Robots, Railguns and Rube Goldberg

I'm starting a weekly video feature over at the teaching blog (part of a bigger educational media project, but more on that later). For now I'm collecting fun little STEM videos that teachers can start off the week with or use as filler or as writing prompts.

The purpose is generally more inspirational than instructional. Unless it fits in with the day's lesson plan, it is difficult to get across a complex scientific concept in a five minute YouTube clip. You can, however, get across quite a bit of cool, particularly with a young audience.

I'm looking for short videos with big visual impact (visual enough that most can be played with or without sound). They should appeal to a broad age range (ideally K through 12) though probably for different reasons -- we'd like a high school physics student to watch at the first video and think about majoring in engineering; we'd like first graders to watch at the video and think "wow, a robot cheetah!" (because the second reaction has a way of leading to the first).

Right now, the candidates are mostly from engineering but I'd like to broaden that later. I've been mainly looking at MIT and IEEE. There is also a TED, which somewhat violates my principles but when you see it, you'll understand. I'm also looking for off-beat Rube Goldberg devices and pre-industrial tech. I'm not crazy about the two historic examples here. I love the tech but the videos leave something to be desired.

Does anyone have any suggestions for filling out this list?





Robotic Cheetah








Magnetic Hair












Squishy Robots








Three Strokes of Upward Lightning







Small cubes that self-assemble




Soft autonomous earthworm robot at MIT







A Swarm of One Thousand Robots





Atlas Human-Powered Helicopter - AHS Sikorsky Prize Flight






Levitating Superconductor on a Möbius strip





















Soft Robot Uses Explosions to Jump







1,000,000,000,000 Frames/Second Photography - Ramesh Raskar







OK GO








Hero's steam engine






World record trebuchet at Warwick Castle

The perils of bell curving

This is Joseph.

As people probably know, I tend to be a bit of a Marissa Mayer fan.  However, this one item from a recent piece is definitely something that I would prefer not to see in a company that I worked for:

Mayer also favored a system of quarterly performance reviews, or Q.P.R.s, that required every Yahoo employee, on every team, be ranked from 1 to 5. The system was meant to encourage hard work and weed out underperformers, but it soon produced the exact opposite. Because only so many 4s and 5s could be allotted, talented people no longer wanted to work together; strategic goals were sacrificed, as employees did not want to change projects and leave themselves open to a lower score.

The problem with this sort of approach is twofold.  One, it makes people reluctant to join elite teams and groups (where they get to be rated as sub-average) instead of letting these teams mentor and nurture up and coming talent.   Two, since any ranking is partially a political process (only so much of the data can be objective, especially since the manager presents the evidence), it encourages political infighting among the managers of closely related teams.

Neither process is ideal.  Furthermore, you have to pay extra to compensate for the fear and uncertainty around a low ranking (base salary gets more important when bonuses can be variable). 

Now I am not saying that this process cannot work effectively under some circumstances, but it has notable downsides that need to be thought carefully about. 

Modern War

This is Joseph.

Paul Krugman:

Angell’s case was simple: Plunder isn’t what it used to be. You can’t treat a modern society the way ancient Rome treated a conquered province without destroying the very wealth you’re trying to seize. And meanwhile, war or the threat of war, by disrupting trade and financial connections, inflicts large costs over and above the direct expense of maintaining and deploying armies. War makes you poorer and weaker, even if you win.

I think that this insight is something we should think about a lot more.  In the modern and interconnected economy, even the peaceful absorption of a state (think East and West Germany) can be fraught with difficulty.

Once imperialism looks cost ineffective, it really does change the best ways to deal with opponents.  It's no longer the case that Alsace-Lorraine is really what we want to be fighting for -- instead it is the human capital where a lot of the value lies and it is hard to loot that.  This is not to say that resources are not a good thing (the resource curse can be over-rated) but that they are no longer the sole thing that drives a states economic power. 

This is likely to be a good development, in the long run. 

Today's Principal Agent Problem

This is Joseph

From Mark Miller:

The fix moving through Congress would revised the Employee Retirement Income Security Act (ERISA) to grant plan trustees broad powers to cut retired workers' benefits if they can show that would prolong the life of the plan. That would mark a major change from current law, which calls for retirees to be paid full benefits unless plan assets are exhausted; then, the PBGC steps in to pay benefits, albeit at a much lower level. The bill also would increase PBGC premiums paid by sponsors, from $13 to $26 per year.

and

The big problem here is that the plan fails to put retirees at the head of the line for protection. When changes of this type must be made, they should be phased in over a long period of time, giving workers time to adjust their plans before retirement. For example, the Social Security benefit cuts enacted in 1983 were phased in over 20 years and didn't start kicking in until 1990.

