Aug 27, 2015   [permalink]

On Being Wrong

How inevitable an error?

Nobody is right all the time—not even you!—and depending on what you do for a living and whether you're a married male, you may actually be wrong more often than you're right. Of course if you wash dishes or mop floors to earn your keep, then as long as you show up for work and attempt to look busy, there's a limit to how wrong you can be. However, shame on you. Get a real job, pussy! Washing dishes is easy and even sort of fun, but there's someone out there with three kids and an IQ of 90 that needs that job way more than you do.

Now, actual technical jobs fall into two broad metaphorical categories, which I'll call "airline pilot" and "fighter pilot." As an airline pilot you have a known route to follow, and while minor variations may be necessary in the face of changing weather and traffic patterns, there's always a voice from the control tower telling you what to do. 99% of the time, the job does not require any real judgment or interpolation or creativity—just perfection. The other 1% of the time, you're in some sort of anomaly condition and have to fall back on trained-in responses. Engine malfunctioning? Shut it down and proceed to the Engine Out protocol. If you're fit to be an airline pilot, there's simply no excuse for being wrong on any of this stuff, and the consequences of an error can be very large indeed, involving death and destruction for crew, passengers and bystanders, grief and heartbreak for hundreds of families, and possible bankruptcy for the parent company. Only about 0.001% of the time do you really need to innovate—like when you fly through a flock of birds that smash the guts of all four engines at once—and most of those times you fail and die and no one ever really quite knows what happened. Very occasionally, airline pilots will land it in the Hudson River or eke out a 30% fatal crash in an Iowa cornfield, and retire as heroes—something they never aspired to be and probably still wish they weren't.

Airline pilot type jobs include journalism, teaching, production engineering, power grid management, the maintenance of legacy code, and most types of surgery, and if you're on one of these career paths it behooves you to train hard, show up sober, and fly by the seat of your shut-up-and-do-it-correctly.

If you're a fighter pilot the world looks rather different, because virtually everything you do is a one-off, and innovation is a flat-out requirement if you expect to survive. Missions are rarely repeated, and if you're a naval fighter pilot then not even takeoff and landing are routine, because the deck is always pitching and rolling and the boat is never twice at the same GPS coordinates. The job is inherently a lot riskier because even when people aren't literally trying to shoot you down, you're still dodging mountains and radar installations in real time, while hunting some tiny target against the vastness of Earth and sky. Taking risks is the whole point of the job!

On the other hand, most failures simply involve missing the designated target, and even catastrophic errors will pretty much just kill you and your gunner and a hundred-million-dollar aircraft. No medals, no lawsuits. Fighter pilots aren't expected to be perfect, and indeed other than a continued heartbeat and some kill icons painted on the side of the aircraft, there's no scale by which perfection could even be measured.

Fighter pilot type jobs include police work, commercial R&D, emergency medicine, oncology, writing software for novel applications, and anything involving the adaptation of new materials or new technology to existing industries such as architecture or vehicle design. There's a lot of trial and error, a lot of blind groping toward half-conceived goals, and yes, a lot of panicky escape maneuvers when you realize you've fucked up and have, like, forty seconds to get yourself out of trouble before the ground comes up and smacks you. There's a lot more glory in these jobs than in the airlines, but also more stress and heartache. Careers tend to be shorter, with messier endings and a much higher chance of coming away scared and humbled. Still, the same is true for professional athletes, and there doesn't seem to be any shortage of takers for those jobs. And as with sports, some of these jobs offer at least a chance of getting wealthy or famous or both, whereas with most airline jobs you'll be lucky to retire with your pension and IRA intact.

Of course, in the real world most jobs aren't pure airline or pure fighter, but somewhere in between. The intensity level may fluctuate over time and space, and you may spend a lot of time waiting around for something to happen before you even know which type of job you're in. Still, the distinction is a useful one, and if you know which one suits you better it's a major help in guiding important life decisions.

My first job out of college was a dream come true, designing the guidance and navigation software for a new generation of commercial rockets. We sent one to Mars, another to the moon, and several to geosynchronous orbit. It was pure fighter pilotry, and indeed, while the work was thrilling and rewarding it was also very demanding, and the team I was working with made a hundred-million-dollar error that eventually killed the project. Not me personally, you understand, but arguably I do share one fortieth of the blame. The history of spaceflight is littered with examples like this, so I'm in good company here, but it still sucked, and it's one of the things I'm least proud of in my life. It does not appear on my résumé. But when you're doing something new and big and exciting, that's the risk you run, and if nobody's willing to take the risk then nothing new will ever happen. C'est la vie.

In the wake of this failure, I found myself dumped into a very similar project, but for military rather than commercial customers. We were launching spy and communication and weather satellites, which is not so different except that they were mostly secret, so either we didn't know what we were launching, or we knew but couldn't tell anyone, which seemed a lot less glamorous.

More importantly, the software for this rocket had already been designed. It was mature, and needed very little maintenance, so each launch was a matter of feeding in the desired trajectory, running simulations to make sure nothing strange would happen, and then pressing the GO button. It was routine, airline pilot work, and over six years of increasingly unhappy effort I finally realized it simply wasn't the right place for me. So I left the relative security of Big Aerospace for a small startup company and never looked back.

Sometimes you need to dwell on both sides of the fence before you know where the greener grass really is. On the other hand, I've seen a lot of good people tear their hair out in fighter pilot environments, because their meticulous perfectionism clashed with the inherent uncertainty and short timelines of product development. A lot these folks would be perfectly happy as college professors or Fortune 500 IT managers, and I hope that's where they've ended up, because in the fast-paced world of technology development, if you know exactly what you're doing, you're underperforming.

More anon.

[ comments | add a comment ]