Monday, July 26, 2010

Work and the Internet

For context about why I care, here is a slight paraphrasing from grant proposal I recently wrote:
For the last eight years, I have been working to develop an area of computer science called "human computation," which studies how to harness the combined power of humans and computers to solve problems that would be impossible for either to solve alone. This growing academic field now has an annual workshop, a community with researchers from the top computer science programs in the world, and has directly influenced the popular online trend of crowdsourcing, in which crowds of people are enticed to perform work over the Internet. Subsequent to the development of this area, for example, Amazon created Mechanical Turk, a marketplace for human computation tasks (or “human intelligence tasks” as they call them), which is now used and studied by hundreds of researchers worldwide. Since then, other similar services have emerged where workers are paid to perform micro-tasks that are hard for computers.

An example of human computation is reCAPTCHA, in which people help digitize books by typing CAPTCHAs on the Internet. To date, over 750 million unique people—more than 10% of humanity—have helped transcribe at least one word through reCAPTCHA.

All human computation systems must have a way to motivate the users to participate. In the case of reCAPTCHA, the value proposition is as follows: by typing a CAPTCHA, the user gets access to a desired resource like a free email account or tickets to a concert, and in exchange they perform ten seconds of work that is utilized to help transcribe a book. In the case of Mechanical Turk, users are paid a few cents to perform each task.

A discussion that I've had with multiple people over the last few years is whether systems like Mechanical Turk, in which real money is exchanged, should be legislated so that workers are fairly compensated. You see, the average hourly rate of most workers in such sites is usually well below the minimum wage of most third world countries. As a concrete example, the minimum wage in Guatemala is approximately $1/hour, whereas it's not rare to see tasks on Mechanical Turk in which the effective hourly rate is $0.30/hour. (It's amazing that many of the workers on Mechanical Turk come from inside the United States.) Some labor economists would tell you that this is ok: if people are willing to work for such low rates, who is to stop them? However, most countries have some notion of a minimum wage in their laws, including the United States, so in essence as a country we do not believe in an unregulated labor market.

Recently I have heard more than one company saying something like: "We use Mechanical Turk because otherwise we would have to pay people $7/hour to do this task." In other words: "We use Mechanical Turk to get around the minimum wage laws." As wrong as it may sound to some, this is currently ok. In the United States, "independent contractors" are typically not covered by minimum wage laws, so while I'm not a lawyer I believe using Mechanical Turk to get around minimum wage is as legal as hiring independent contractors instead of full-time employees.

But the question remains: Should sites like Mechanical Turk be regulated? Perhaps not today, but if the Internet or crowdsourcing really is the future of work, we should at least be thinking about it.

Here are some issues that make this complicated:

  • Labor markets like Mechanical Turk are truly global, with workers coming from many different countries. Can the same minimum wage be applied to all?

  • Most countries have immigration work laws that prevent people without the proper visa to work inside that country. Should these still apply when the work is performed over the Internet? In many cases it's not even possible to tell where the worker is located, so are these laws even enforceable?

  • Assume we decide as a country that labor markets like Mechanical Turk should be legislated and a minimum wage is imposed. Some of the work on human computation involves transforming tasks into enjoyable games so that people perform them in exchange for entertainment. Is it ok to pay people less (or nothing) if the task is fun?

  • What about writing a review for a book online or rating a video? These are concrete pieces of work that benefit the Web sites, but that nobody seems to object to doing for free.
  • Saturday, June 19, 2010

    Research versus Teaching

    My previous post stirred some people's emotions. Reading the comments, it seems part of that came from the tension between teaching and research in modern American universities.

    In most countries, the role of universities is solely to educate their students. That's true of many colleges in the United States, but not of Research I universities. The majority of American universities you've heard of belong to this category: Harvard, Yale, UCLA, Stanford, MIT, CMU, Princeton, etc. In addition to teaching, these institutions have another equally important role, which is to produce high-quality research that benefits society. Indeed, many of the game-changing discoveries or inventions in the last century have been entirely or partly developed at American Research I universities: The Internet, Google, the cure for polio, vitamin D milk, even Gatorade. To a large extent, this is where Nobel prizes are won, and where the future is invented.

    American Research I universities are also mostly responsible for educating the smartest people in the country (or even the world), both at the graduate and undergraduate levels. This includes most doctors, lawyers, politicians, US presidents, dot com billionaires, and yes, even Lady GaGa.

    Combining these two very important roles may have benefits, but it also causes an unspoken tension. Is the job of a professor primarily to educate or to do research?

    The interesting thing is that everybody seems to have a different opinion about this. Students and their paying parents, of course, think professors are there solely to educate; Professors mostly think they are there to do research; and university administrators seem to change their tune depending on whom they’re talking to.

    As usual with me, I have more questions than answers. Should research and education be combined in this manner? Should professors primarily concentrate on research or teaching?

    Regardless of what should happen, I can tell you that at least from a tenure-track professor’s point of view, the system at the vast majority of Research I universities is extremely biased towards the research side. Most of my friends at other universities chose to be professors because they want to do research without being pressed by economic outcomes like they would in a company, and consider teaching a bearable chore that they must do to get the freedom and prestige of being a professor. The hiring of faculty (at least at the ~15 Research I universities that have offered me a job) pays almost no attention to the potential quality of the candidates as teachers. The tenure process also puts teaching in the back seat. So in essence, professors are largely not selected, evaluated, or rewarded based on teaching.

