First off, a brief apology. I made a New Year’s resolution to finish at least one post a month, and, well, you can see how well that ended up. Back in January, I changed jobs, and I’m now CTO of a company that’s building a platform for digital cash as humanitarian aid. I wasn’t expecting my available Copious Free Time to increase, exactly, but I was sort of hoping it wouldn’t be quite so much of a Major Life Transition. Ha ha!
People will tell you that managing engineers is nothing like being an engineer oneself. This is mostly true, but like most important career lessons, the ways in which it’s obviously true are its least interesting aspects. Management is, more than anything else, a change in perspective. As an engineer, you’re behind the wheel of your project, and it’s your job to get it to the finish line. As a manager, your job is to get your entire team to the finish line, and their projects assembled into whatever Voltron configuration your specification calls for, on time and on budget.
This isn’t where all conflict between engineering and management comes from, but it sure does account for a lot of it. The engineer’s top priority is getting it done right. The engineer’s manager also cares about getting it done right, but cares more about getting it done at all. When these priorities come into conflict, so do the people they’re attached to.
But differences in perspective aren’t all bad. They can obscure, but they can also illuminate. A Hindu folktale illustrates this:
Four blind men, making their way along one day with their canes and dark glasses, encountered an elephant. One of them touched its trunk, one touched its leg, one touched its side, and one touched its tail.
“An elephant is like a python,” said the first blind man.
“No, an elephant is like a tree trunk,” said the second blind man.
“No, an elephant is like a wall,” said the third blind man.
“No, an elephant is like a rope,” said the fourth blind man.
And they never could agree on what an elephant was really like.
The blind men can’t agree on what an elephant is really like — but the person who hears the story can listen to all four blind men’s perspectives and synthesize them into a more complete mental image of an elephant. The arguing men are foils for the listener, who is meant to draw a lesson from the characters’ folly.
A couple of months ago, I was on the phone with one of my teammates, a guy I’ve known for about twenty-five years. (We lived on the same street in high school, were on the debate team together, and stayed in touch over the years.) He’s an expert in geographic information systems and data collection, and I hired him to design and implement a simulation engine that we can use with real-world map data. This lets us test our application in a simulated version of the environments where we’ll actually deploy it. He’d been focusing on the geographic and econometric aspects of the simulator, whereas I was looking at the project in terms of its components.
“We’ll need a network simulator,” I said, “to be able to model what happens under all different kinds of network conditions. And the econometric simulator, which you’re already working on. And a traffic simulator, to model people moving around on foot or in vehicles. I found –”
“Holy shit,” he interrupted. “I can’t believe it.”
“You’re absolutely right. I was really starting to worry about how we were going to keep track of all these agents, because the econometric simulator is already starting to get a little top-heavy. But you’re right, we can use a traffic simulator as a separate component. I can’t believe I didn’t think of that.”
I have had these kinds of kicking-myself moments before, and they are no fun. The polar opposite of fun, really. But if I could give him my perspective, I could end the misery.
“Don’t sweat it,” I told him. “There’s an Alan Kay quote that applies here: Point of view is worth 90 IQ points. You’ve been working on this from the bottom up, down in the gory details of spending patterns and shape files. I’ve been looking at it from the top down, thinking about how to put it all together. I saw it because of the direction I was looking at the problem from, that’s all. So don’t feel like you’ve messed up by focusing on just that one tree and not seeing the forest. That tree is your baby. Taking care of the forest as a whole is literally my job.”
Maybe that sounds like one of those meaningless reassurances that’s only supposed to make people feel better. It is, however, entirely true. If I’d had a head full of facts and figures that I were trying to distill into patterns a simulator can replay, I probably wouldn’t have seen it either. Nor is top-down necessarily the point of view that gives the effective-IQ boost, either! Just as often, a person looking at a problem from the bottom up can identify the optimal substructure that makes a large problem easy to break down and solve, even if from the top down it looks like an implacable monolith. Nor, for that matter, is structure the only useful change in perspective! Switching from near mode to far mode, or vice versa, or between System 1 and System 2 or vice versa, can often bring details you were missing into sharp relief.
Let’s talk about that sharpness in a minute. First, though, a digression.
I made one other resolution this year: Get better enemies.
By and large, individual people make pretty terrible enemies. Our monkey brains are wired to experience prevailing over an opponent — even just a perceived opponent, and possibly even just a perceived prevailing, if social media is any indication — as pleasurable, but the pleasure is short-lived. You won, great. Now go do everything else you were neglecting in favor of the fight.
If that individual person also happens to be a figurehead, then the triumph is even shorter-lived, as the United States military keeps failing to learn in the Middle East. Synecdochically conflating Osama bin Laden with radical Islamic terrorism made for compelling prime-time narrative, but lopping off that head — or any other, for that matter — has utterly failed to kill the hydra. Any organization with a bus factor of 1 — that would collapse if a single key person were disabled or eliminated — is by construction fragile. It’s easy to want to think that social organization is simple and orderly and key persons are easy to identify, but, well, it’s nice to want things.
