Online Education and Self-Driving Cars

Last week, I noted that former Stanford professor Sebastian Thrun enrolled 160,000 students in an online computer science class. That inspired him to set up a new company, Udacity, to pursue online education. A new article in Bloomberg BusinessWeek adds some additional color to the story.

Barrett Sheridan and Brendan Greeley answer a question many folks asked about the students: how many actually finished? Answer: 23,000 finished all the assignments.

Second, they note that professor Thrun is also at the forefront of another potentially transformative technology: self-driving cars:

Last fall, Stanford took the idea further and conducted two CS courses entirely online. These included not just instructional videos but also opportunities to ask questions of the professors, get homework graded, and take midterms—all for free and available to the public.

Sebastian Thrun, a computer science professor and a Google fellow overseeing the search company’s project to build driverless cars, co-taught one of the courses, on artificial intelligence. It wasn’t meant for everyone; students were expected to get up to speed with topics like probability theory and linear algebra. Thrun’s co-teacher, Peter Norvig, estimated that 1,000 people would sign up. “I’m known as a crazy optimist, so I said 10,000 students,” says Thrun. “We had 160,000 sign up, and then we got frightened and closed enrollment. It would have been 250,000 if we had kept it open.” Many dropped out, but 23,000 students finished all 11 weeks’ worth of assignments. Stanford is continuing the project with an expanded list of classes this year. Thrun, however, has given up his tenured position to focus on his work at Google and to build Udacity, a startup that, like Codecademy, will offer free computer science courses on the Web.

I wish Thrun success in both endeavors. Perhaps one day soon, commuters will settle in for an hour of online learning while their car drives them to work.

P.S. In case you missed it, Tom Vanderbilt has a fun article on self-driving cars in the latest Wired.

Can One Professor Teach 500,000 Students At Once?

That’s what former Stanford professor Sebastian Thrun aims to do.

Sound impossible? Well, he’s already taught a class of 160,000 students. As Felix Salmon recounts:

Thrun told the story of his Introduction to Artificial Intelligence class, which ran from October to December last year. It started as a way of putting his Stanford course online — he was going to teach the whole thing, for free, to anybody in the world who wanted it. With quizzes and grades and a final certificate, in parallel with the in-person course he was giving his Stanford undergrad students. He sent out one email to announce the class, and from that one email there was ultimately an enrollment of 160,000 students. Thrun scrambled to put together a website which could scale and support that enrollment, and succeeded spectacularly well.

Just a couple of datapoints from Thrun’s talk: there were more students in his course from Lithuania alone than there are students at Stanford altogether. There were students in Afghanistan, exfiltrating war zones to grab an hour of connectivity to finish the homework assignments. There were single mothers keeping the faith and staying with the course even as their families were being hit by tragedy. And when it finished, thousands of students around the world were educated and inspired. Some 248 of them, in total, got a perfect score: they never got a single question wrong, over the entire course of the class. All 248 took the course online; not one was enrolled at Stanford.

And I loved as well his story of the physical class at Stanford, which dwindled from 200 students to 30 students because the online course was more intimate and better at teaching than the real-world course on which it was based.

 Inspired by that experience, Thrun has now founded Udacity, a private online university. As Nick DeSantis of the Chronicle of Higher Education reports:

One of Udacity’s first offerings will be a seven-week course called “Building a Search Engine.” It will be taught by David Evans, an associate professor of computer science at the University of Virginia and a Udacity partner. Mr. Thrun said it is designed to teach students with no prior programming experience how to build a search engine like Google. He hopes 500,000 students will enroll.

Teaching the course at Stanford, Mr. Thrun said, showed him the potential of digital education, which turned out to be a drug that he could not ignore.

“I feel like there’s a red pill and a blue pill,” he said. “And you can take the blue pill and go back to your classroom and lecture your 20 students. But I’ve taken the red pill, and I’ve seen Wonderland.”

