in China, including those at Baidu, Tencent, Alibaba, Lenovo, and Huawei. He was also named one of the 100 most influential people in the world by TIME magazine. I got my hands on his New York Times bestseller AI Superpowers: China, Silicon Valley, and the New World Order last year but only got a chance recently to dig into it.
In AI superpowers, Kai-Fu Lee reveals and discusses how China has suddenly caught up to the US at an astonishingly rapid and unexpected pace.
Decades ago AI was a field limited to academia, research labs and science fiction books. Today it has leapt to the forefront of human minds and conversation. People had an idea that AI was about making robots and machines think, but there was no connection between it and everyday life. That is no longer the case today.
From our internet browsing to shopping - AI is now a part of our everyday lives, whether we know and like it or not. It powers much of business and personal applications that we rely on.
In the 1950s, pioneers of AI set a lofty but well-defined mission of recreating human intelligence in a machine. This drew in some of the greatest minds to pour their expertise. In a few decades AI had divided itself into two camps (1) The rule-based camp that encoded a series of logical rules - if X then Y- and it worked well enough for simple and well-defined games but fell apart when the universe of choices became more complex and expanded. (2) The Neural Networks camp that tried to reconstruct the human brain itself. This approach mimicked the brain’s underlying architecture, constructing layers of artificial neurons that could transmit and receive data. You fed a lot of data of a certain phenomenon (such as cats) into the neural network and let the network figure out the patterns.
Neural networks did not gain much popularity in the past century but what resuscitated it in recent years were the two ingredients that were scarce in the previous century 1) Computing power and 2) Data
Deep learning is exciting because of its core power- the ability to recognize a pattern, optimize for a specific outcome and make a decision. Its application and success has left companies like Google and Facebook scrambling to hire more deep learning experts and left the rest of us with the sense that we are on the precipice of a new era. One in which machines will radically empower or/and violently displace humans.
Kai Fu Lee argues that many people think we are living in an age of discovery given the medias breathless reporting of breakthroughs every day but that is misleading. The reality is these are milestones that are the applications of the past eras breakthrough. We are now in an age of implementation. Much of the difficult abstract work of AI research has been done, and now entrepreneurs and businessmen are rolling their sleeves to get down to the dirty business of turning algorithms into business.
Turning Algorithms into business is where Kai Fu Lee goes into great detail about how the US and China have harnessed this new technological wave. Most of us look towards silicon valley and think that is the only ideal booming tech entrepreneurial ecosystem but Kai Fu redirects you towards china where some leading giants today such as Ali Baba, Tencent, Baidu had their origins. He writes in detail about how this culture developed in China. From the era of copy cats where Chinese would copy US apps and sites pixel to pixel to the era of modifying and pivoting products tailored to Chinese audience.
Below Screenshot of Kai-Fu Lee’s presentation at O’Reilly AI Conference 2018 shows how Chinese used Western design to leapfrog unique products and then innovated. Now we are beginning to look at an era where Western companies are creating local replicas of Chinese origin!
“In stark contrast, China’s startup culture is the yin to Silicon Valley’s yang: instead of being mission-driven, Chinese companies are first and foremost market-driven. Their ultimate goal is to make money, and they’re willing to create any product, adopt any model, or go into any business that will accomplish that objective. That mentality leads to incredible flexibility in business models and execution, a perfect distillation of the “lean startup” model often praised in Silicon Valley. It doesn’t matter where an idea came from or who came up with it. All that matters is whether you can execute it to make a financial profit. The core motivation for China’s market-driven entrepreneurs is not fame, glory, or changing the world. Those things are all nice side benefits, but the grand prize is getting rich, and it doesn’t matter how you get there.”
In Lee’s opinion, to successfully deploy AI, you need to align four assets or forces: Computing power, Data, Data science expertise and Policy - all 4 of which China has harnessed making it so well-positioned at the precipice of a new era.
According to Lee, there are 4 waves of AI development: Internet AI, Business AI, Perception AI and Autonomous AI. The first two waves are already with us. Internet AI has a powerful grip on our eyeballs and wallets. It is largely about using AI algorithms as recommendation engines. Ever noticed how amazon knows exactly what purchase to recommend, how you seem to go down into a rabbit hole of YouTube videos and how Netflix knows exactly the right thing to recommend for bingeing. Algorithms are deciding the ads we see, the videos we are offered and the news we read. The key is data. The more we interact with the AI through clicks, views and engagement, the more it understands and the better it gets at giving you what you want.
