On Learning From Mistakes. And Successes.

Doh!They say you can learn more from your mistakes than your successes. I’ve recently come to the conclusion that this is bullshit. In my experience, if you’re motivated, you can learn as much from what you do right as from what you do wrong by deliberately thinking about what just happened. In the case of physical activities, the sooner you do so, the better.

Consider for example an athletic endeavor, say, soccer. If you kick the ball, and it falls short or curves the wrong way, you can learn by pausing and saying to yourself:

“Ok, self, what just happened? How was my body positioned? What was the angle of my foot, leg, my hips, and my chest? How much force did I apply? Where were my arms? Where was I looking? And which one of those things will I change to try making the next kick better?

On the other hand, if you kick the ball and it goes precisely where you wanted it to and with just enough power to get the job done, you can learn by pausing and saying to yourself:

“Ok, self, what just happened? How was my body positioned? What was the angle of my foot, leg, my hips, and my chest? How much force did I apply? Where were my arms? Where was I looking? And which of all those things should stay the same to make the next kick just as good?”

In other words, if you’re motivated to learn, your relative success or failure on any particular effort does not matter. What does matter is what you do with the information you’ve gained.

Tagged , , ,

Google’s Brilliant Employment Strategies

Hiring is a difficult and laborious process, generally because it’s very hard to determine which particular person will actually perform well on the job. (I spent years studying this in grad school, and have spent years implementing it in my office. Believe me – It’s very, very hard.) However, research suggests that there’s one trait that universally predicts job performance better than any other, and that’s general intelligence. Basically, the smarter you are, the better you’re likely to do on performance ratings. As such, companies that select staff based on how smart the applicants are will have employees with higher performance ratings, and if performance ratings are correlated with the marketplace performance of the company, hiring and retaining smart people should result in a stronger bottom line.

Consider for example how Google goes about getting new crew members: They hire brilliant people. Not only do they want to know where you went to school and if you graduated with a degree, they want to know what your GPA was while you were there. (Fortunately for my ego, I haven’t put in a resume.)

Picture of one of Google's self-driving cars

Credit: Google

 

Once they’re on board, Google keeps their brilliant people entertained with personal projects. “Google screws around with a lot of far out, even cockamamie projects — well, because it can and it’s fun. Two good reasons that few can fault.” It’s not hard to imagine that having one out of every five days of work to devote to your own ideas would be very intellectually stimulating for someone with the brainpower to have brilliant ideas.

Google provides services on campus so that their brilliant people don’t get distracted by the everyday minutiae like getting their oil changed or their hair cut. And they require all employees to get at least 120 hours of training and development every year, to keep them abreast of what’s going on in their field. (There’s also the unlimited sick leave and 27 days of paid vacation every year. And don’t get me started on the free gourmet lunches and dinners.) You put all these factors together, and you get a gigantic company staffed entirely by brilliant people that repeatedly gets acknowledged as one of the best companies to work for in America. ‘Nuff said.

 

Tagged , ,

How Less Government Is Actually More Government

It used to be that the government did things itself. Put a man on the moon? Form NASA and get to work. Spy on the Soviets? Create the CIA and start sending fake nuns to St. Petersburg (Leningrad for you historical types.) Foment a revolution? Send the marines into the Bay of Pigs. (Oh, wait — That one we didn’t do ourselves. Sorry.)

The U.S. Capitol buildingOne of the big, highly unappreciated advantages of this approach is that, regardless of the inefficiency and red-tape inherent in the process, it provides accountability, both before and after the fact. If the NSA or the FBI or anyone else in the employment of government wants to directly collect information about American citizens, they have to deal with pesky things like warrants, court orders, and civil rights lawsuits.

Nowadays, however, the government is doing far, far less in the way of project execution itself, and the implications should be deeply troubling to anyone who doesn’t want the NSA looking at pictures of their butt taken the last time they were on the crapper. Why? Because the NSA doesn’t have to take the pictures, of course! Not only are big companies like Google, Apple, Facebook, and every other one you’ve heard of collecting this data in massive quantities and providing it to the government, they’re actively discouraged from admitting they’re doing so. The NSA is even building a massive data center in the middle of the country, right next to twelve major data centers and several cross-country fiber optic networks, to make it easy for these companies to send the data their way.

But wait! There’s more!

Because these are private companies we’re talking about, they’re not acting as government agents. Ergo, all those little annoyances like search warrants and civil liberties are completely obviated. And as it turns out, they’re even willing to collaborate with their competitors to make it easier for the government to do one-stop-data-shopping.

Consider for example this article about U.S. mobile phone carriers planning a massive centralized database of cell phone numbers and IDs. The purpose is ostensibly to make it easy to shut off services to a phone reported as stolen. However, it wouldn’t surprise me if this system were also used by the government to quickly and easily find any particular cell phone user in the country, without having to figure out which carrier he or she is using. (Also note how this article hit the papers shortly after the government lost the ability to do warrant-less location tracking via GPS device, which will likely eliminate their ability to do so via the victim’s suspect’s cell phone.)

I’m all in favor of a smaller, less expensive federal government. However, if we don’t introduce some accountability into this system of the feds using private companies to do their dirty work, we’re all going to live in a far less free society in the very near future.

4/27/2012 UPDATE: Slashdot is covering a story about how the FBI has an office located inside a non-profit whose specific mission is to pass information about us and our online activities from private companies to the feds.

