The list of things describing why I’m leaving graduate school to found a startup.

February 18th, 2009

I originally decided to not write this post as each attempt became a longwinded dissertation on things that no one other than me cares about, but since someone apparently does care, I figured that this would be a good opportunity to document this point in my life.

In order to satisfy the curious, here are my reasons for leaving graduate school to found a startup.

First, a disclaimer: The following list describes items that are of relevance to me and me alone. My situation is different than yours, and my reasoning is not applicable to your situation. If you’re a graduate student reading this, do not read this as me trying to persuade you to leave graduate school for a startup.

Now, the list:

1. I can build a better career in industry than I can in academia. Philip Greenspun has a sobering take on academic (science-based, in particular) careers.

2. The nature of academic employment almost guarantees that I have to leave Chicago to work as a researcher. There are a few open slots for researchers that do my kind of work, and a huge number of superbly qualified people to fill them. Realistically, I’d have to look outside Chicago for employment if I wanted to work as a researcher. Researcher salaries do not include enough compensation to make up for what I would lose if I left this city.

3. The golden chalice of tenure holds absolutely no appeal to me. Why kill myself (metaphorically) while my tenure clock ticks if I won’t appreciate it when I get it?

4. I can do my work more efficiently when I have full control of the process.

Here’s a good example of this. Over the past two years, I’ve been working hard to bring context-aware computing to mainstream users and developers. One of the challenges of establishing the value of this software is proving in some way that it’s actually useful. To achieve this goal, I’ve taken two simultaneous approaches. The site linked above takes you to the open source page for the software where you can download and try it yourself. I’ve also been working on a proper formal study to see whether the technology is useful for people with disabilities.

I was up and running with the open source approach as soon as I had code that I was comfortable releasing. It took me many months of writing and paperwork to get to the point where I could even ask people to participate in the formal study linked above. The main reason that getting the study running took so long was that I had to submit what I was doing to the local institutional review board.

Now, I understand why IRBs exist and I fully support their mission. However, for something as simple as creating new software for users with disabilities and asking them to see if it was useful, the IRB added an inordinate amount of work to my plate. In the IRB’s defense, they turned around the paperwork pretty quickly, but I still spent months putting that paperwork together.

In contrast, the open-source approach of releasing software and asking for feedback has done more to improve the quality of my work, open new and interesting questions, and suggest new avenues for investigation than the formal work. However, since everyone sending me comments and feedback did not sign a consent form, none of that feedback is publishable, which means that (in an academic sense) it effectively does not exist.

5. Publish or perish. The measure of one’s worth in academia are the papers that one publishes. I’ve been told that I need X number of publications before graduating in order to secure a job. Some of my professors (thankfully, not my advisor) have gone as far as stating outright that there really isn’t a point to doing something if a publication isn’t an end-result.

I have two major problems with academia’s obsession with publications:

a. It creates a “gold rush” mentality where researchers try to stake their claim on a “hot” topic in order to get the publication credit for that area. I’ve done work in the past couple of years that I would love to share, but haven’t because I’ve kept quiet so that the work doesn’t get “scooped” before there’s a paper to cite. I’ve also read really poor papers that are clearly cynical attempts to become the paper that everyone cites by virtue of being the first one. (Huang, et al. 2005 is a great example of this phenomenon.)

b. Academics don’t own their written work. Since journals and professional organizations are the ultimate publishers of research papers, they wield a tremendous amount of clout. They typically do the rational thing (for them) and take ownership of everything they can with minimal or no compensation those generating the content. Ownership implies control, and if you are not a paying customer of the journal or association, you may not be able to access the work.

Quick reader poll: How many of you were able to download Huang’s paper that I linked to above? If you’re outside of a university or research institution, I’m willing to bet that you could not get to it from the link that I provided. If you used a search engine, then you probably found many contraband copies. However, I’m willing to bet that each copy contained the following text on the first page:

I did an experiment earlier this year to see if the ACM would be willing to let an open-source project host copies of relevant papers. In my case, I wanted to host a copy of Fogarty & Hudson’s SUBTLE paper on the Pennyworth site. I wrote a nice letter to the ACM asking how I would get permission to do this and how much it would cost. My e-mail apparently was devoured by a black hole as I never received a response.

(In the meantime, I am happy to link to Prof. Fogarty’s local copy, and pray that the URLs don’t change and the copies of the papers stay online.)

Rather than write endlessly about this, I’ll let Steven Few’s experiences with the IEEE stand in for the remainder of my complaints. In short, I don’t fancy becoming an unpaid slave of the modern academic publishing industry. I will share and distribute my own work on my own terms.

Back to the list…

6. I’m more motivated by creating concrete solutions to actual problems instead of playing rhetorical ping pong. In terms of evaluating myself, people who install and use my software generate much more satisfaction for me than people reading the theoretical justification behind its design. I’ll take 10 everyday users over 10 preeminent readers any day of the week.

7. I’m at an ideal point in my life to launch a startup. My expenses are low, my enthusiasm is high, and now is the perfect time in the business cycle to get started building a sustainable enterprise.

8. I am much more likely to achieve my major life goals with a startup than I would be in a traditional academic career. The risk/reward ratio for my new startup is much more favorable than a career as a researcher employed by someone else. And even if the startup fails, I’ll have gained a collection of skills and experience that make me even more valuable as part of someone else’s organization.

9. I have a chest of ideas that I’ve been sitting on that the startup allows me to pursue. I’m comfortable making (and testing) the business case behind each. As a researcher, I’d be forced to come up some sort of theoretical justification before beginning. The justification that I can simply do something better is sufficient for me.

10. As an entrepreneur, the only people that I have to justify myself to are my users and clients. Life’s too short to be worrying whether I’m spending it doing something useful, and I believe that the market is a more honest feedback mechanism on this front than tenure committees, anonymous reviewers, and conference organizers. If my ideas are worthwhile, the market will notice and and I’ll thrive. If my ideas are crap, the market will ignore me and I’ll be forced to adapt.

~

This post turned out much longer than I planned. Honestly, I intended to keep it short and simple!

In the interest of fairness and balance, I just added to my todo list to write a blog post 6 months after I go solo in June describing everything that I enjoyed about academia and miss. I have enjoyed my experience there tremendously, and highlighting the “why I am leaving” reasons unfairly eclipses the things that were pulling at me to stay. However, I’m not in a good position to highlight those at the moment, so following the principle of “you don’t know what you’re missing until it’s gone”, I’m waiting to write that post at an appropriate time in the future.

Update: Please see a recent coda to this post here.

This entry was posted on Wednesday, February 18th, 2009 at 10:51 am and is filed under Academia, Personal. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.

Comments are closed.