Cycling Hobbies to Day Jobs: A Personal Assessment of Starting a Startup

When I was exposed to programming in an official capacity (university) I hated it. At, and before that time my hobbies included learning about computer networking, hacking around in Linux, and modding computer games, much of which I later realised involved 'programming' (compiling drivers, changing and compiling game code, writing scripts, etc.). Later, after I made this association, and more importantly when I saw the creative power of the medium, my spare time was filled primarily with programming projects and the rapid assimilation anything and everything that would help me create better or more interesting programming artifacts. Work products from my spare time crept into my student life (code examples, awesome assignments, etc.) and I was quickly employed to do it for a living.

This recurrent process of transitioning 'hobby' to 'day job' has been a common theme with my interactions with computers. For example, I was exposed to networking technologies (which I disliked intently) when I spent my spare time playing games, which progressively overtook my interest in games after participating and organising in LAN Parties and Modem variations of the games I was playing. Presently my 'day job' is research in Artificial Intelligence, which was my hobby when I worked as a software engineer (about the most interesting thing in from the study of programming I could find). Observing this trend, one may be inclined to consider the exploitation of this organic process: what is your hobby (something you love, enjoy at the least), and how can it pay the rent (become a day job)?

Postgraduate life is awesome, although having to deliver work product in an official capacity can suck the fun from the experience without discipline. For example, I dislike publishing my research (an academic career killer), although I love writing it up, particularly for the refinement of my ideas that it affords. The important aspect of having a stake in the research project (essentially being the sole stakeholder) is that you actually care about the end product. As a result, I have found that my hobby and my day job have merged. This is not completely accurate, as I have invested time in researching the hell out of interesting pocks of work that do not directly contribute to my project that may be considered a 'hobby' in the 80/20 consideration of work (for example complex adaptive systems, satisficing, human computation, and all kinds of methodologies). From a more abstract perspective, my hobby over the last 3 years may have been a form of project management, while may day job has been to 'do research'. Besides reading about esoteric facets of computer science, and stressing about research project schedulling, I have pondered the latest web technologies and principles of being involved in a technology startup.

A rational assessment suggests that: Software Engineer + Intelligent Systems Research = Job Building Intelligent System Software. I'm cool with that, although I question whether the only options for such work are academia and big (read: someone else's) business. The responsibility of management and control has impaired by desire (not ability mind you) to do someone else's bidding. (Un?)Fortunately there are many resources out there for people thinking along these lines, the least of which is the page-rank prominent Paul Graham's Essays. Frankly, I enjoy his writing style, the optimism, and I agree on much of what he has to say regarding education and technology, although I'm pragmatic enough to not take his retrospective advice too literally. Specifically, I subscribe to the messages of 'work on what you love' (I live and breathe 'always produce'), 'having the guts to try', and 'work on hard problems', which he revisits many times in his writings.

Regarding the 'work on what you love', I have had the fortune, and now the bias of thinking that if you work had enough on your personal passions, that good things will come, like getting a day job. An observed flip side is that I keep needing new hobbies, this is because when fun work is structured with responsibility and expectation, it's no longer as fun. Regarding 'having the guts to try', it's a no-brainer for anyone ambitious, although it is a hard lesson to take on board, especially when pushed by someone who's current day job depends on it. By the way, I think it is brilliant to start a startup that helps people start startups. If only you could mitigate the risk of running such a thing remotely (franchise?), I'm sure the world wide market would be huge! Anyway, having attractive opportunities fall in your lap removes the need to make this important and hard decision. I am at this crossroad now, and I equate it to the crossroad I faced nearly four years ago of whether to give up a perfectly good job (and relatively linear career) as a consultant to research AI. I think rationally decision making is not the problem, rather it is the objectivity of the process, specifically the emotional impact of the unknowns involved. Finally, regarding 'working on hard problems' Hamming's advice has influenced me since the day I first read it, and hat tip to anyone spreading the good word.

Two cases pushed me over the line to help me to decide to pursue my own technology startup before seeking classical employment: (1) As a tool, intelligent systems are an underutilised powerhouse which translates to competitive advantage (Graham's popularised example makes this point), and (2) An ex-colleague deliberated on the decision, gave it a go, and successfully launched a product (Elimatta) in 3 months (localised confirmation that it can be done). Additional no-brainier points include my own financial security, the low and continually decreasing cost of bringing web products and services to market, and my lack of broader commitments.

I am left with questions regarding the relationship between hobbies and day jobs, particularly given my biased perspective. My suspicion is their present merger will hold for sometime yet as I'm sure starting a startup, much like graduate research, does not promote consideration of additional pleasurable workload.