Fueling a furnace with banknotes?
Today, I have found out that yet another of my academic friends accepted a position with Google and will join the company after completing his teaching assignment this semester. Because his current position is visiting and non-tenure track, the Google job will offer him permanent employment with much needed job security in this volatile job market. That’s why I heartily congratulated my friend.
Nevertheless, by accepting this position, my friend has essentially dropped out of the academic game. Because I know him as a capable researcher with interesting ideas and insights, I wonder if working as a programmer for Google is the best application of his abilities and training.
There is a Russian saying that can be roughly translated as “to fuel a furnace with banknotes” (топить печку облигациями), which refers to wasting valuable resources, particularly those capable of yielding tangible future benefits. Of course one can fuel a furnace with banknotes, but scrap paper will do just fine. Training a Ph.D. level scientist incurs enormous financial and human costs, both for the individual and the society at large. I would like to hope that Google will provide a challenging and nurturing environment to fully leverage my friend’s superior Computer Science expertise and abilities. And indeed many prominent researchers feel completely fulfilled having left academia for Google (e.g., see Matt Welsh’s reflections on the subject). Nevertheless, I wonder if all developer jobs even in Google can truly justify the investment of a Ph.D. degree.
As recently as three years ago, my friend was adamantly committed to pursuing an academic career. He accepted a postdoc position and was working diligently on strengthening his CV to compete on the academic job market. What was it that convinced him to leave the world of academic research? I can only guess. But I am sure that observing the realities of modern day academia (a shortage of open academic positions, the very tight federal research dollar, the immense hurdles one must overcome to build a successful tenure case, all come to mind) must have played a role .
Once again I do hope that things work out well for my friend, and that he finds his Google job as fulfilling as many former academic researchers do. However, I see Ph.D. level scientists accepting programming jobs as a general trend on which I would like to comment. I believe that in hiring a Ph.D. in Computer Science to work as a coder would be indeed like fueling a furnace with banknotes. By a coder I mean a person who primarily writes computer code for a living. Such an arrangement is a great deal for the programming shops that hire Ph.D.s–a person who has spent 5-9 years building complex research artifacts and delving into the complexities of CS theory in most cases is a competent programmer. (Another question altogether is whether Ph.D. training can adequately prepare one for the challenges of day-to-day Software Engineering)
Of course not all industry jobs involve programming only. Ph.D. training does prepare people to tackle complex problems that require systematic and multifaceted problem-solving skills and enormous perseverance. But does the previous sentence reflect the reality of most modern software development jobs often claimed by Ph.D. level computer scientists? I would like to hope so, but I am not sure.
Besides what is the impact on fundamental research and the overall competitiveness of this country of all these highly training scientists choosing not to pursue a research career? Do we have a glut of Ph.D. level scientists in computing to afford not to care about this issue? If this is the case, isn’t overproduction of Ph.D.s is a good use of this country’s limited basic research budgets?