Monthly Archives: March 2018

Break it yourself

(With apologies to Andrew Bird. You can listen along.)

I give you a holy word:


(Great, now I have to apologize to Kurt Vonnegut, too.)

A few weeks ago I did something accidentally that I regret, but I think it's just as well that it happened.

That's the remainder of my time at Orbital Sciences, and the short first phase of my career. Ten years ago today was my last day there. I remember the last day, walking around, collecting signatures for a revision of the Orion Launch Abort System Jettison Motor spec. Then handing in my badge and driving off down VA-28. I quit that job to move to Texas for a girl. I love telling that story. I've got the whole routine down, all the way down to the mirthless laughter and asking the audience to be sure that write it down: don't do that.

Coincidentally, two days ago I bought a new car. I've had the other once since—drumroll—26 October 2006, four days before I started that job at Orbital. That's a less painful story. Why did I buy that car? Because the one the preceded it broke down in the Appalachian Mountains in western Maryland, and I got towed to a Pontiac dealer in Cumberland and traded it in.

What else is left? Maybe it's time to turn in this 703 area code on my mobile phone. What else is left? The bike that I wrecked when riding back from the office in 2007 got stolen in Burbank in 2014. I don't mind. It still had scars on the frame and seat. What else is left? Until recently I had a travel mug that I bought on my first work trip to Arizona even though it hadn't kept a proper seal in ages. What else is left? Somewhere I've filed my first paystub and my first performance review and raise (I think the percentage on that first raise is roughly the same as the sum of all subsequent raises).

What's the one thing I wish I had kept? I had an email from Dave Thompson, the CEO of Orbital, thanking me for an article I had written in the American Astronautical Society's magazine. Didn't think about taking that at the time.

I don't know if it's an American thing, or a male thing, or a youthful thing, or whatever it is, but I understand that the "correct" answer when talking about your life is: no regrets. I wouldn't change a thing. &c. Buddy, I'd change a lot of things. Leaving would be #1.


Honestly, after ten years, I thought that I'd have an Objective answer to the question of whether I should or shouldn't have. It would make this note a lot simpler to write, that's for sure, but it would also make the thoughts I have about it quiet down and disappear. It's not even an interesting thing to get stuck on. I mean, I know several people my own age who have died or almost died from breast cancer, killed themselves, died or almost died from plane crashes, etc. Those are problems.

("Dare the plane to crash / Redeem the miles for cash / And we'll dance like cancer survivors / When the prognosis was that you should have died")

Professionally, it's been a toss downhill since then. The job in Texas was bunk, even though I got to sit on a MER console for two Space Shuttle launches and landings. I stayed there for a year and three days (the moving bonus vested after one year on a Tuesday).

Moved to Lowell, Mass. Spent short of two years on that job after the company laid a bunch of us off. The job itself was so-so, but living in the Boston area covered for that, especially volunteering for 826 Boston. After the layoff, and the extended limbo after it, that was the thing that kept me alive, having some kind of external purpose.

And that interstitial period was strange. I can't fully account for it. I've thought before about examining it, considering it from different angles to see if there's any sense to be made, giving it the written treatment. It's like a pond that reflects the sky, giving no intimation of the world beneath until I dive into it. But I'd rather not. Swimming in fresh water gives me the creeps anyway. So I'll just toss it out here and leave it.

I spent 10 weeks in India. I climbed Elephant Tusk while backpacking around in Big Bend and the Guadalupe Mountains in Texas. I fixed the plumbing at the Panamint Hilton after heading up to Half Dome. I ran 100 miles (yeah, that first one, ran more like 70 miles) and qualified for Western States. Anyway, those were trivial, listable things. There was time together, but a lot of time alone.

There's no sense to be made from it all. No narrative. (Here's an interesting article: Galen Strawson, "Let’s ditch the dangerous idea that life is a story", Aeon Magazine.) But all of that time, sixteen months of it, does reside somewhere inside of me like a patch of cool shade. It's not a malevolent feeling, it's just like that small hitch before you breathe in and sigh.

Move to Wisconsin because that's where I got hired. Platteville. Maybe 11,000 people including the university. Two months in, before even my things were delivered from Mass to Wisconsin, get transferred out to our sister company in Sylmar, California—which is in the Valley, but a weird corner of the Valley where the vaqueros ride their horses unironically down the road—for two weeks. Then two weeks back in Wisconsin. Repeat. For two years. Go on a weird distance running binge. Meet Chen. Effectively move to St. Louis while commuting to California. Get married. Move out there to Burbank. Basketball on Wednesdays. Tianjin food on Saturdays. Go to China. Move back to St. Louis when she went to business school.

