Brian Keng

  • Making Universities Obsolete?

    I recently came across a blog post by Matt Welsh about Making Universities Obselete.  It described how Sebastian Thrun’s Udacity is going to change how higher eduction works.  His points in the article were that there were three key failures in the current model for higher education:

    1. Failure #1: Exclusivity
    2. Failure #2: Grades
    3. Failure #3: Lectures

    Much of the discussion was talking about how Udacity might correct some of these failures and usurp the traditional forms of higher education.  I respectfully disagree — at least not according to the arguments stated.  One key aspect is missing from this model that the current model of higher education does really well: playing to incentives of the stakeholders.  Let me address the three points individually.

    Grades
    The biggest presupposition they make with any of these new models for higher education is that most students want to go to school to learn.  This is a commendable idealism but couldn’t be further from the truth.  I strongly suspect that the main motivation of students is to get a good job (high paying, growth opportunities, good location etc.).  If this is the incentive, then we should consider what is in the best interest of companies looking to hire graduates of these institutions.  Ideally, they want to hire the best candidate with a high degree of confidence.  Obviously this is not a simple task.  This requires evaluating each candidate for using strict hiring guidelines and processes.  This just doesn’t scale well when a single job may have hundreds of applicants.  Practically, employers need some easy measure to filter out candidates.  Yes, they might miss the diamond in the rough, but they save that time in hours and hours of filtering out poor candidates.  Enter grades.  Reducing your entire education to a single number or letter average may seem disheartening to the poor student who has toiled to achieve a respectable degree, but it’s dead easy to filter out people.  Below B+, out.  A+, interview immediately.  Everyone else in between, look at their resumes.  Like it or not, the purpose of grades is exactly to reduce your talent, ingenuity and performance to a single number so that you can be binned.  It’s obviously not accurate, nor precise, not to mention rife with abuse, but it’s the trade-off for how easy it is to use.

    Exclusivity
    Like any fashion retailer knows, your brand is one of the most important assets when you are trying to sell a high end product.  The same goes for a person, a company or a school.  If you brand gets diluted, its value (deserved or not) goes down.  This is precisely why many schools are exclusive.  If MIT suddenly accepted ten times the amount of engineering students, how would the quality of their out-going students fare?  My bet is that the average performance would greatly decrease.  This causes a huge negative effect on the brand that, in the end, does the school, alumni and graduating students a disservice.  It’s like if we produced ten times the amount of diamonds, that 1 karat ring that you got your wife all of a sudden doesn’t mean the same thing (although it’s great for people looking to pick up a diamond).  Brand — and maintaining it — is an important factor that drives the incentives of everyone already affiliated with the university.  This means that for the most part, all these existing parties will want exclusivity over a diluted brand.

    Lectures
    As for the final failure of universities, I think the original article is spot on.  Lectures are a very limited way to convey information, but it’s still not clear if there is a scalable model(s) that works as well.  Giving an iPad to all students definitely is not the solution.  Some universities are already providing video taped lectures for students to view online but this is not the end all or be all of learning.  Most likely different types of material work best when taught in different ways.  I believe that this is an evolution that will (albeit slowly) incorporate newer technologies.  Lectures are mostly just an older model that scales well and does a decent job.  Newer methods are welcome and hopefully we can see some great innovation here.

    Incentives, incentives, incentives…
    Charlie Munger is one of my heroes, in part, because he has some much wisdom without any of the BS.  He’s a huge advocate of incentives and you should be too.  Most people talk about what should happen instead of what will happen.  It’s unfortunate that we don’t live in an ideal world, but that’s the reality.  The only way to make things change for the better is if we make policies, models, and rules that play off the natural human instinct to be swayed by incentives.  Of course, I haven’t talked about any solutions, I’m not sure I have any.  I’m just reminding you that regardless of whatever innovations, ideas, or ideology you have, none of it is going to have any relevance without somehow working incentives into the picture.

  • Incentives and Marks

    Incentives can be powerful things especially for the worse.  Here’s one professor’s story on why he’s not a professor anymore.  A particularly interesting excerpt is on the subject of “scaling” marks:

    “However there are lots of ways round this little problem. One of them is doctoring the marks.  Except its not called ‘doctoring’ its called ‘scaling’ and its done by computer. You scale the marks until you get the nice binomial distribution of fails and firsts. You can turn a fail into a II(ii) with scaling. Probably you want to be generous because otherwise students might not elect to study your course next year and then your course will be shut down and you’ll be teaching Word for Windows. Scaling was universal and nobody except the external auditors (who were lecturers who did the same thing themselves) got to see anything but the scaled marks.”