My problem with this approach is that it does not appear to give the people who manage the plans (i.e. the corporate entities) any incentives to manage the plans successfully.  The focus on short term performance and the linkage of these metrics to managerial compensation, creates a severe principal agent problem -- the group responsible for managing the plan doesn't bear any of the cost if the plas end up dramatically underfunded.


It is often the case that such schemes do not end well, as it can be very easy to boost profits by underfunding the pension scheme.  Under these rules, the potential long term cost to the business is quite mitigated.  Do we really want incentives aligned like this? 

Movies in the modern age

This is Joseph, trying something new.

I was very interested to read this piece on the focus on sequels:

I believe that what studios see when they look at the bumper-to-bumper barricade of a 2015–20 lineup they’ve built is a sense of security — a feeling that they have gotten their ducks in a row. But these lists, with their tremulous certainty that there is safety in numbers, especially when numbers come at the end of a title, represent something else as well: rigidity and fear. If you asked a bunch of executives without a creative bone in their bodies to craft a movie lineup for which the primary goal is to prevent failure, this is exactly what the defensive result would look like. It’s a bulwark that has been constructed using only those tools with which they feel comfortable — spreadsheets, P&L statements, demographic studies, risk-avoidance principles, and a calendar. There is no evident love of movies in this lineup, or even just joy in creative risk. Only a dread of losing.

I must admit that is interesting that the Lord of the Rings and the Hobbit became a six movie extravaganza and that the Marvel/DC comic book movie line-up is . . . daunting.  Sure, some of these movies are surprisingly good (see the latest X-men movie).  But the focus on doing more of the same is exactly what led to a long period where TV was pretty much a dead art form.  Now roles are reversed, and it might well be that movies have a long slump before something new and daring pops up.

Happy Holidays

Since 2009, Mark and I have used this as a forum for a lot of different topics.  From a blog that started about Epidemiology and a little bit of Dungeons and Dragons, we've become instead an eclectic opinion blog, talking about a wide range of fun topics.  Education reform was probably the first deviation, but since then we've seen media criticism, financial advice, and a general willingness to be skeptical about strong claims.

It's been a great ride and we look forward to continuing it in 2015. 

Incentives and proper incentive alignment: a never-ending series

This is Joseph.

Mark has promised a post on this topic, so I thought that I'd start the show by revisiting the issue with Amazon and warehouse workers.  From Megan McArdle:

Should you get paid for standing in line? Workers at an Amazon warehouse thought they should. I kind of agreed. But the Supreme Court disagreed, holding 9-0 that the Amazon contractors could be forced to stand in line to clear security at the end of their workday but did not have to be paid for their time.

The issue here is that the wait could be lengthy.  That added unpredictability to the end of the shift and, if time in line was unpaid, the employer has no incentive to improve matters.  Screening costs and the more efficient it is then the more costly it would be.  Increasing waits to exit protected Amazon profits by reducing shrinkage (which is good) but cost workers (standing in line).

It is legal outcomes like this that check my libertarian streak,  If the law isn't constructed to handle these sorts of external costs on workers (and everyone else) then it makes the preconditions for basing the social contract on contract law to be questionable -- even if this ruling should happen to be correct on the merits.

Sony's hypercube -- they keep finding new sides to cave in on

First they caved in when Rogen and Franco came in with a bad idea for a movie that was certain to be more trouble than it was worth.

Then they caved in due to threats and cyber-attacks from North Korea.

And finally(?) today they caved in to criticism and bad PR and announced a small and largely symbolic theatrical release.

Is this really the end? I can't think of anyone else to whom they could cave, but we do have about thirty-six hours till the theaters open on Christmas Day, I have faith in Sony executives' ability to surprise me.

The Hobbit: The Battle of the Five Armies

This is an off-topic post by Joseph

I just wanted to point out this review.  I think Howard Taylor hits most of my high points and he is quite astute about the book versus movie piece.  The movie is better if you just pretend it isn't adapting a book and take it on its own merits. 

Accountability is for little people continued -- more thoughts on the Sony hack

I'm not quite sure what to call it, but most business reporting suffers from the journalistic equivalent of regulatory capture. There are exceptions of course – I can think of lots of sharp, independent writers covering this beat – but the vast majority of what you read in the business section reflects the viewpoint and very often the spin of the companies being covered.

This "journalistic capture" becomes particularly apparent when companies have massive screw ups. You get lots of stories about how various disasters were unforeseeable and/or unavoidable. The part about the highly paid C-level executives being grossly incompetent has a way of being buried.

