Recent reads (Mar-Apr 2024)

I am late for my book reporting. These are the books I read mostly in March. 

Going infinite (Michael Lewis, 2024)

I had the highest of praise for Michael Lewis several times, and called him one of my favorite authors. I feel sorry to write a bad review for one of my favorite authors.

Lewis faced significant backlash from people who questioned his impartiality in portraying Sam Bankman-Fried (SBF), raising concerns about the fairness of his depiction. My complaint is not that. I came to this book for captivating story-telling, which, unfortunately, falls short.

This book pales in comparison to Lewis's past works. We are talking about the guy who made me read thick books about finance and stock trading worlds which I had zero interest in as if they were fast-paced thriller novels. So I was really looking forward to reading this book to get that reader's high, but I was disappointed.

I am familiar with the blockchain and cryptocurrency world through my research and my internal debates about its value. I had been following the debacle from outside. SBF is a fascinating character and there is a great story here. But the book fails to deliver that gripping narrative. I felt bored. The storytelling in the book failed to connect. While I've lauded Lewis for his masterful storytelling, his usual magic was absent here.

It is hard to pinpoint the problem. Maybe it's the lack of Lewis's trademark elements: the clever analogies and sharp one-liners that elevate his work. Their absence is undeniable. Or perhaps the book was rushed, published before it was fully developed, and we got to see draft 2 instead of draft 4.  

This is still a good book. If you had not been following all this hoopla, this is a good book to catch you up completely. There were also some nice parts of the book e.g., about SBF interview and time in Jane Street. But even those parts could be better. On a side note, Lewis's podcast on the SBF affair was fascinating, and I enjoyed listening to that. 

It is probably unwise for me to comment on what went on and how much malice was there and what was the root cause. But here is my take. 

I don't disregard explanations about this being due to mismanagement, rather than intentional fraud. These people had been too reckless with the money. The cryptocurrencies scene had been mostly a field of speculation and scam. They got too acclimated to that. And there is also the power poisoning and lack of commonsense at high levels.  "As you go above, commonsense gets sparse like oxygen." On grand strategy (John Lewis Gaddis, 2018)

Secondly, it is clear SBF is on the autism spectrum. I have seen some people who are brilliant in a theoretical field, but unable to function normally and pay bills. Smart people in one area can be very incompetent in many others. Regardless of reasons, the trial is over and SBF will be saving 25 years time for violating the law (and getting caught for it).

Talent code (Daniel Coyle, 2009)

While I wasn't a fan of the author's "The culture code" book, I found this book both informative and engaging. This book exlores many examples of "talent hotbeds" – environments like tennis academies, music schools, and even Michelangelo's workshop – that seem to foster exceptional talent. Have these places figured out the secret of accelerated learning? Is the secret competent/master teachers?

Or maybe their success is due to mostly the scenius effect. But that is not too far from the book's thesis anyways. The book says that: Analyzing talent hotbeds worldwide reveals talent to be a mixture of deep practice, ignition/motivation and master coaching.

In particular, the book zooms into the deliberate practice aspect, highlighting it as the cornerstone of talent development. The core message of the book is that: Deliberate practice (of making, catching, and correcting mistakes) fosters talent because it stimulates Myelin growth!  

Myelin is the substance that insulates neurons and accelerates the transmission of signals. The book argues that according to euroscience research the difference between talented and mediocrity is the layers of Myelin wrapped around the neurons in the brain. This increased Myelin allows for increased "bandwidth" or speed of firing neurons. The more a neuron is fired the more insulation it demands hence the increased wrapping of myelin. So the secret to talent is deliberate practice with the aim of correcting mistakes one at a time. Breaking down a difficult task into its component parts and slowly learning each one until it all comes together. I had written related things in my "Master your tools" and "Learning a technical subject" posts.

I really like the rapid feedback idea in the book, especially as covered by the futsal arena examples. If you are interested in the design of distributed systems, TLA+ gives you rapid feedback!

This was a well-written book exposing some resarch on the topic as well. But I can also make a cynic case for the book, by flipping this, and saying that this was catchy commercialized writing, and there were more fluff than substantive research. It would be fascinating to see more research on the topic, and exposition of this research. Imagine the breakthroughs we could achieve, if only we could channel 1% of the resources/effort/interest being directed to researching/developing machine learning to researching/developing human learning.

PS: The author has a video presenting his book here.

The soul of a new machine (Tracy Kidder, 1981)

This book talks about mini-computers. This is just before micro-computers, aka, personal computers era. I immensely enjoyed this book. 

For this book, Tracy Kidder, a Pullitzer winning author, embedded within a skunkwork mini-computer project team at DataGeneral, and wrote their story as they launched Eclipse MV/8000. This is both a burnout and victory/accomplishment story. 

The writing is top notch. The book engages and transports the reader to 1980 Data General. It felt like Kidder is a McPhee style writer. Kidder wrote in a 1994 essay, "In fiction, believability may have nothing to do with reality or even plausibility. It has everything to do with those things in nonfiction. I think that the nonfiction writer's fundamental job is to make what is true believable."

Here is the brief synopsis of the book from Wikipedia. 

Tom West, the leader of the Westborough designers, starts a skunkworks project. Code-named "Eagle", it becomes a backup plan in case Fountainhead fails, and then the company's only hope in catching up with DEC. In order to complete the project on time, West takes risks: he elects to use new technology, and he relies on new college graduates (who have never designed anything so complex) as the bulk of his design team. The book follows many of the designers as they give almost every waking moment of their lives to design and debug the new machine.

Tom West practices the "Mushroom Theory of Management" – "keeping them in the dark, feeding them shit, and watch them grow." That is, isolating the design team from outside influences and, instead, using the fear of the unknown to motivate the team.

I think this would be a good supplemental reading for a computer architecture course to provide that motivation/ignition component. This link says that it was used like that for sometime. The page has great photos and additional material that was not in the book. Through reading the book, I also learned the meaning of microcoders. Pat Helland had been referring to kids as microcoders, and I thought it was only because the kids are physically small. I learned the double pun (of course a single pun would not be worthy of Pat) in that this job title actually existed. Microcoder is people who wrote the microcode for new computer architectures.

Reading the book I realized that engineering and engineers did not change much over the next 5 decades. You can recognize the engineer characters in this book in your teams even today. You can appreciate the debugging mysteries they tackle, the way they cooperate and communicate with each other. I was also amused to see how a young engineer wrote a very nice (and deterministic) simulator to help microcoders start working before the hardware is ready, and that work basically saved the project and help it ship in time.