Recommended Reading

I ‘ve realized that there are a set of books I’ve been recommending to others pretty consistently since reading them, and a recent spate of doing so made me think to post them here.

The first book on my current “everyone should read this” list is The Righteous Mind, by Jonathan Haidt. I’m unlikely to think about moral decision making and differences of perspective and prioritization the same way after reading it. Early in the book Haidt describes our system for “moral decision making” using the analogy of an elephant and a rider. The elephant – our intuitive decision making system, akin to what Kahneman calls our “fast” system – leans toward things it likes and away from things it doesn’t. The rider – our rational (or “slow”) system, he says, isn’t there to steer the elephant rather to explain the elephant’s actions. So when faced with a moral conundrum we make a decision, then construct post-facto rationalizations in support of that decision. The book goes on to explore how our decisions and rationalizations are shaped by the different weights we place on a set of core “pillars.” Thoughtful and thought provoking through out.

The next book up is shorter, older and a bit harder to find – Albert Hirschman’s Exit, Voice and Loyalty: Responses to Decline in Firms, Organizations, and States. Written in 1970, it remains both relevant and seemingly the definitive writing on the subject. Hirschman argues that when we find ourselves in a relationship with an organization (company, club, country) we come to disagree with, we have only two real tools at our disposal; voice (raising our voice to cause change) and exit (severing our ties with the organization). The book spends its scant pages exploring how we choose to apply those tools in different situations, the impact that loyalty has on our decision making, and how different “ratios” of voice and loyalty affect the organization in question.

Last for the moment is Normal Accidents: Living With High Risk Technologies, by Charles Perrow, first published in 1984 and updated in 2011. Perrow looks at systems in terms of their “interactive complexity” and “coupling.” Interactive complexity is about the number and degree of relationships between parts of a system, and coupling (tight or loose) is about how easy it is for failures in one part of the system to cause cascading failures in other parts. I found the core ideas to be an enlightening lens to look at all sorts of systems through.

It is what it is

As someone utterly incompetent said recently, “it is what it is.”

A month ago, I characterized London as “slowly creeping toward normal.” Today, London has been placed on the COVID “watch list” – which as far as I can tell means Boris and Co. will sit and watch as we cede the gains made in containing the spread of the virus.

It ended up taking just about a month to get my BRP back from the DVLA, and a few days later my provisional driving license arrived. License number in hand, I was able to schedule a theory test. I took the first available, which turned out to be over a month out – the 3rd Monday in October. So I’ve got plenty of time to forget (and review) the material. And, of course, I can’t schedule the practical exam until I’ve passed the theory test. This country could stand to learn a thing or two about pipelining.

Once I had my license, I reached out to a handful of driving instructors. Most of them were either on sabbatical or were fully booked. I ultimately booked an hour with one who was neither. We got on reasonably well, and he was good at pointing out habits (like palm steering) that I’ll need to suppress for the exam. Once I can book a test, I’ll book a test, and he and I will spend another couple lessons in his Kia before I sit it. At this rate, I’ll be lucky to have it done by Christmas.

At his suggestion, I took a couple practice “hazard perception tests.” They show you a video from the point-of-view of a driver, you have to click on the screen when there’s a “developing hazard.” Your score depends on how early you recognize the hazard and react. These tests were clearly not made for motorcyclists. My first attempts I scored nul points – each of my clicks was just before the scoring window opened. Of course that farm equipment traveling parallel to the road is a hazard, waiting for it to turn into my path seems counter-productive. So before taking the test I need to practice reacting later. Somehow that seems very British.

Life has otherwise found a routine. I’ve been running three days a week, though we’ll see how devoted I am as the wet winter starts. Dawnise does the weekly shop and I help her mule it home. With apologies to Casablanca, “she buys the food, I cook the food, we eat the food. It is fairly convenient.” Most weekends we visit our local cafe for brunch, and we read the news from America with a mix of frustration, sadness, fear and resignation.

I’ve continued keeping track of noteworthy COVID-related articles I’ve read, though there have only been a handful in the past month that have made the list. And I’ve posted a few other thoughts that haven’t been about “life in London” and so haven’t been sent to this list.

So I’d say we’re still doing well, overall. We’ve had some rough days, for sure – when it’s been hard to remember we’re both on the same team – but more good days than bad ones.

I’ve started, several times, to write something about the election, or American politics more generally, but each time I’ve tried I end up deleting the draft and walking away. I’m at a total loss about how to have positive impact on any of those issues, but it’s reasonably clear that shouting into the electronic void won’t change anything.

So we’ll vote (absentee, it’s much better than mail in voting). And hope that enough of the country agrees with our perspective to vote with us. And hope that the popular vote carries the day. The whole thing seems dangerously close to the definition of insanity.