The last post covered an almost comic level of incompetence in Sony's IT department. That is probably the bigger part of the hacking story, but there were other questionable decisions that led up to this fiasco, starting with the decision to greenlight the Interview in the first place.

Mark Evanier, who has been around the industry for decades, has a very good post on the subject. His treatment of the freedom of speech issues is extremely sharp, but it is his discussion of the movie itself that is relevant to this post.

Thursday night at the screening I attended, there was what we call an Industry Crowd, meaning the entertainment industry. I heard much talk about the whole matter and I kept hearing — this is the rumor mill speaking now — that everyone at Sony thought the film was awful and that they were just hoping to get it into theaters and make some bucks before reviews and word of mouth killed it. It's common knowledge the film's release was delayed from last August because Sony demanded changes.

I'm not suggesting that good films deserve to be defended and bad ones don't. But before the hacking and threats, Sony had the right to decide the film was a lox that wasn't worth releasing. Some execs at Sony felt that way; that the film shouldn't be released…or maybe wasn't worth the problems it might cause. (No one in the film business is dense enough to think a movie about assassinating a foreign leader couldn't possibly get anyone upset.) And they had the right to make that decision. I'm suggesting they still have that right.

Evanier is almost certainly correct here, but, from a business standpoint, was making this film a good idea in the first place?

Let's be clear, this decision was never about art or making a statement. We're not talking about Dr. Strangelove; at best we're talking about a Hope and Crosby "Road to" picture with gross-out gags . The only considerations were financial and the only political element was the studio politics involved in telling a couple of big, spoiled stars that they couldn't make a vanity project. From Sony's point of view, the Interview was probably a bad idea for a movie and was likely to create all sorts of problems (keep in mind, this is a Japanese company which makes concern about North Korea a bit more immediate). It appears that the main argument for making the movie was that it kept the stars happy and the studio didn't have to make Green Hornet II.*

As I said before, none of this in any way diminishes the severity of the criminal acts involved, but there's blame enough to cover both malicious and the negligent. In theory, we shouldn't have to worry about the latter because the market for top level executives is supposed to be efficient -- we are told that companies get what they pay for when they pay the big bucks.

With that in mind, take a look at this graph from Fusion.

 (The metric used is “market posture,” which measures each level of employee pay compared to the market median for that level.) According to the chart, SPE pays its level 10 employees 113.1% of the median, but only pays its level 1 employees 92.2% of the median.

Assuming level 10 is the top of the scale, it is difficult to see how that above average pay has translated into above average executive performance.


* The Green Hornet probably did turn a profit (between the massive marketing budget and the peculiarities of Hollywood bookkeeping, it's difficult to say for certain), but the box office was not great and the reaction to the film effectively killed the anticipated franchise.

I'm more comfortable blaming the victim a little when the victim has a market cap of twenty billion

While in no way taking away from the magnitude of the criminal acts involved in the Sony hacks, it is important to remember that upper-level management gets such high salaries in part because they are supposed to anticipate threats and take steps to minimize their potential impact.

At Sony, not so much...

The new trove appears to include a collection of documents the hackers came across on the Sony Pictures network that had “password” in their titles, and includes digital keys for everything from Sony computers and servers to magazine subscriptions and YouTube accounts for Sony movies. (As much as we’d like to log into This is the End’s YouTube page, we haven’t actually tried any of these passwords to see if they work.) It is generally a bad idea to store all your passwords in a document on your computer. It is an even worse idea to title that document something like “My Passwords.”
The hackers leaked a new file that includes a collection of all the documents Sony Pictures employees used to store passwords

Sony Pictures employees and former employees are flipping out about the leak and the unexpected debut of their personal information on screens across the world. But some former employees, who asked to remain anonymous, have told us that they’re disappointed but not surprised by the massive hack given Sony Pictures’ long-running lax attitude toward security. They say that employees highlighted specific vulnerabilities on company websites and systems that were never addressed.

“Sony’s ‘information security’ team is a complete joke,” one former employee tells us. “We’d report security violations to them and our repeated reports were ignored. For example, one of our Central European website managers hired a company to run a contest, put it up on the TV network’s website and was collecting personally identifying information without encrypting it. A hack of our file server about a year ago turned out to be another employee in Europe who left himself logged into the network (and our file server) in a cafe.”

Part of that joke was an org-chart straight out of a Dilbert cartoon.

The information security team is a relatively tiny one. On a company roster in the leaked files that lists nearly 7,000 employees at Sony Pictures Entertainment, there are just 11 people assigned to a top-heavy information security team. Three information security analysts are overseen by three managers, three directors, one executive director and one senior-vice president.

