Book review: “AI Superpowers”

Dr. Kai-Fu Lee is the chairman and CEO of Sinovation Ventures, a China based tech focused investment firm. Previous to becoming a full-time investor, Lee held positions at Google, Microsoft and Apple. A large part of that career, Lee spent working on data and Artificial Intelligence (‘AI’), both in the US and in China. In “AI Superpowers — China, Silicon Valley and the New World Order” Lee bundles his experiences and insights to describe the progress that China and the US have made and are making in the field of AI.

AI Superpowers contains a heap of valuable insights as well as predictions about the impact of technology power that both the US and China have been racking up. These are the main things that I took away from reading AI Superpowers:

• US and China, contrasting cultures — Lee starts the book by writing about the contrasts in business culture between the US and China: “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.” Lee goes on to explain that the ultimate goal of Chinese companies 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.” This mentality help to explain the ‘copycat’ attitude that Chinese companies have had historically. Meituan, for example, is a group-discount website which sells vouchers from merchants for deals which started as the perfect counterpart of US-based Groupon.
• “Online-to-Offline” (‘O2O”) — O2O describes the conversion of online actions into offline services. Ride-sharing services like Uber and Lyft are great examples of the new O2O model. In China, Didi copied this model and tailored it to local conditions. Didi was followed by other O2O plays such as Dianping, a food delivery service which subsequently merged with the aforementioned Meituan company, and Tujia, a Chinese version of Airbnb. Lee also mentions WeChat and Alipay, describing how both companies completely overturned China’s all-cash economy. More recently, bike-sharing startups Mobike (see Fig. 1 below) and ofo which supplied tens of millions of internet-connected bicycles, distributing them across them about major Chinese cities and now across the globe.
• China catching up quickly in the AI department — Having read the story of image recognition algorithm ResNet, and how its inventors moved from Microsoft to join AI startups in China, I can see how China as a country is quickly catching up with the technology stalwart that is Silicon Valley. One of these image recognition startups, Face +++, has quickly become a market leader in face / image recognition technology, leapfrogging the likes of Google, Microsoft and Facebook along the way.
• The four waves of AI — In AI Superpowers, Lee argues that what he calls the “AI revolution” will not happen overnight. Instead, AI will wash over us in four waves: internet AI, business AI, perception AI, and autonomous AI (see Fig. 2 below). This part of the book really struck a chord with me, as it brings to life how AI is likely to evolve over the coming years, both in terms of practical applications and use cases.

Main learning point: I’d highly recommend “AI Superpowers” to anyone interested in learning more about how China and the US are furthering the development of AI and the impact of this development on our daily lives.

Fig. 1 — Screenshot of the Mobike bike-sharing app — Taken from: https://technode.com/2016/07/07/mobike-uber/

Fig. 2 — The four waves of AI — Taken from: Kai-Fu Lee, AI Superpowers, pp. 104–139:

• First wave: Internet AI — Internet AI is largely about using AI algorithms as recommendation engines: systems that learn our personal preferences and then serve up content hand-picked for us. Toutiao, sometimes called “the Buzzfeed of China”, is a great example of this first wave of AI; its “editors” are algorithms.
• Second wave: Business AI — First wave AI leverages the fact that internet users are automatically labelling data as they browse. Business AI, the second wave of AI, takes advantage of the fact that traditional companies have also been automatically labelling huge quantities of data for decades. For instance, insurance companies have been covering accidents and catching fraud, banks have been issuing loans and documenting repayment rates, and hospitals have been keeping records of diagnoses and survival rates. Business AI mines these data points and databases for hidden correlations that often escape the naked eye and the human brain. RXThinking, an AI based diagnosis app, is a good example in this respect.
• Third wave: Perception AI — Third wave AI is all about extending and expanding this power throughout our lived environment, digitising the world around us through the proliferation of sensors and smart devices. These devices are turning our physical world into digital data that can then be analysed and optimised by deep-learning algorithms. For example, Alibaba’s City Brain is digitising urban traffic flows through cameras and object-recognition.
• Fourth wave: Autonomous AI — Autonomous AI represents the integration and culmination of the three preceding waves, fusing machines’ ability to optimise from extremely complex datasets with their newfound sensory powers.

Related links for further learning:

1. https://www.cnbc.com/2018/09/07/chinas-meituan-dianping-confirms-4point4-billion-hong-kong-ipo.html
2. https://techcrunch.com/2017/10/10/tujia-raises-300-million/
3. http://www.forbesindia.com/article/ckgsb/how-tujia-chinas-airbnb-is-different-from-airbnb/48853/1
4. https://en.wikipedia.org/wiki/Mobike
5. https://towardsdatascience.com/an-overview-of-resnet-and-its-variants-5281e2f56035
6. https://www.faceplusplus.com/
7. http://www.iflytek.com/en/
8. https://www.mi.com/global/
9. https://www.happyfresh.com/
10. https://www.grab.com/sg/