That Wonderland will be a serious challenge to traditional chalk-and-talk universities — and a wonderful opportunity to democratize knowledge around the globe.

(ht: Alex Tabarrok at Marginal Revolution)

How Will Colleges Innovate?

That’s the question that Jeffrey Selingo poses over the The Chronicle of Higher Education (ht: Jack B.):

[I]f current economic trends continue, much of traditional academe is going to be forced to change. Families can no longer use their house as an ATM. States are making tough choices about the size of government, and public colleges are often left at the end of the line. And now the federal government is likely to cut back on many of its fiscal promises to deal with an out-of-control deficit.

The bottom line is that we’re likely to face a future where students and their families pay a lot more of the cost of a college education out of pocket. Without grants and loans as a safety net, students are probably going to make different choices than they do now (read: less expensive choices). We’re likely headed toward a future where smaller, struggling colleges need to move to new models of doing business, while elite, wealthy colleges continue to support the current model.

Selingo then summarizes several ideas that were bandied about in a meeting of academic “disruptors” and disruptive innovation guru Clayton Christensen of Harvard Business School. They include:

  • Disaggregated universities, in which colleges would purchase courses from other colleges (or, I suppose, sources like Khan Academy) rather than produce them themselves, and
  • Modular universities, in which colleges would provide much more focused degree offerings.

Also on the agenda: rethinking the often-anachronistic academic year and the scholastic currency known as the credit hour.

To be sure, none of these ideas are particularly earth-shattering. But that may well be the larger point. America’s higher education “industry” might well reap substantial benefits from adopting organizational ideas that are already old hat elsewhere in the economy.

It’s Back-to-School Season, Time to Lay Your Bets

According to an article over at the Huffington Post (ht Natalie), students at 36 colleges will have a new option when they start classes this fall. Thanks to an outfit named Ultrinsic, students can now bet on whether they will get good grades. Students put up money at the start of the semester and then get payoffs at the end depending on how they do.

Calling it a bet isn’t completely fair, however, since the payoff creates an extra incentive for students to do well. So think of it as a combination of betting (if you think your odds of doing well are better than Ultrinsic thinks) and using a financial incentive to get your future self to study a bit harder. Naturally, Ultrinsic emphasizes the incentive perspective in describing its “Reward” product:

Do you like getting good grades? The right amount of cash should provide you with the needed motivation to pull all-nighters and stay awake during the lectures of your most boring professors. At Ultrinsic.com, you will be able to earn cash while working to achieve your academic goals.

Obligatory note to my new crop of students: all-nighters are generally not an optimal learning strategy.

Like a race track, the company offers packages that pay off not only if you do well on a single course, but also if you perform well in multiple courses or over an entire semester. If a new freshman is really feeling motivated, he or she can also put down $20 up front for the opportunity to win (earn?) $2,000 for maintaining a 4.0 GPA throughout college.

And if students are feeling risk-averse, they can also buy insurance against bad grades. Bomb that final and get a cash reward.

Somehow I doubt many students will want to buy such insurance. Or that Ultrinsic will want to sell it given the risks of moral hazard. Perhaps Ultrinsic will screen for “pre-existing conditions” (like failing a related class) in order to limit the adverse selection. Or just offer such high premiums that only a few extremely risk-averse (or mathematically-challenged) students will apply.

The incentives product, however, seems much more promising. Indeed, it resembles some other efforts to help people modify their own behavior through financial incentives. See, for example, the folks at stikK.com whose service allows users to create their own incentives. For example, you could commit to give $500 to your favorite charity if you fail to lose 10 pounds by Christmas. Even better, you could commit to give $500 to your least-favorite charity if you fail to drop the pounds.

Ultrinsic is just applying this logic to college grades … and kindly offering to take a cut when students fall short.

