Summer 2025 Book Recommendations
Nuclear War by Annie Jacobsen
Can you think a book is fantastic and simultaneously also hope that you wished you never read it? Well, this book did that for me. Jacobsen has done copious research and interviewed people who would talk to her (many will not/cannot due to security reasons). In this book she explains how a nuclear war would unfold. While the story itself is (thankfully!) fictional, her take on how it would play out is certainly not her imagination. Most nuclear powers - especially the US and the Soviet Union (now of course Russia) - have extremely detailed plans on how to react to a nuclear attack. With either remarkable sang-froid or a complete lack of self-awareness both countries have planned down to the minute and second, how to do the counter-strike. And add China to the mix now. The timeline of making decisions in this reaction is not weeks or days or even hours but mere minutes. A nuclear launch in North Korea headed towards a California nuclear plant and a second one towards Washington D.C. kicks off the book and the US president has mere minutes to decide what to do while also trying to get the heck out of Washington D.C. You should (maybe?) read the book but suffice to say that in about 3.5 hours after Kim Jong Un launches a nuclear strike against the US it is basically game over for ALL of us. Billions (not millions) will die in hours and the rest shortly thereafter. 300,000 years of human history will come to an end by the 6pm news. There is no real way to avoid this chain of events once that first strike is launched. I wish I didn't know this.
Why Machines Learn by Anil Ananthaswamy
A absolutely wonderful and timely book. Why Machines Learn is not a casual read - and that is a great thing! Ananthaswamy's goal is to try and peer behind the sometimes-indistinguishable-from-magic curtain of ML/AI to get a better handle on what it is about the architectures and algorithmic design that enables machines to learn. He does not shy away from diving into the math to explain the algorithms that power AI systems today. For this book, the author interviewed some of the biggest names in AI and used their perspectives to come up with a cogent and engaging explanation of one of the hottest topics in the world today. To be sure there are other books like my recommendation from last year of Cade Metz's Genius Makers. If that book was a gentle stroll through the neighborhood park this is a half-marathon run through the hills. Not impossible but certainly more demanding - and well worth it. If you stick with it and get through the math, then you will be rewarded with a deeper understanding of ML/AI than the vast majority of people out there. Ananthaswamy has a gift of making the math accessible and even fun to dive into.
The Thinking Machine by Stephen Witt
Another timely biography, this time of the person who is perhaps more responsible for the current AI boom than anyone else. Jensen Huang - he of the black leather jacket - gave Witt access to Nvidia's company leaders and his friends and this book is a well-rounded 360-degree view of the man and the company. Nvidia is not a hot new AI company like OpenAI or Anthropic. It was founded in 1993 and has been led by Huang since Day 1 which is a record for long lasting CEOs. But OpenAI and Anthropic - and Meta and xAI and Google - and every other serious frontier model company out there could not exist without Nvidia. Nvidia started off very much as a graphics company. I remember my first Dell computer from 25 years ago having an Nvidia graphics card. What Nvidia had was an architecture that helped them do math really really fast with parallelism. To use an analogy, Nvidia's GPUs were a super-specialized type of knife while CPUs are like a Swiss Army Knife. Most of the time, what you need to pack is the swiss army knife and so it was with CPUs becoming huge - witness the rise of Intel for example. GPUs did graphics and math way better and faster but they were still stuck in niche fields - computer gaming, hollywood animation studios, scientific modeling etc. Along came the Bitcoin and crypto currency revolution which needed GPUs to do math super fast. Nvidia graphics cards sudddenly became hot property. But it was when the AI revolution began and the algorithms and math behind it needed gigantic matrix multiplication operations, that Nvidia's time had truly come. It's not just Nvidia's chips and hardware, it is also the software ecosystem, most notably CUDA, that is the hungry alligator-filled moat that Nvidia has built. Seeing many years ahead and patiently navigating to this moment in the sun is what Jensen's vision is about. At the same time Witt is quite blunt about Huang's overdrive, his temper and his competitiveness. Read this book for the detailed story.
House of Huawei by Eva Dou
When I think of iconic tech companies, they are almost always American - IBM, Microsoft, Apple, Intel etc. My awareness of Huawei came more from all the news of it being banned - in the UK, in India, and in the US. I was vaguely aware of their smartphones but began to pay more attention when the US DoJ arrested Meng Wanzhou, Huawei's CFO (and the daughter of its founder, Ren Zhengfei) - in Vancouver airport and sought her extradition to the US. Not something that one sees everyday. This book begins with this event - the arrest in Dec 2018 to tell the story of Huawei. From its founding in 1987 through the Chinese economic reforms, the admission of China into the WTO in 2001 and eventually its domination of the global telecom industry - the very domination that has led to fears of backdoors in their software and hardware and the bans in the US, UK, and India. The real story though is about its founder, Ren Zhengfei, and the company's journey from a small company in Shenzhen to a global behemoth. In many ways Huawei is an embodiment of the rise of China itself. One wonders with the recent bans of the exports of advanced chips like Nvidia's GPUs to China, if in the next 5-10 years, Huawei might not become the next Nvidia. After reading this book, I believe that is a distinct possibility.
The Boundless Sea by David Abulafia
For a change of pace from AI and GPUs and Huawei, I decided to read this book. I should start by saying this was about 1100 pages long - over 40 hours as an audiobook. It took me the good part of a month to get through it. And still the author David Abulafia, a historian and professor at Cambridge University felt compelled to say that he didn't feel that this book was an "attempt to be complete or comprehensive"! For a book this long and detailed it always held my attention. It is extremely well-researched and engaging and tells the story of the world from the vantage point of seas and voyages. And what a story it is! I won't attempt to summarize 1100 pages in a paragraph but I will say that as a meta-observation: the countries that get the most mentions and attention are (in rough order): India/China, Ancient Rome, Malaya, Spain, Portugal, and Britain. And then other European powers such as the Dutch, the Russians and so on. The United States merits its first mention only after about 90%-95% of the book is done. It was a whole new perspective on how young the US really is in this broader historical context. Trade and commerce is the thread that weaves the entire story together. A lovely book and it has now tempted me to try and start on more of these 1000+ page books once again!
About |
Projects |
Reading |
Writing