Keep in mind, this is more than three years after Sony suffered "one of the largest data security breaches in history."

Just to be clear, the great majority of the upper-level executives I've encountered (no C-level, but quite a few directors and VPs) have been smart, hard-working and conscientious. I certainly don't want to make a blanket condemnation, but stupid, incompetent people do sometimes make it through, and if they get to a high enough rung, it is amazing how small the consequences are for their screw-ups. Accountability is for little people.

On a completely unrelated topic.

In 2005, Sony Pictures Entertainment was audited to ensure the company was keeping in line with federal regulation regarding information security practices. The auditor found, among other things, that Sony had deliberately engaged in insufficient digital security practices, including allowing employees to use basic proper nouns as passwords instead of requiring them to use a complex system involving random letters, numbers and punctuation marks.

If Sony were a bank, the auditor said, its lackluster security practices would put it out of business.

Sony’s then-executive director of security information Jason Spaltro pushed back: If a bank was a Hollywood film studio, he said, it would already be out of business.

“It’s a valid business decision to accept the risk (of a cyberattack),” Spaltro told CIO Magazine in 2007. “I will not invest $10 million to avoid a possible $1 million loss.”

As mentioned earlier, a few years after that interview hackers would steal  personally identifiable information from 77 million Sony PlayStation accounts. What happened to the executive of security information who gave that embarrassing interview?

By the way, Jason Spaltro — the executive from the beginning of this article who suggested the company not spend $10 million to combat a potential $1 million risk — still works at Sony. He has since been promoted to vice president of information security — one of the top executives tasked with ensuring things like the Sony Pictures hack don’t happen. He makes close to $700,000 a year: $300,000 base salary and a $400,000 initiative-based bonus. We know this because hackers published his employment information last week.

Checking in on the last next big thing

I'm working on a longer piece that involves Uber and I got to thinking about the big discussion we had about Groupon back in 2011. Many of the arguments currently being used to justify Uber's $40 billion valuation (rapid growth, huge potential market, a foundation of new economic and technological paradigms) were also being used to justify faith in Groupon.

I don't want to make too much of the analogy -- they are very different companies with very different business models -- but it is still useful to stop and think about how that bet worked out.

[If you want to get a head start on the Uber thread, check out these exceptionally thorough analyses from Talking Points Memo and NYU Finance Professor Aswath Damodaran.]

A quick note on Kružno, official game of the village of Kružno*

I've got a couple more post I'd like to do on Kružno, the abstract strategy game I developed a few years ago, a post on abstract strategy games and another on trying my hand at small scale manufacturing (and why I ended up using chess pieces in a checkers variant). Unfortunately, I'm a bit pressed for time and I really wanted to get something out today, so this will have to do for now.

You can find the rules here and the game itself here. The game takes about three to five minutes to explain, perhaps a bit longer working from the instructions (some parts are a bit unclear and need to be rewritten).

If you are, or someone you know is, a teacher (particularly at a school where money is tight) who would like to make more use of games in the classroom, let me know and I'll try to set you up with some game sets or at least some extra boards.


* That's also a subject for another post

Perhaps claims that are almost certainly false should be fact-checked more thoroughly

If you read Talking Points Memo or the Washington Post, you've probably heard about this story from New York Magazine by up and coming journalist Jessica Pressler:

Late last year, a rumor began circulating at Stuyvesant that a junior named Mohammed Islam had made a fortune in the stock market. Not a small fortune, either. Seventy-two million.

An unbelievable amount of money for anyone, not least a high-school student, but as far as rumors go, this one seemed legit. Everyone at Stuy knew that Mohammed, the soft-spoken son of Bengali immigrants from Queens and the president of the school’s Investment Club, was basically a genius.

As the news spread, Mo’s stock went up. The school paper profiled him, Business Insider included him on a list of “20 Under 20,” and Mo became “a celebrity,” as his friend Damir Tulemaganbetov put it on a recent Friday night at Mari Vanna near Union Square. “A VIP!”
...
Like [Jordan] Belfort, Mo started with penny stocks. A cousin showed him how to trade. He loved the feeling of risk—the way his hand shook making the trade—but he swore it off after losing a chunk of the money he’d made tutoring. “I didn’t have the balls for it,” he said. He was 9.