That last part seems chaotic but it's the stable part.

When I started this I didn't want to make a big list of things. I wanted to make sense of them. Because ten years ago I felt like the trajectory was pointed upward, like a rocket. I had a network. I was working on a human spaceflight project. Failure wasn't an option because I wasn't going to lose but because I just didn't realize it was an option. I miss that feeling—although I wouldn't have felt it at the time, so I guess it's more proper to say that I miss the position I was in. And I want it back and I don't want it back. I have that feeling more than you would believe. Things are good, but the magic of possibility isn't there. Maybe it would have been like that anyway. Maybe it's like the endurance running was—the magic of discovery in the first few races, discovering new places on the map of yourself, but most of the time and effort was in the training in between the races where there was usually the same old ground.

I thought I would come up with a clever ending to this while writing it, something Hollywood to say as I flung it away, but it never came. Fine, fine. It fits. I'll keep it.

When: peak, trough, and recovery

"...if we recognize that our cognitive abilities don't stay the same over the course of a day, that they change in predictable ways, and that when we do something depends on what we're doing, then we should be moving our analytic work to the peak, our administrative work to the trough, and our insight and creative work to the recovery. And it's that simple. And what the research tells us also is this: that time of day explains about 20% of the variance in how people perform on cognitive tasks."

—Daniel Pink, "Daniel Pink — How to Make Better Decisions and Be More Creative", The Tim Ferriss Show, 2018-03-26

I wish I knew what research Daniel Pink was talking about here. Obvious solution: reading his new book, When. I was going to avoid non-fiction for a while but there's a copy sitting at the St. Louis County Library, so...

Anyway. What he mentioned regarding the peak, trough, and recovery—if you've been alive for a full day or more, you could match these terms to their respective segments of your day without any further explanation—makes sense to me. There are periods of the day that are better for suited for different types of work. It's obvious. No one needs to be told that. But listening to the explanation revealed to me a bit that I hadn't considered: why not plan for those periods that you know will be there. Do the work that requires deeper thinking in the peak. Do the drudge work that doesn't require much thinking in the trough. Do the creative work in the recovery. It's obvious after being pointed out.

So: lunchtime run to the library tomorrow.

The particulars of a comprehensive entity

Speaking more generally, the belief that, since particulars are more tangible, their knowledge offers a true conception of things is fundamentally mistaken.

—Michael Polanyi, The Tacit Dimension

This is a problem we get ourselves into in systems engineering. The allure of specifying all of the ins and outs of a system to design is irresistible. Just one more requirement—that's all we need to complete the specification. Just one more, just one more. Just one more.

Just one more.

There are never enough requirements. Yet often there are too many. That's what this line, and the larger passage, from The Tacit Dimension reminded me of. Systems engineers sometimes go off on these quixotic searches for the perfect set of requirements. We all know it doesn't exist, but the drive to find it is still there. Somewhere on the journey, though, we pass some threshold where each requirement becomes better—perhaps more correct individually, or more precise—but the set of requirements becomes worse. The individual components of the design specification are beautiful, but the integrated whole is ugly.

When I'm in a good mood and I want to describe what I like about systems engineering, that's how I describe it: the individual pieces can be brilliant works of technical art, but if we can't make them work together then they're no good.

A week in review, 2018-W12





Alison Klayman, Take Your Pills


What would you do with an ROI of a million dollars?


The machine doesn't learn anything

Tom Simonite, "AI Has a Hallucination Problem That's Proving Tough to Fix", Wired, 2018-03-09.

I'm still chugging along in CS498 Applied Machine Learning. For good or ill. One of the surprising things to me about machine learning is that the machine doesn't really learn anything. There's no understanding. You try to decide which model works best, direct those algorithms at a pile of data, and then try to optimize some of the parameters in the model as the algorithms are graded on their predictions.

The machine doesn't learn anything. I guess, after just a little consideration, that point is incredibly obvious. But I'm susceptible, as are others I'm sure, to believing that machine learning and artificial intelligence are so powerful that they're basically magic.

From the article:

Humans aren’t immune to sensory trickery. We can be fooled by optical illusions [...] But when interpreting photos we look at more than patterns of pixels, and consider the relationship between different components of an image [...]

