Some books were read: 2020
Here are some notes on a few books I enjoyed and learned from in 2020, not all of which are books published that year.
Deep Learning for Coders with fastai and PyTorch, Jeremy Howard, Sylvain Gugger. The book’s example are written in Jupyter notebooks all available online, and on a computer at least, the Kindle edition is well rendered. Apart from being a useful overview of both PyTorch and Jupyter, this is a good book for non-specialist software engineers to get started with deep learning. The data ethics chapter is usefully grounded. It’s clear the book is written by people who teach, per their fast.ai course: the book as a result is well structured, introducing concepts in sound order with each chapter ending with a good set of questions for revision/retention. For example, data ethics is introduced early in chapter 3, and training is split across introductory and more advanced chapters (the training introduction in chapter 4 is as an accessible an overview of training as you’re likely to find). There is content on image processing, latent spaces, natural language processing and text classification; despite the title there’s a neat overview of random forests. Later, the book goes into more advanced topics like ConvNets (CNNs) and Long Short Term Memory (LSTMs). The back half also covers some ML formulae written out in Python code, which is going to be more intuitive to programmers. One thing I really appreciated throughput the text were the references to academic and breakthrough papers. I’ll confess to having read this in review style (for work reasons) rather than as a textbook in depth. Nonetheless I’m looking forward to reading it in 2021 in more detail. It’s also nice to see a really good book covering PyTorch. Highly recommended for anyone getting into practical deep/machine learning.
The DynamoDB Book, Alex DeBrie. DynamoDB is maybe the most popular cloud managed non-relational datastore, one which I gather is heavily used inside Amazon. Givejn that, it’s a surprise there’s little to no general guidance outside personal experience and videos by Rick Houlihan. Alex DeBrie fills that gap. Not only is the book chock full of detail, and comes with complementary material, it’s well written. Important is that the book is opinionated while remaining engaging: DynamoDB requires a different approach to relational practice, and frankly does it not solve some things at all well. (Put another way, there are lots of ways to fail with DDB). DeBrie cuts through this by being unequivocal in his approach to design. The coverage of the single table pattern alone makes it essential for AWS practitioners, and I’d welcome more books by this author, who has a great voice.
Software Engineering at Google: Lessons Learned from Programming Over Time, Titus Winters, Tom Manshreck, Hyrum Wright. I liked this book a lot. It speaks to code in the long term. It’s a welcome counterbalance to the cult of the new and that much of our shared lore is informed by lessons learned from consultancy, which may tend to overfit on initial development and rewrite exercises. I’m unsure if I’m ready to recommend it as a general engineering text. It may take a few more passes to determine what is survivor bias, and what generalises outside one software organisation, albeit a very important and very influential one. That said, it certainly provides a different and helpful contemplation on software development and its maintenance. One for monorepo enthusiasts, if only to know what they’re getting into.
Invisible Women: Exposing Data Bias in a World Designed for Men, Caroline Criado Perez. I’m aware of some design bias from Industrial Design (eg crash test dummies being male sized), but the extent of bias against women in society outlined in this book is astonishing. The level of detail and research is relentless: page after page after page of examples converging on designs for the average man with serious consequences for women. Outstanding read.
One Strategy: Organization, Planning, and Decision Making, Steven Sinofsky, Marco Iansiti. At its peak, shipping Windows 7 involved about 4,000 people targeting about 450M users. During the time One Strategy’s collection of internal blog posts were written by Steven Sinofsky (which also cover Windows Live and Explorer releases), much of what we take as givens today about strategy and delivery at scale were not codified. As such this is a ring-side seat into someone figuring a lot of it out for the first time, live, with little to no point of reference to draw on. The present tense of the posts makes for a very different, and more dynamic read, than the typical business book written from hindsight. It shows how writing can scale to form alignment around a strategic vision. A problem with the book, the Kindle edition at least, is having to distinguish Sinofsky’s writing from Iansiti‘s commentary. This got tedious and it would be good to see a revised edition that demarcated the content. Nonetheless, well worth it for the insight into the Windows partner ecosystem (more generally how MSFT thought about partners at the time), the general theme of integrity across the posts, and the specific post “Cutting is Shipping”.