Business AI takes advantage of the fact that enormous companies have been labelling and storing data for decades, and it helps them exploit that data for better results. Issuing loans, catching frauds, insurance risks, diagnosing diseases, finding a correlation between symptoms and diagnosis. In some countries Business AI is now helping judges give out fair sentences based on evidence and studying all the past cases with similar evidence.
Perception AI is digitizing our world through sensors and smart devices. It is what blurs the lines between the offline and online worlds. Now you can order your takeout through amazon Alexa instead of punching the order on the phone. Ali Pay has pioneered pay-with-your-face services in some stores where your face is linked to your online wallet. Want to pay for that big bag of cheetos and forgot your wallet? Scan your face and before stuff it with cheesy cheeto dust goodness.
Almost all smart home applications fall under perception AI and their future potential is dizzying. Frankly I can’t wait the day my refrigerator will be able to scan all the groceries I am short on communicate it to my supermarket shopping cart. I just hope it can get Dhaniya and Podina right.
Autonomous AI is the culmination of all three previous waves. It will unite the superior thinking of machines with physical extensions. But lets not narrow our vison with just autonomous cars as an example. Robots carrying out harvests on farms, robots in Amazon warehouse picking and packing orders, and drones carrying out search and rescue missions are all powerful examples of autonomous AI.
The difference between autonomous and automated comes down to the ability to make decision and improvise based on changing conditions. By giving machines the power of sight, sense of touch and the ability of optimize form data we have dramatically expanded the number of tasks they can handle.
When we think of the consequences of AI - we picture robots taking over the world with humankind at risk. Much thanks to hollywood for that. Lee argues the actual imminent and long-term impact of AI will be massive unemployment. PricewaterhouseCoopers estimates AI will increase worldwide GDP by $15.7 trillion dollars by 2030 with 70% of that going to the US and China. Since AI comprises algorithms that are easily installed where needed and maintained remotely, this revolution can take place at unprecedented speed. Jobs that don’t require social interaction or advanced robotics are most at risk: both White collar and blue-collar jobs. For example, Radiologists are at risk, Psychiatrists are not. Insurance executives are at risk, Public Relations directors are not. Dishwashers are at risk, dog trainers are not. Companies are constantly investing in technologies that will save time and optimize operations during downturns, so a recession would speed that up.
Lee evaluates prescriptions for dealing with an AI world with too many people for too few traditional jobs. There are suggestions such as retraining, reducing work hours and guaranteed income. All three are sound suggestions but will eventfully reach their limits. There is a need for policies with which governments and industries can invest in human-oriented service jobs. AI will do the thinking, but it has no emotion.
My postgraduate was in Robotics and AI following which I pivoted into non-tech/blended roles. One of the questions I get asked most frequently is why Robotics or AI and why in Pakistan of all places. I tend to throw this question back, “Why not?”. It still surprises me how under prepared we are for the AI revolution. You’d think the private sector would be making some investments in the field, but most companies, even multinationals, are skeptical about its implementation in a traditional market. It becomes a vicious cycle where skepticism engenders reluctance to implement which continues to keep data scarce and that scarcity can’t improve systems or learn the features of a traditional market to be effective which fuels skepticism and so it goes all over again.
China was much like Pakistan in that regard, but one major factor they had going differently for them was its government support in the area. And that triggered something akin to a domino effect which kickstarted action all the way to private, entrepreneurial and investment arenas.
Why read?
Kai Fu Lee is a technical genius who has led AI research/adoption and efforts in both west and east. He understands the complexity of issues that will stem from the explosive development of AI. His insights into the diverse cultural, political and technical factors that will frame the future of a new world order makes this book a must read for anyone interested in AI.
Kai-Fu Lee’s presentation at O’Reilly AI Conference 2018
Kai Fu Lee on the Lex Freidman Artificial Intelligence Podcast.
The other two books I read this week were:
The Silent Patient which I found disappointingly overrated. Avid readers of thrillers and mystery genres may find the writing and plot cliched
Folding Beijing (a dystopian novella) which I found was a brilliant quick read
This theatrical audiobook of the hobbit is so much more than a person saying words. It is incredibly immersive and a joy to listen. Enjoy!
-Saima
Thank you saima. It was a good read