    One way to look at it is to view the system abstractly as an optimization problem.  If you have 10,000 undergraduate students per year trying to optimize their marks, all of which are funding (at least in part) the pay of the university and the professor in question, which goal is going to be optimized first?  One professor who wishes to maintain the high quality of standard, or the 10,000 (per year) undergraduates all trying to maximize marks while minimizing work?  Gone are the days when I think of university as a place solely for higher learning.  (Though it’s important not to draw broad strokes on the few who still are fighting the good fight.)

    Knowing this, it’s almost crazy to think that one person can change such entrenched institutions as universities.  But perhaps that’s exactly what we need more of — crazy people:

    “Here’s to the crazy ones, the misfits, the rebels, the troublemakers, the round pegs in the square holes… the ones who see things differently — they’re not fond of rules… You can quote them, disagree with them, glorify or vilify them, but the only thing you can’t do is ignore them because they change things… they push the human race forward, and while some may see them as the crazy ones, we see genius, because the ones who are crazy enough to think that they can change the world, are the ones who do.”
    Think Different, narrated by Steve Jobs

    So go on, get your crazy on today.

  • Anna Karenina

    “Happy families are all alike; every unhappy family is unhappy in its own way.”
    from Anna Karenina by Leo Tolstoy

    This quote is  reminder that there is exactly one way to have a happy family — have everything go right.  On the hand, there are millions of ways in which a family can go wrong: attraction, personality, money, religion, and the list goes on.  I think this also applies more generally to what’s termed the Lollapalooza Effect where things conspire together to reach that critical mass where amazing things start happening.  Leave just one ingredient out and the magic doesn’t happen.  The corollary to this is that even if one thing is missing, there’s no rule that says you can’t fill in that gap yourself.  No one has everything going right from the start — they make it happen.

    Happy families may be rare, but they don’t have to be.

  • Learn by Doing

    There are dozens of articles about learn by doing or experiencing or taking action or any number of synonyms.  This is as opposed to learn by reading, listening, planning etc.  The basic gist of this idea is that there are some ideas that can’t be learned by reading a book, they can only be learned by actually doing the activity.  I think most people generally agree but I’d like to offer up some of my experience and a caveat I’ve observed.

    In graduate school, I’m a teaching assistant for a course called Algorithms and Data Structures that offers both a undergraduate version and a graduate one.  The undergraduate version is pretty vanilla with reasonable level questions that some students intuitively get but others struggle with; the usual spread between any reasonably difficult undergraduate coarse.  But the graduate version is much more insidious.  The professor purposely puts questions on the homework (and to some extent on the exam) that are not just hard but tortuous.  These are the kinds of questions that you can literally spend hours thinking about only to end up in the exact same place you started in.  It’s like a maze involving sorting, searching, dynamic programming, greedy algorithms, max flow and any other nasty “basic” algorithms you can think of.

    So it’s no surprise that the students complain.  They wish it could be easier and sometimes a few ask for help on how to make it easier.  I answer them but I’m sad to say it’s never the answer they want to hear.  The answer I give them is that you really have to learn by doing.  That is, you only get good at these algorithm questions after you’ve done a lot of them.  That’s the way you build intuition, by doing — NOT by someone more experienced than you telling you all the answers.  It’s always tough to hear because I know they are stressed out with three or four other courses all with their own deadlines and of course they’re worried about their marks (even though the latter isn’t so important in graduate school).  And alas (I hope) they realize that there is no shortcut, sometimes you have to cross through the bush to get to the other side.

    And here’s where my caveat comes in: the one thing that seems to be missed when touting the “learn-by-doing” strategy is that it can be really hard.  Either because it’s really scary (e.g. public speaking), or maybe just a lot of work (e.g. algorithms).  The best teachers I know intuitively understand this and make you take that key step in understanding by yourself.  It may be hard but no one said learning was easy.

  • Magic

    “If you knew how much work went into it, you wouldn’t call it genius.”
    Michelangelo

    It’s that mysterious force that seems to make the impossible real, the difficult easy, and the genius accessible.  The funny thing is that magic is only magic when you don’t know how it was done.  Once you do, it’s gone.  Forever.  And that’s why perhaps so many choose not to learn how magic works because then they can stay in their fairy tale worlds where some prince/wizard/genius will come to save the day.  That’s a nice world to live in with one caveat.  If you never know how the magic was done, you’ll never be able to make it happen.  And that, in my opinion, isn’t such a nice world.  Genius takes hard work, talent is overrated, and magic is a word people say when they don’t understand something.  Or not.  It really depends which world you’re living in.

Top Posts

A Little Bit of Slope Makes Up for a Lot of Y-InterceptJuly 25, 2015
Scott Adams’ Financial AdviceJanuary 25, 2015
A Slot Machine RiddleDecember 11, 2012
Lucky or smart?July 28, 2013
You can't outrun a bearApril 9, 2016

Past Posts