You Get What You Measure

One of my graduate school professors was well known in our class for what he called his Behaviorists Bowling Team shirt. According to him, it says “You get what you measure” across the front, and “The best predictor of future performance is past performance” on the back. The second saying has only limited application (if it were completely true, there’d be no point in practicing something, as you’d never get any better.) However, the problem of getting what you measure is like the third law of thermodynamics: you can’t get out of the game.

Consider for example this story about police surveillance cameras being installed in Toledo. As the victim of three break-ins (one attempted), I think the cameras are a good idea to counteract the effects of fewer officers patrolling the streets. Data-driven policing, however, is another matter — search for “compstat” on this page and you’ll see what I mean. The short version: if the data show the frequency and severity of crimes going up, the pressure comes on to bring the numbers down. Unfortunately, there’s two ways of doing so: by arresting and prosecuting the criminals (hard), or by under-reporting crimes and/or intentionally mis-classifying them as less severe than they actually are (easy.) Guess what actually happens.

Hello Robots, Goodbye Civil Society

The TerminatorRobots can operate 24/7/365. They don’t complain about the work environment, don’t ask for raises, and don’t get paid extra for overtime. They don’t form unions, ask for promotions, require health insurance, or play office politics. They’re never late for work or inappropriately dressed, and they’re patently incapable of theft, sexual harassment, or discrimination. And they’re getting smarter, faster, and (here’s the important part) able to do very complicated things only humans could do before. It’s my prediction that, much sooner than most people realize, there will be a robot designed to do virtually any job that currently requires a high school education. And they’ll be cheap enough that leasing enough of them to replace hourly workers will be a no-brainer for business owners.

Consider for example this recent article in the WSJ about the use of robots in hospitals. They are already replacing the janitorial and housekeeping staff because they cost half as much. I’ve been reading in IEEE’s blogs about robots that can build structures (even from the air), stock warehouses,  and drive cars, and they’re working on farming robots that can identify and pick ripe fruits and vegetables. It doesn’t take a rocket surgeon to see where this is going.

If it turns out that some jobs are just too difficult for a robot, telepresence robots will allow low-wage workers in other parts of the world to do manual labor in the U.S. from wherever they are. Will there be opportunities in robot maintenance? I’ll predict that there will even be robots that can diagnose and repair other robots in the not-to-distant future.

When you put this phenomenon together with the incredible imprisonment rate in the US, there’s going to be an awful lot of poorly-educated young people with criminal backgrounds who have nothing better to do than sit around and get pissed that they have no future. Put that together with the massive quantity of firearms in this country and the relative ease with which they can be obtained, and the only reasonable conclusion is that we’re sitting on a powder keg. We need to start thinking now about what we’re going to do about it.

UPDATE: Robots can now debone chickens, according to the Wall Street Journal.

UPDATE: Robots can now construct homes and buildings, according to Geek.com.

UPDATE: Robots are taking over manufacturing jobs in China, according to R&D Magazine.

Save Your Comments For The Blog

I had an epiphany the other day, right when I was about to write up an extensive comment on an article at another site. It occurred to me that, if it was important enough for me to go through the hassle of opining somewhere else, I might as well make it a blog post instead. As an added bonus, this obviates the problem of having to come up with material, the lack of which (for those of you who haven’t tried this yet) presents a serious challenge to a blogger.

Consider for example this article, where the author argues that data derived from split (or A/B) testing “never lies.” I was about to lay down the law* in a comment about the philosophy of science (knowledge of which I acquired through numerous arguably wasted but reasonably entertaining years in grad school), when I realized that I’d be spending an inordinate amount of time creating material that a) very few people would read, and b) fewer still would associate it with me. Whereas, if it’s on my blog, presumably visitors are there to read the articles, and it’s easy for them to find out who I am.

For those of you who aren’t bloggers, I’d encourage you to leave your comments below. The rest of you should know what to do by now.

*Data without theory, while interesting (and a useful starting point) is essentially meaningless, because you can’t replicate the results. In other words, just because you flip a coin and get heads ten times in a row, you might not want to bet against tails simply based on the outcome of your previous observations unless you have a specific reason (such as you suspecting the coin has two heads) to believe you’ll be correct. If you don’t have a belief to be falsified (e.g., “I bet they’ll like this design best because…”), you can’t make forward progress in ferreting out the variables that determine why a particular design is better than another, and you’ll end up going around in circles on design decisions.

See what I mean? Blogging material!

Tagged , , ,

The Lack of a Common Threat

If we all want to get along well enough to succeed, we need to be under threat.  Not just any threat — We need a common threat.

Consider for example the US Congress.  Back in the good old days when I was growing up (70’s and 80’s), they got along pretty well.  Not well enough to avoid yelling at each other, mind you, but well enough to compromise on just about everything that required passage (“required” as in ‘the nation will fail if we don’t do this’.)  Of course, they had a good reason: if we didn’t keep up the appearance of strength, the Soviets would blow us to smithereens.  Several times over, in fact.

There’s nothing like mutually assured destruction to focus the mind.  It makes it much easier for a congressperson to say to him- or herself “I may not like this solution, but if I can’t find a way to get along with the other side, we’ll be exposing a vulnerability in our clearly superior democratic, capitalist system.”

Nowadays, we (and in particular, our government) aren’t under a common threat, so there’s no incentive to compromise.  And until we are, the political gridlock and partisan warfare will continue to be the norm in Washington.

Tagged , , ,