As I concluded the final pages of Kai-Fu Lee’s book, somewhat annoyed by the weakness of the ending, I wondered what this world is going to look like when AI replaces many of our routine tasks and work lives.

I can tell you what our world looks like today: interminable gridlock on Toronto streets. Thousands of cars making trips that needn’t be taken at all. Polluting the air. Throwing up pointless tons of CO-2 emissions.

Why couldn’t people make the same trips with VR glasses to virtual workplaces and eliminate the physical consumption of resources? And schools. And malls. And govt facilities.

AI could help build these virtual places. ( )

generally interesting but also somewhat repetitive. Also it felt really strange to read an AI book that didn't include the word "ethics" even once - maybe I missed that part ( )

Wider in scope than the title suggests

I enjoyed reading this book because of the high level view it provides of several important tends in tech in the last decade and likely the next one: the growing significance of the Chinese tech ecosystem and of machine learning based applications.

The background section explaining the rise of and competition between Chinese tech companies for the Chinese market is probably good enough reason to get the book, because the stories in it challenge many unstated assumptions in the West about the nature of Chinese tech companies today. ( Eg. That they are copycats which map neatly to an American prototype, or that they are almost like government sponsored monopolies, or even that American companies' failure to dominate that market are mostly due to government action ).

I found his take on which industries are more susceptible to automation in the coming decades based on data categories, and his appraisal of the strengths and weaknesses of the American and Chinese tech ecosystems thought provoking. For example, that Chinese companies will have or already have the data edge in industries that mix digital and physical world operations, due to vertical integration and amount of users,
And that America has the edge in top talent. This later factor would only come into play in the case of a true breakthrough (eg in the amount of labelled data needed), so it is unpredictable. In other industries he sees them as more evenly matched.

There's more in it than those two topics. Later sections make the case that the effects on employment may be more relevant to the world at large than the rivalry angle in the title.

I liked the organization of the book and the clarity of the writing, though it comes at the cost of some repetition. In terms of things that seem missing: more digging into the constraints different ecosystems operate under, eg privacy and labor laws and their enforcement, surveillance, as well as trade secrets. He mentions there is a privacy law in China as well, but there's no digging into details comparing them or what it means in practice. ( )

Classical case of bad China business writing:
* Reduce world to a binary contrast between ‘Caricature USA’ vs ‘Caricature China’;
* Ignore the political context and the CCP’s behaviour and aims entirely;
* Fall for the ‘Amazing China’ propaganda trap, engage in baseless speculation based on anecdotes and pilots and PR stunts.
Questionable economics, non-existent labour theory.
Alright introduction to debate on impact of AI. ( )

I fail to see why this book gets such high ratings. Yes, it's written in a very readable way. Despite the author's experience in A.I., he never makes it too technical. There are not enough graphic charts, but the ones that are there are very useful.

Unfortunately, all of this is of course not enough to write a positive review. Why not? Well, for starters, there's almost no mention of India. I'm a aware that this book is meant to be a comparison between the two tech giants, but India is up & coming, and I believe that it will surpass China very soon. And now, the worst part: this has got to be the most repetitive book I've ever read. I'll give you an example: In chapter 5, The Four Waves of AI, Lee mentions the Chinese company Iflytek a whopping eight times, and this is only on the first two pages!

I still don't understand how I made it through... ( )