Having just finished Kevin Kelly’s What Technology Wants and Ray Kurzweil’s The Singularity Is Near, I’ve been doing a lot of thinking about the pace of change, especially in the field I’m most familiar with: web-based startups. When I started WebOnTap in 1998, here’s what we spent our first six months doing:
- Choosing our software platform. PHP was untested, none of us knew perl, and ASP was a joke (Windows hadn’t conceived of .Net yet). We chose J2EE (“Java beans” were the latest big idea from Sun) because it was new and hot and we figured we’d be readily able to find developers who knew it or were willing to learn (we were wrong). Licenses for the SunOS, Oracle’s database, and the Java development environment easily exceeded $15,000.
- Finding servers to host our product. Machines weren’t powerful enough (we thought) to run database/application/webserver on a single box, so we built separate database, application, and hosting machines, and created failover backups for each tier. It took months to apply and get approved to join Sun’s program for startups which got us “loaner” machines, which eventually would have cost us upward of $40,000. We ran our own DNS and mailservers.
- Finding a place to host our product. The co-location services we explored were far too expensive, since they were tailored for enterprise firms. Managed hosting was a fledgling concept, and the idea of shared hosting was laughable. So the boxes lived under a desk in our two-room office on the second floor above a deli and a chinese restaurant in Newton, Massachusetts. In the summer the room got so hot everyone worked in shirtsleeves.
- Hiring a systems admin. One full-time salary was dedicated to the guy responsible for keeping the machines running.
Total cost: $100,000
Total time: six months
When we started Cauzoom in 2011, we got all that (with processing speeds two orders of magnitude faster) from a hosting company we found on the Internet based on ten minutes of research. There wasn’t even a question about using the LAMP stack (Linux/Apache/MySql/PHP), and to this day I don’t know what kind of machine our stuff runs on, or where it’s physically located (Florida, I think). Over two years I can recall two sub-twenty minute outages, and I don’t know the names of any of the people watching over our boxes.
Total cost: $20
Total time: ten minutes
I won’t even mention the design tools and helper applications we paid big dollars for, or the marketing platforms and APIs we didn’t use (because they didn’t exist). In 2011 we rented dozens of supporting systems and tools for cheap or free – all in, we paid maybe $50/month.
I realize that recounting this in detail will simply underscore how old I am (twentysomething entrepreneurs were in elementary school at the time). I rolled my eyes at the forty-year-old Digital and BBN guys I hung out with back then when they’d tell me stories of working on the VAX, or using Archie to search FTP archives. In many ways what we did then is now largely meaningless – the brands and code and services we used and built have all been superceded by several generations of progress.
But one thing it does is make me appreciate how far we’ve come, and how quickly. Fifteen years wasn’t that long ago in real terms. Think of any prior fifteen year period and you’d be hard-pressed to find as much of a difference in technology development. 1960-1975? 1940-1955? 1890-1905? This of course is the point that Kelly and Kurzweil are making: everything is speeding up.
When it comes to web-based business, the real wonder is how fast one can move. The distance between concept and execution has never been smaller than it is right now, and will only get smaller. I find it difficult to imagine what kind of environment we’ll be working in ten years from now. Will the Facebook API and Google Analytics seem as archaic as the SunOS does today?
Now that would make an interesting post.