It was a while before he was ready to try again. In the meantime, he became a scholar of modern finance, studying up on hedge-fund managers. He was particularly enamored of Paul Tudor Jones. “I had been paralyzed by my loss,” Mo said. “But he was able to go back to it, even after losing thousands of dollars over and over. Paul Tudor Jones says, ‘You learn more from your losses than from your gains.’ ” Mo got into trading oil and gold, and his bank account grew. Though he is shy about the $72 million number, he confirmed his net worth is in the “high eight figures.” More than enough to rent an apartment in Manhattan—though his parents won’t let him live in it until he turns 18—and acquire a BMW, which he can’t drive because he doesn’t yet have a license. Thus, it falls to his father to drive him past Tudor Jones’s Greenwich house for inspiration. “It’s because he is who he is that made me who I am today,” Mo said.

[I emphasized the part about “high eight figures.” It becomes important later.]

Islam fills in some more details on his website:

My interest in finance started at the tender age of 8, but I really started trading when I was 10 years old. I paper traded for a year or so and finally opened a real trading account using my own capital that I saved from entrepreneurial endeavors. 

So we're looking at six or seven years to go from tutoring money to more than fifty million dollars trading part-time. That's a difficult statement to believe but Pressler and her editor at NY Magazine managed to swallow it, as did the New York Post which ran "High school student scores $72M playing the stock market" and CNBC which had Islam and friends scheduled to appear until...

From the New York Observer:

Monday’s edition of New York magazine includes an irresistible story about a Stuyvesant High senior named Mohammed Islam who had made a fortune investing in the stock market. Reporter Jessica Pressler wrote regarding the precise number, “Though he is shy about the $72 million number, he confirmed his net worth is in the ‘high eight figures.’ ” The New York Post followed up with a story of its own, with the fat figure playing a key role in the headline: “High school student scores $72M playing the stock market.”

And now it turns out, the real number is … zero.

In an exclusive interview with Mr. Islam and his friend Damir Tulemaganbetov, who also featured heavily in the New York story, the baby-faced boys who dress in suits with tie clips came clean. Swept up in a tide of media adulation, they made the whole thing up.

The Washington Post picks it up from there:

Then on Monday afternoon, Mo backed out of a spot on CNBC, confessing the $72 million figure was inaccurate. “The attention is not what we expected — we never wanted this hype,” he said. New York, however, stood by the story: “Our story portrays the $72 million figure as a rumor … we did not know the exact figure he has made in trades. However, Mohammed provided bank statements that showed he is worth eight figures, and he confirmed on the record that he’s worth eight figures.”

Just to be clear, the NY editors initially pinned much of their defense on the distinction between $72 million and “high eight figures,” and on the easily faked bank statement provided to the fact checker. Apparently the fact-checker didn't ask to see additional documentation or to speak to the parents (who are, according to the Observer, really pissed).

New York Magazine finally came out with a retraction but even here they don't seem to have learned their lesson:

After the story's publication, people questioned the $72 million figure in the headline, which was written by editors based on the rumored figure. The headline was amended. But in an interview with the New York Observer last night, Islam now says his entire story was made up. A source close to the Islam family told the Washington Post that the statements were falsified. We were duped. Our fact-checking process was obviously inadequate; we take full responsibility and we should have known better. New York apologizes to our readers.

The process would have been inadequate if the claim had been credible; for a story this incredible, it was grossly negligent.

This is not an isolated case. Standards for journalistic accuracy have been dropping for at least a couple of decades, but perhaps more troubling is the disconnect between fact-checking and credibility. Even at the venerable BBC, the most unbelievable of claims is not subjected to any extra scrutiny. Fact-checking has become a CYA process, not something you do because you want to get the story right.

Infrastructure part 2

This is Joseph

In the same vein as some of my recent comments on infrastructure, the latest example of tricky projects is happening on the west coast.  Consider:

Transit advocates are often accused, absurdly, of engaging in a “war on cars”. If we were indeed committed to such a war, I’m not sure we could have come up better with anything than this. The overruns will likely cannibalize WSDOT’s budget, including all manner of road repair and construction projects (some of which are necessary and useful) for the foreseeable future. If, as appears increasingly likely, the viaduct must be shut down before the tunnel is ready, transit will become even more crucial for accessing downtown, and far fewer cars will be able to do so with any efficiency at peak travel times.  Meanwhile, Sound Transit’s tunneling project for light rail, using well established, off the shelf tunneling technology and conservative cost estimates, chugs along ahead of schedule and under budget, and Seattle just voted itself a tax increase to fund more bus service.

I don't like the terms luddite and ddulite (I prefer technophile and technophobe), but it is a classic example of a series of expensive decisions caused by trying to save money through a clever technological solution.  I return to my thoughts in the previous post, that the real point of concern is our inability to use off the shelf technology effectively in terms of infrastructure.

From an economics point of view, we really have a case of misaligned incentives.  How do we make more projects look like Sound Transit and less like the projects that have been encountered in the viaduct replacement?