Doing the same thing, expecting a different result.

Tit for Tat: American Democracy and the Prisoner’s Dilemma

It should come as a surprise to no one that the party in power in the US is pushing to install a new supreme court justice with haste – before an election they stand a non-zero chance of losing.

The fact that this same group vociferously and successfully obstructed a similar appointment four years ago is being broadly cast as “evidence of hypocrisy.” If it’s hypocrisy isn’t the point, the point is they understand the game they’re playing, and are playing to win. Call it (calling out?) hypocrisy doesn’t change anything.

The prisoner’s dilemma, in case you’ve forgotten, is a simple game with rules often phrased something like this:

You and a partner-in-crime are arrested. You’re kept apart, unable to communicate. You’re both told that if you both admit to the crime, you’ll each face 3 years incarceration. If you both stay silent, there’s sufficient evidence to convict on a lesser charge, so you’ll both face one year in jail. If one of you implicates the other, who stays silent, the silent implicated party will be jailed for 5 years, and the betrayer set free.”

To people who study games the window dressing (crime and jail) don’t really matter – what matters is the relationship between cooperation, betrayal, and rewards.

It should be evident that both players end up best if they cooperate and stay silent. But if one party betrays the other, the betrayer suffers no penalty and imposes a larger one on the one they betray. So for each individual player, betrayal looks like a better outcome than cooperation. The net result is that mutual betrayal is, in a sense, the most likely outcome.

A single play of the game is interesting as a study in trust and self-interest. And individual political outcomes are often intuitively understandable by treating them “roughly” as a prisoners dilemma. But what we should be thinking about are repeated plays of the game.

Implicit in the relationship between cooperation and self-interest, and often overlooked, is the equality between the actors. Betrayal from either actor penalizes equally. This equality means “if you hurt me, I can hurt you.” And so it’s fairly easy to convince ones self that if you’re playing the prisoners dilemma over and over against the same opponent, a good strategy might be to initially cooperate, and then “do unto them as they did unto you.”

Tit for tat.

Intuitively most of us would call this strategy “fair,” and it turns out to be both simple to describe and highly effective at achieving cooperation.

Tit-for-tat clearly isn’t generally possible in American politics, due to party power dynamics if nothing else – but I wonder how the party and partisans would behave if it was.

Always ask why

The family story (“legend” just sounds pretentious) is that one of my most valuable, and likely most irritating, habits was instilled and encouraged at a young age by my uncle Denis. “Always,” he advised, “ask why.” I’ve always imagined it must have caused my parents some consternation that his was a direction I chose to listen to.

Fast forward a bunch, and I’m working at one of the A’s in FAANG (the river, not the fruit). Like any large company, it has a culture. Some of its norms and customs have been instituted deliberately, others have accreted, all have evolved. Adherence to those customs is a large part what it means to “be one of us.” While I disagree with some of those customs, in the aggregate I think they’re positive. One such positive, in my view, is that we look critically at our mistakes, try to understand root causes, and to capture those observations and lessons in written form and publish them, so that others might learn vicariously. And part of that process is to ask the five whys.

The nuance, and the value, in the five whys approach is is asking ‘the right’ question at each iteration, stress-testing the answers, and following the chain of causality, even if – especially if – it leads in unexpected directions. When it works, the illumination can be blinding.

So what, I hear you ask? Well, I’m reading more and more in the “main-stream media” asserting that social media may just might be hurting us. What I don’t read much of, if any, is thoughtful analysis of the how and why. Mostly, the argument goes, it’s because they have all the data – and with enough data, “they” (whoever “they” are) can make us do what they want. Some sort of social-media mass hypnosis.

This, to put it crudely, sets off my bullshit detector.

I don’t think it can be wholly explained by Facebook, or Twitter, or Mark Zuckerberg and all the data they have about us. They have the data we give them. If an individual tends to engage with content that’s dishonest, narcissistic, misogynistic, xenophobic, or antiscience – the algorithms didn’t make them do it. The person is making the choices. Clicking the links. Clicking some more.

I also don’t think it’s Trump. Don’t get me wrong, he’s pathologically dishonest, narcissistic, misogynistic, xenophobic, and antiscience, but he didn’t create any of those things. And regardless if he was elected to office because of those traits or in spite of them, he didn’t invent anything. He tapped into something existing. Exploited it.

So the algorithms feed us what we show them we want – it’s just operant conditioning – and we’re conditioning the algorithm. Living in a filter bubble certainly doesn’t makes us better, more rounded, people – but feedback loops amplify an input signal. They don’t create it.

So where does that leave us?

If you’re one of those people, like me, who have concluded that social media is a clear and present danger to democracy, ask yourself why. And keep asking. Because you can’t fix a problem until you understand it, and I believe we need to fix this problem.