    This is not to say that there are no good teachers among the faculty at Research I universities. Many of the faculty both here at CMU and elsewhere are outstanding instructors and work very hard on their teaching. However, they do so out of pure love (and possibly a misconceived sense of duty), because the system is not set up for this.

    Since I don’t want to get in trouble again with the commenters, I will end with a few disclaimers. First, I do spend a significant amount of time on my teaching (as evidenced by having won the teaching award). Second, there are very good institutions that educate smart people in the US that are not Research I universities and that concentrate solely on teaching.

    Wednesday, June 16, 2010

    Outsourcing My Research Group

    A PhD student at Carnegie Mellon costs approximately $80,000 per year. (Research programmers and post-docs cost about the same.) Given that PhD students have to take classes for the first couple of years and are therefore running at 50% capacity, this means that each effective person in my research group costs on average $100,000 per year.

    I'm from Guatemala. For $100,000, you can hire 4-5 extremely competent full-time engineers there (even accounting for the 50% overhead rate inside CMU). My question today is: would it make sense to take 5 engineers instead of a PhD student next time I have extra money?

    I understand that CMU PhD students have a much higher IQ than the average programmer, and that for certain tasks you can't just rely on programmers, but if the exchange rate is 5 to 1, I think the experiment is worth a try.

    And from there, it's a slippery slope: why not just move my whole research group to India or China, since a large fraction of our PhD students come from there anyways?

    Part of the goal of being a professor is mentoring, and I love that part: I am not saying we should get rid of PhD students, but that perhaps a mix of some outsourced coding and PhD students would be a better investment for everybody.

    Disclaimer: 100% of my PhD students are working on projects of their own choosing, and if anything my biggest flaw as an advisor is not giving them enough direction (instead of micromanaging them).

    Wednesday, June 9, 2010

    America's Next Top Nerd

    I want to make a computer science reality TV show. Like in most reality shows, the participants would compete to win a prize, but since they would be CS nerds, we won't aim for something as crazy as getting married to a famous person (maybe it can be something like becoming Facebook friends with a girl).

    Anyways, instead of athletic or beauty competitions, I want the participants to compete by solving computer science problems. My question is this: what are some good CS problems for TV? I have some thoughts, but I'd also like to hear what others have to say. Ideally the problems would: (1) explain a cool CS concept, (2) be accessible to a PBS-type audience (i.e., no PCP proofs), and (3) have something that can be filmed.

    Monday, June 7, 2010

    Startups and CMU

    Is Carnegie Mellon a good university to go to if you plan on working for or building a startup? That's the question I was recently asked by somebody from the Quora team. I'm frequently asked similar questions because I started a company out of CMU that never moved out of Pittsburgh even after it was acquired by Google. Following another brief conversation about this on Twitter today, I decided it would be good to have an open discussion about startups and CMU. I'll start with my personal opinion, but hope that others pitch in.

    Personally, I think CMU is a great place to start a company. Granted, I'm biased because things worked out for me, but here are some objective reasons why I think more people should be starting technology companies out of CMU:

    1. Talent Pool. CMU graduate and undergraduate students are truly world class. The computer science PhD program is ranked #1 by USNews, and according to my recruiter friends, CMU is the #1 or #2 school by quantity of hires from companies with uber selective hiring standards such as Facebook or Google.

    2. Less Competition for this Talent. I won't claim that CMU students are strictly better than students from e.g. Stanford, but I will claim this: the competition for hiring a top student to join a startup at CMU is much much lower than at comparable schools in California or Boston, since it's not the case that everybody and their mother has a startup in Pittsburgh (last time I went to Silicon Valley, even the guy that served me at Starbucks, who overheard my conversation with a VC friend, started pitching me his company!). At CMU you'll have your pick of top talent to start a company with.

    3. Opportunity for Different Ideas. Don't get me wrong, I love Silicon Valley. But it always strikes me when I go there how much everybody thinks exactly the same as each other (most even dress the same as each other). While some herd mentality is ok, I think there is huge benefit to being outside of that bubble.

    If I were to start another company, I would do it out of CMU.

    Monday, May 24, 2010

    Give Me Back My Name

    Dear Person Who Took My Name on Twitter and Linked to a Porn Site,

    Yes, I know, I was stupid. I should have taken @LuisvonAhn years ago when I first signed up. But you have to understand, I didn't know Twitter would become such a big thing. I was just checking out reCAPTCHA on their registration page, and typed the first username that came to mind: @freakbit. (No comments, please.) Now, Twitter is big, I wanna start using it, and I am stuck with @freakbit. Would you consider giving the name back?

    Sincerely,
    The real Luis von Ahn

    PS: Are you also the person who has the fake Luis von Ahn on Facebook? If so, please give that back as well! I won't accept your friend request.

    Update: People from both Twitter and Facebook were super helpful, so now I have @LuisvonAhn as my Twitter name, and there is no more FB impersonator :)

    Thursday, March 25, 2010

    Bored: Need Some Scandalous Gossip

    Many people have emailed asking why I haven't posted on this blog for a while. Well, here's why: I've been busy working on a bunch of things including a super duper secret project. I have a big mouth, so I best not say anything else.

    I'm now taking a little break and I need some entertainment. I tried the TV but there was nothing good. Soooo... how about we try this: readers, use the handy anonymous comment box below to tell the world and me some gossip. Now, this gossip should be maximally entertaining to me, so here are the rules: (1) It should be juicy. (2) It should be related to computer science or technology in general. (3) Extra points if it has to do with CMU or any university. Let's see what you got.