Ai Superpowers

3.5
This one started off strong explaining cut throat business wars in China but it eventually fizzled into a memoir (albiet moving) with lots of tips on technology design based off his life experience. An interesting read in retrospect with AI being a commercial product today, one wonders how this book wasn't at the fore when it was released.

A bit repetitive in some places, but as someone who works with AI and machine learning, I think the values and framework he talks about are absolutely essential.

The future of A.I. through the perspectives of its two biggest players: the US and China. Super interesting look at how their different cultural and political norms influence how they collect data, what they fund, and where their research is focused.

China surged onto the A.I. scene after DeepMind's AlphaGo defeated its Go champions. Encouraged and supported by the government it spent years learning and mimicking - the “copycat years” - but has now reached A.I. independence. No other nation has as many smartphone and app users generating data. No other nation is as lenient and nonrestrictive about its data laws. The combination of these two factors makes China the most furtile ground for future groundbreaking A.I. applications and research.

Kai-Fu Lee also talks about how A.I. industries will disrupt the current job market, and he ends on his hope that we'll figure out a way to use artificial intelligence as a tool and possibly the reason to become more focused on human and social connections. This aspirational section of the book feels almost naive, the same way the first section of the book feels very devoid of any critical look at China's politics. But nonetheless, this book gives a great glimpse at a field (and country) that will probably domineer so much of our future.

I guess the book is meant for a lay audience so I didn't learn all that much. There was some cultural insight into China (helping me lose a few preconceptions), but not much new about AI. There's a lot of starry-eyed predictions of how AI and data will improve everything, but no details or original examples, just repeating the well-known ones. (Also, he's name-dropping companies a lot, which I can only assume are his investments).

Thankfully, the author is aware of the real issues (societal problems due to automation, not AGI and robot overlords). I kinda liked the summary of various predictions of jobs to be automated soon. His critique of universal basic income as just a patch that the Silicon Valley elite wants to use to keep the masses docile was the most original idea in the book for me. But again, when the author tries to give some ideas on solutions, they're vague (and naive?) ideas: a class of ethical investors should develop who are content with linear returns and government should somehow reward socially-beneficial activities.

A surprising ending reminding us about what is really important in life, and how AI might actually help us to leave our economically-centred lives behind to chase it. Connections, community and culture.

One of the most well-written and interesting books I have read where I disagree with almost everything.