Saturday, February 18, 2012

Training for our Future

The education system, by its very nature, is rooted in the past.  But when it comes to advanced educational degrees particularly in the sciences, where the typical entering graduate student may not be in the real job market for another decade, a curriculum that is focused on the future may be more advantageous.  In the last few decades we have seen an explosion in technological advancement, which the futurist; Ray Kurzweil, calls the singularity.  This singularity is characterized by an exponential curve in the rate of technological advancement that will eventually get to a point where we as humans will no longer be capable of predicting the future.  It also suggests that in past generations, predicting the future was not the necessity that it may be today, due to the slow rate of change.  Therefore, an education system did not have to direct much effort toward explaining to students about the rapid changes that would take place in years to come.  For example, farming practices may have changed gradually over time, but the mainstay of a farmer’s education could be focused on learning the tricks and techniques of their fathers.

We no longer live in that time.  Within my own lifetime, starting in 1980, I have seen a shift from land line phones to ubiquitous cell phone use.  I have seen the shift from no computers in most homes, to at least one if not several in every home or for that matter in your pocket or purse.  Likewise, the changes can be seen in all sectors of employment.  Which means, that by the time I have completed my graduate degree, the field will have radically changed from the time I entered my undergraduate degree (which in my case is actually two separate undergraduate degrees, not completed sequentially or followed directly by graduate school, i.e. roughly a decade of time has elapsed thus far).
One might ask how you can focus on the future when it hasn’t happened yet?  And the answer is that humans are poor at predicting the future with good accuracy, but we are much better at it than our ape-like ancestors or pretty much all of the other organisms on this planet.  Furthermore, every year that I exist, we get better at it.  And this is where I find a flaw in Kursweil’s argument.  That is that, technology does seem to advance on an exponential curve as he points out in his book ‘The Singularity is Near’ but it does not follow that we will come to a point where we will not be able to predict it.  We as humans will use our ingenuity to find better predictive models.  One such model is that proposed by mathematician, Alberto-Laslo Barabasi, which describes a division of graph theory that explains and may predict network interactions.  These network interactions don’t just apply to the coming and going of individuals in a social network but could be applied to metabolic systems, to economic markets, or even neuronal connections in the human brain.  This type of predictive modeling and others may eventually be used to predict your behavior in the way that Google is already doing by following and predict your marketing preferences online.  So how does this all relate back to the science and technology education system?  One could imagine that if marketers can predict the proper market for their product pipeline then educators should be able to predict the demand for various professions within at least a range of say one to two standard deviations.
 It is also fathomable that these same predictive algorithms, none of which I will delineate here, may not only predict future employment options and funding sources for future scientists, but may indeed be able to predict which technologies will be available and what skills will be needed to be competent in these future markets.  Today's system falls woefully short of these goals.  I postulate, that a change in higher education design, which incorporates predictive algorithms to streamline programs and guide new students,would produce a more fluid and effective model for producing trained professionals in a rapidly advancing world.  This system would also be more cost effective at the institutional level and for the individual seeking a direct route to employability.


  1. I am not a big fan of bright font on black backgrounds, kinda gives me a headache.

    Having spent a little time in research myself I have realized that while you can make broad "exponential" statements like "such and such body of knowledge doubles every three years" there is a large amount of that "knowledge", especially in theoretical sciences that is unconfirmed or may even be downright fallacious. So while knowledge increases, so does noise and the need for experts to discriminate. Just something to think about.

    I am convinced that processes like growth and knowledge accumulation are best modeled by michaelis-menton type curves that model systems that reach a point of saturation (fastest possible computation speed, complete knowledge, maximum size given energy input, etc.) I've run across these curves in such diverse subjects that I feel like I see them everywhere now...

    I would like to remind the readers that innovations have historically not sprung from empires (think Egypt or Persia). Empires have a vested interest in maintaining the status quo and feed only existing institutions. It is in melting pots on the frontiers, where people are trying to build societies and change their lives for the better, and where they are free to ignore the empires and their institutions, that true innovation takes place(think Greece and America).

  2. (I will take the formatting suggestion under advisement)

    Both of Nick's points are valid. However, the point made about michaelis-menton type curves may only apply in the short term. The exponential curve is a michaelis-menton curve flipped on its head and in my initial post, I pointed to the work of Kursweil whose argument about accelerated returns can be found in his book and at the following link ( In which he states that when technologies reach their 'michaelis-menton' plateaus, then a new technology takes its place, see computation speed versus price as a function of Moore's Law of computer curcuit design which (Kursweil points out) is not the first but fifth in a series of technological advances leading to reduced computing price-performance.

    As to Nick's second argument about the best innovations occurring on the frontiers of society, one must consider that in advanced technical fields scientists and engineers are creating a metaphorical frontier in which regulatory agencies must struggle to keep up.


As always, this blog is also a forum, discussion is appreciated.