Blink - Malcolm Gladwell

What a fascinating book! As in my two previous reviews, Malcom's Gladwell's Blink is a random choice from my mom's bookshelf. I recommend reading it with an open mind, then reading others' criticisms.

Blink explores the psychology of the adaptive unconscious, the part of our subconscious that makes first impressions about people and situations in fractions of a second. Pairing psychology research with biographical stories, Gladwell explains how we make intuitive judgments based on the tiniest "thin slices" of information. The book covers a tremendous variety of astounding successes and abject failures of the adaptive unconscious in cases such as art museums, battlefields, food tasting, car sales, and police operations.

Given that the book is 15 years old, some of the research from which it draws conclusions is now out of date, most notably with respect to the Race IAT*. On the other hand, per the book's Wikipedia article, criticism contemporaneous with Blink's publishing focuses on the idea that the book could mislead a reader into thinking of the subconscious as magical and omnipotent (with a bit of training). I think these latter criticisms are overblown, as Gladwell goes to great lengths to explain how and why the adaptive unconscious is often unreliable. In the examples where the adaptive unconscious is successful, Gladwell explains the massive amount of conscious, dedicated training and effort that went behind it.

By comparing so many applications of rapid cognition, Gladwell is able to synthesize key lessons that apply both to individuals and to society. These are summarized in the afterword, and I'll condense them further here. First, rapid cognition is extremely brittle, and can lead to catastrophically poor judgment if we are stressed and untrained; the corollary is that expertise built through years of experience (i.e., training) enables one to be make good rapid judgments less clouded by stress. Second, too much information can overwhelm our intuitive judgments and lead to poor decision making, as Gladwell illustrates through Millennium Challenge 2002 and the Battle of Chancellorsville. The book, then, urges us to combine these lessons to act, by identifying implicit biases and changing the circumstances in which they cloud our judgment.

I'll elaborate more on this third point, because it is the most exciting conclusion of the book.  Gladwell illustrates the idea beautifully with an example from classical music. Women used to make up a tiny percentage of musicians in orchestras. In the past few decades, by have musicians audition behind a screen, conductors have been forced to focus on musicians' skills and not their appearance. This has led to a massive increase in the number of women in orchestras purely based on merit. In other words, with a bit of conscious thinking, we can negate implicit biases, without necessarily relying on affirmative action.

Just like The First Idea, I think this book is critical to read for people like me who work in engineering and robotics. It is well known that these fields are dominated by White and Asian men, and one way to interpret this is as a result of implicit biases. By going through non-political examples, Blink presents implicit bias in a less confrontational way than one may experience in sensitivity training. Furthermore, Blink offers a glimpse of a path forward for ensuring merit-based promotion. In other words, I think this book is a gentle way to ruffle one's own feathers and thereby inspire positive change. For me personally, it has made me more curious about how researchers, such as Prof. Ayanna Howard and Dr. Kenneth Gibbs Jr., are concertedly trying to open engineering to a more diverse set of people.

Altogether, I strongly recommend Blink, and hope it jolts your curiosity about your own mind as much as it has mine.

Update (10 Dec 2020): Check out Charles Isbell's awesome NeurIPS keynote for ways that big thinkers in AI and robotics are identifying and addressing implicit bias. One of the lines in the talk, from Cynthia Rudin, is basically exactly the lesson of Blink: when we identify biases, we have to address them.

* Implicit biases certainly exist, but elucidating them through rigorous study remains challenging. Gladwell cites this famous 1998 study about the Race IAT by Greenwald et al. Many others have attempted to reevaluate IAT significance in the last two decades, most notably Prof. Philip Tetlock and coauthors. Separately, one interesting study from the Netherlands attempted to remove confounding factors and found that the racial bias may not be the main factor at play. I think that Greenwald himself is sometimes frustrated with media portrayals of IAT-based findings, especially in the context of sensitivity training programs. It seems undeniable that our culture and implicit biases have changed over the past 20 years. But, I think Gladwell's main point holds: when we can identify such biases, it is our responsibility to change them.