Algorithms To Live By: The Computer Science of Human Decisions, Brian Christian, Tom Griffiths. The book hinges on the idea that we encounter difficult algorithmic problems all the time, whether it’s figuring out how many people to interview, how to file and find books in a library, or even sorting laundry. The book covers a lot of ground—coverage of scheduling, explore/exploit and Bayesian/Laplacian prediction are highlights. Each chapter is grounded in real world examples. This is the best popular science book I’ve read on the nature of algorithmic computer science and why it’s useful. Brilliant book, highly recommended.
Six Impossible Things, John Gribbin. I picked this up as a vacation read. It’s mercifully short, for a poplar science book and you can read it in a few hours. It covers six plausible interpretations of quantum physics, and how each comes with the most bizarre, even unacceptable consequences. Clear and accessible explanation of the unexplainable. It’s been some time since I’ve read anything by Gribbin and this is a ringing reminder on just how good a writer and conveyor of physics he is.
The Ultimate Engineer: The Remarkable Life of NASA's Visionary Leader George M. Low, Richard Jurek. George Low worked on the space program from the late 50’s to the mid 70’s and was chief of manned spaceflight. Despite his accomplishments and the singular regard he was held in as an engineer, he is not as well known as he should be. Low was gifted, hardworking and accomplished. At one point he ran a 300 person organisation as a lead engineer, working on anything from contracts to funding to aerodynamics, something that’s difficult to comprehend today where we tend to split management, business and functional contribution into roles. One thing that rings clear is Low’s attention to real detail, and identify the right detail to pay attention to. The book is thorough and while the focus is on the space program, the arc of Low’s life is well captured in this biography. Jurek makes you feel like you saw into the man and his accomplishments without getting caught up in the hero worship and romanticism of the space race. The more I read about post WWII systems and engineering the more I think we’ve lost something on building complicated systems reasonably quickly. A wonderful and inspiring book.
Frank Ramsey: A Sheer Excess of Powers, Cheryl Misak. Frank Ramsey is not known nearly as well as he should be, given his contributions to mathematics, economics and philosophy. This is a good overview of his short life (Ramsey died at 26), that mixes his family background, life, and work, expertly. Cheryl Misak, an accomplished philosopher and academic in her own right and has done a fantastic job of covering his work at different levels of detail via panels written by experts from various fields, and painting a picture of a no-nonsense but largely likeable figure, without falling into the trap of treating genius as otherworldly. Even if your interest in philosophy and math is limited, the book is a good read on the Bloomsbury Set and post WWI Cambridge life.
Leadership is Language, L. David Marquet. Like Marquet’s acclaimed Turn The Ship Around, I’ve ended up reading this a few times this year, and have 40+ highlights and growing marked in Goodreads. Marquet continues with what seems to be his grand theme, how to cede control and decision making authority while increasing standards. What’s notable relative to the better known Turn The Ship Around, is the introduction of mechanisms and constructs that can be applied. The emphasis is on personal communication and not process, is what marks it out from the usual Toyota/Lean/Sigma fare. Instead, how spoken (and written) communication occurs is framed as a form of active leadership. Much said here is not new. Moving away from industrial command and control management is well trodden ground: what marks things out here is highlighting how habitual that style is. What’s novel is the focus on doing and enabling this through communication rather than process. Marquet has quickly become my favourite writer of this kind of business leadership book.
Blood Meridian: Or the Evening Redness in the West, Cormac McCarthy. I make no bones about being in the McCarthy is a titan camp and remember first reading Blood Meridian over twenty five years ago like it was yesterday. There is arguably nothing in fiction like the legion of horribles passage. The book is absent of psychology. There’s no inner landscape for the characters who are neither good or bad. They are just actions and words coursing through a biblical primeval west. The violence is often commented on but by modern standards that is overstated. What the book is is dense the way poetry is dense. It’s gruelling rewarding reading that will result in back tracking to re-read and appreciate passages—Blood Meridian is the Soulsborne of fiction. One of the great novels.
The Big Sleep, Raymond Chandler. It’s difficult for perception not to be dominated by Howard Hawks 1946 classic film, but the book is a classic in its own right. The novel’s Marlowe is more caustic, cynical and charmless, far less less admirable than Bogart’s cinema portrayal. Then again, the novel’s Marlowe has an infinitely richer mind’s eye, especially for the character of Los Angeles itself. Today it’s hard to tell whether Chandler is a master of the first person interior narrative, or that he’s been endlessly imitated in the 81 years since this, his first Marlowe novel, was published. Best read as literary fiction and not just the crackling detective noir it surely is. Still, just who did kill the chauffeur?