The Weird Economics of Cellular Calling Plans

Yesterday’s New York Times has an amusing article about the complexities of cell phone pricing (ht: Carolyn):

HERE’S a consolation prize to the millions who recoil in bafflement from cellphone companies’ labyrinthine price plans, with their ever more intricate arrays of minutes, messages and megabytes: Economists don’t understand them, either.

“The whole pricing thing is weird,” said Barry Nalebuff, an economics professor at the Yale School of Management. “You pay $60 to make your first phone call. Your next 1,000 minutes are free. Then the minute after that costs 35 cents.”

I frequently use cell phone pricing to illustrate key ideas in my microeconomics class. For example, why is it often impossible to get a signal for my beloved iPhone? Because the marginal cost of downloading data is zero, and AT&T’s bandwidth is often overwhelmed. And why do cell phones usually charge a monthly fee rather than just a per-call or per-minute fee? Because they can get more revenue by offering a low (or zero) per-minute price coupled to a high monthly fee.

Still, as the article suggests, the complexity of cell phone pricing can sometimes seems inexplicable. Which reminds me of one of my favorite TV ads, starring Catherine Zeta Jones and a group of economists who are determined to explain cell phone pricing to consumers:

For my interpretation of the ad, see this post.

Opium Economics in Afghanistan

If you are troubled by opium production in Afghanistan, Jeff Clemens at Harvard has some bad news for you: eradication efforts are doing little to reduce opiate production. (ht: Tyler Cowen at Marginal Revolution). Moreover, to the extent they are having an effect, it’s to drive up prices and thus enrich the farmers who illicitly grow poppies.

I mention this not only because I find it interesting, but also because it nicely illustrates one of the ideas that I teach my microeconomics students. When you think about policy interventions – in this case poppy eradication efforts – it’s important to understand both the qualitative impacts of the intervention and the magnitude of those impacts.

Your basic supply and demand model will tell you, for example, that eradication efforts will shift the supply curve left (up), resulting in higher prices and lower production. To gauge the relative importance of those two changes, you need to know something about demand. And in this case, the key fact is that demand (from other countries, not Afghan consumers) responds very little to price. In the lingo, opium demand is very price-inelastic (Jeff estimates the elasticity at about -0.09). As a result, efforts to restrict supply translate primarily into price increases, rather than production declines.

The same problem has bedeviled U.S. efforts to restrict illicit drugs. (For example, see this old New York Times editorial about cocaine, which I use in my class – the editorial that is, not cocaine itself.) I haven’t followed the debate in recent years, but my sense is that many observers concluded that demand-side policies (i.e., discouraging consumption) were often a better strategy than supply-side policies. After all, successful demand-side policies would lower both consumption and price, thus lowering profits from drug production.

Given Jeff’s results, I suspect the same may be true in the world of opium production. If policymakers want to reduce consumption, they may want to turn to demand-side policies (assuming, of course, they can design demand-side policies that would have a substantive effect).

Positive, Normative, and … ?

Am I the only one who feels unfulfilled by the standard distinction between positive and normative economics?

I am gearing up to return to the classroom next week, to teach microeconomics to incoming masters students at the Georgetown Public Policy Institute. Anyone who’s experienced the first day of micro class knows what’s coming. After introducing myself and talking about the wonders of economics (which is, indeed, fun, useful, and enlightening), I will launch into the great positive vs. normative distinction.

In brief:

  • Positive is the science side of economics: understanding and predicting the behavior of individuals, firms, markets, economies, etc. In short, the part of economics in which we try to be physicists (or, sometimes, biologists).
  • Normative is the side of economics where we make value judgments, identifying policies as good or bad. In short, the part of economics in which we try to be philosopher-kings.

Both styles of economics are important, particularly in a public policy program. And drawing a careful distinction is vital, not least because of the many people in Washington (both economists and non-economists) who try to dress up their value judgments as science.

I have one problem with this distinction, however: it overlooks a great deal of what economists actually do.

Continue reading “Positive, Normative, and … ?”