Once is chance, twice is coincidence, three times is enemy action: The other kind of individual who makes a terrible enemy is yourself.
Challenging yourself is one thing. That’s friendly competition. But punishing yourself — especially for things like “not being omniscient” or “not having all the answers” — is self-sabotage. You can motivate yourself with a stick, but you can beat yourself up with one pretty effectively, too. When someone else’s perspective illuminates a solution so sharply it makes you wince, if beating yourself up is what you’re used to, it can be really tempting to snap off one of the sharper, pointier bits, fix it to the end of that stick, and see whether stabbing makes for any more effective motivation. (Spoiler alert: It doesn’t. Not in the long run, anyway.)
If individuals make poor enemies, what else is there? I said “three times is enemy action” above, so I must be thinking of something as an enemy. And I am: all the many and varied opportunities to take on someone (yourself included) as an enemy that day-to-day life throws at us are just one kind of attractive nuisance.
Just as the road to hell is paved with good intentions, the road to crappy local maxima is paved with attractive nuisances. Which kinds of nuisances are attractive to which people is a highly idiosyncratic matter; de gustibus non est disputandum. Under conditions of information overload, which we’ve been in for decades and are probably never getting out of barring civilization-scale collapse, any conflict with an attractive nuisance is necessarily a defensive, specifically an evasive, action. You defeat an attractive nuisance by avoiding it, because if you confront it, it’s reeled you in to expending your limited attention interacting with it. Another thing that is Literally My Job Now is helping the rest of my team avoid their own attractive nuisances, if they ask. Which they actually do, sometimes. I’m taking that as a cue that I’m doing something kinda right there, at least so far.
Having entered the domain of Conflicts With Abstract Concepts, it merits thinking about which kinds of abstract concepts make for quality enemies and which ones don’t. The ongoing Wars on Drugs and Terror have been such expensive, miserable failures, with such enormous negative externalities, that to do otherwise would also be self-sabotage. LBJ picked poverty, and mobilized vast federal personnel and materiel resources, to mixed results, and by the middle of the Clinton administration his war machine had mostly been dismantled. On the “fewer negative externalities” end of the spectrum, the War on Cancer continues to fuel basic research, and although much of what we have learned about cancer in the last 50 years amounts to “JFC we have a lot more to learn about how cancer works,” overall incidence rates have gone down and overall survival rates have gone up. This is not a lot of examples from which to generalize, but when you’re spending billions of dollars annually on actions that gain you little or no ground against an abstract adversary, perhaps it is time to take a step back and reflect on what the last half-century of Wars On Things has taught us about target selection.
But this is also a very black-and-white way of looking at the situation. “War on X” has become a metaphor for “committing all available resources, even to the point of austerity, in the effort to defeat X.” It evokes the concerted nation-scale efforts of the World Wars that were still well within living memory for many voters when the term “War on Drugs” was coined in the 1970s. Yet already by the 1970s, and even more so today, actual warfare has evolved into something much more asymmetric. Total war loses against asymmetric warfare, by attrition: eventually the commitment required to keep up the state of total war is simply too costly for the warring actor to bear.
Not every fight has to be a war. If the lessons of history are any example, perhaps a better metaphor is guerrilla action.
This new phase of my career, in effect, is going to be spent in guerrilla action against poverty, famine, and unnecessary complexity in systems design. We are going to be constantly ambushed by nature, both the hurricanes-and-earthquakes kind and the human kind. And even though some of us blind men have felt their way around nontrivial amounts of the elephant, enormous swaths of elephant remain as yet unperceived.
What the blind men in the story forget is something foremost in a guerrilla’s mind: they can move. Which, in the realm of ideas, translates into perspective-taking. The corollary, then, to Alan Kay’s maxim is that if you find yourself feeling like you’re down 90 IQ points, it’s time to start looking for different points of view.
So that’s what we’re doing. We’ve found, these last few months, that incorporating a broad diversity in points of view — both practical and ideological — from different people who focus on different things, leads to a much more effective concentration of effort. It allows us to pay attention to the micro and the macro at the same time, without disregarding anyone’s contribution to the task at hand. It’s easy, from the 50,000-foot view, to get stuck trying to perfect aspects of a system that will never become relevant in practice. That system’s going to have to work in an environment, after all, not a vacuum. Actual, practical solution-making is a work of coordination, both with respect to the problem you’re solving and to the people you’re solving it with. If you can avoid the attractive nuisances that try to sidetrack you — as a first approximation, simply disregarding them works far better than it has any right to — then you’re well on your way to optimizing your own inventory of perspectives in a way that maximizes your ability to use the resources you have.
Give me some time to finish figuring out how to optimize the way I look at time, and I’ll have plenty more to say about all of this. Hopefully sooner rather than later.
Thanks to Giancarlo Sandoval for his editorial advice.