I just started reading a book called The Tacit Dimension by Michael Polanyi (it was a reference in Gary Klein's Streetlights and Shadows). I haven't finished it yet, but one of the leading ideas is: we know more than we can say. Tacit knowledge. And if we can't say it, we can't code it. And if it—whatever factor is relevant to our model—has some bearing on the output of the systems we work with, and we don't know how we know or what we know, then it's too much to expect that our machine learning algorithms can also know.

What were people expecting out of Facebook?

Cory Doctorow, "You can help the web be better in 2018: just ditch Facebook and use your browser instead", Boing Boing

Foster Kamer, "The 2018 internet resolution everyone should have: Forget Facebook", Mashable

I wonder: what were people expecting out of Facebook? I always looked at Facebook the way Kurt Vonnegut looked at existence in Man Without a Country: "We are here on Earth to fart around. Don't let anybody tell you any different."

I think that's a result of starting on Facebook back in 2004 in college, when you had to have a .edu email address to join. At the beginning it was just a text entry. "Kirk Kittell is" was fixed—a literal "what are you doing?" More or less, that's the baseline of what I'm looking for in the site. I just want to see what people are doing. 14 years on the site, and now I have connections with friends from high school, college, early work days, the summer at ISU in France, people I met while traveling in India, some of my wife's business school friends. It's clearly the best way to keep in touch with such a range of people. There is no similar substitute that's that easy. I don't think all of the people I know should make their own web sites. It might be interesting if they did, but even if they created them I think most of them would stop maintaining them anyway. And I don't think I'd be able to maintain checking all of them. That's life.

But as for one of the main points, stepping back to using the browser bar and bookmarks to get to content—I'm down with that. I've been getting back to it myself. Google killing Reader seems to have damped the use of RSS for news, but I get plenty of it through Feedly. Getting news from Facebook and Twitter is garbage. It's rigged. As it should be. Their job is to make you want to go to them for whatever it is that they're selling. That's obviously the pitch.

Back, back, back in the day there were webrings. Do you know what those are? Congratulations. You're old. They were amateur and interesting ways to find amateur and sometimes interesting sites. (John Scott, "Whatever Happened to... Webrings"; "What Ever Happened to Webrings?", Hover.)

I don't remember where I was going with this. Whatever it was, it devolved into some variation of: things were better when we were younger.

I don't believe it was all that much better—there was a lot of crap—but there are a few things I'd like back which have dropped off along the way. Web site discovery via other web sites is one of them. Webrings were super cheesy. They can stay on the shelf. But I think I'm going to bring back my /links page.

Automate and win

Today at work I gave a presentation to our division-wide systems engineering group at work about the why and how of automating some regular tasks that we do in our every day jobs. Over the past two years I had written some code (DXL, Python, VBA, etc.) to make some of the horribly boring and periodic work I do get itself done so I could work on other things I'd prefer to do. So many systems engineers turn into button-pushers and process jockeys. So many know the process but forgot all about the content. I don't want to be like that. Train some software to do the work, and you know the process and how to write software. Win-win.

It took a few years to learn how to code, and now I can use it. It was a painful—a terribly jarring experience for my ego. Kind of like getting out of shape physically. And like getting out of shape it's not hard to get back into shape. You just have to do it. Any idiot can do it—sometimes my life feels like an living testament to the ability of idiots to do things—but it takes time to do it. Unlike getting in shape, it's not clear where to get started. It's pretty clear how to run, but code? Where do you do that?

Anyway, the larger point is that it's not just that I don't want to be like a zombie, I don't want others to be like that either. At Big Corporation, when you start carrying that flag your life is a strange mix of being disliked by your manager (I got a below-average score on my performance evaluation for writing the automation code that saved the project and company money because it wasn't value-added to the project) and getting affirmative feedback from your peers (who know better than to rock the boat, but want the boat to be rocked a little).

I had some fun with the presentation. I can't post the whole thing here because it has some internal info, but I would like to share one slide that might be my masterpiece:

Look at that beautiful transition, from the depths of Lumbergh to the heights of what would you do if you had a million dollars.

And then there was this beautiful transition (you'll just have to imagine that the first three pictures were after a statement of the problem, and the fourth after the statement of the solution).

So it's a fine line between being clever and making a point. I've always tried, in these Big Corporate settings, to catch people a little off guard. There's a kind of protocol: stodgy, predictable, don't make me think too hard just give me some bullet points. Instead I'd like to kick in the door and let them know: there's a different way to do things. I'd like that to be my niche. I don't think one presentation establishes that, but I hope I planted a seed.