Founder Keith McCall on building companies, software, and real-world execution.
There are places we return to—not with our feet, but with something deeper. Memory. Centering. Safety. When I was a young boy growing up in Victoria, BC, we lived in a humble house with something extraordinary in the backyard: a massive willow tree. It wasn’t manicured or symbolic at the time. It was just there—wide, forgiving, and full of possibility.
It was just after the 1996 Atlanta Olympic Games. The dust had settled, the systems had held, and what the world was just starting to call the “World Wide Web” had been pushed, hard, into reality. We were flown in—IBM’s global “Brat Pack”—into New York. The ceremony was held just outside Armonk, near headquarters. A ballroom. Formal. Polished. This was IBM at full gravity.
It was 1995. The internet wasn’t inevitable yet—it was fragile, experimental, and misunderstood. Inside IBM, though, there was a different read on the future. Under Irving Wladawsky-Berger, and led day-to-day by John Patrick, IBM made a move that, in hindsight, looks obvious—but at the time was anything but. IBM created an Internet Division. Not a side project. Not a research thread. A real, focused push into what would become the backbone of everything.
BizSpark gave us access to enterprise-grade software—Microsoft Exchange Server and Microsoft Windows Server—with minimal upfront cost. There were caps, sure. But for a startup, it was more than enough to build. And a small number of licenses to deploy our team to use Microsoft Windows, Office, and Visual Studio - consistent with Steve Ballmer's stated goal for more .NET developers.
That first night, I went down to the classy old bar early—and ended up sitting with Ray Ozzie. I was 26. DB2 background. Relational mindset. I asked what felt like a dumb question: “Why is Notes different from a relational database?” He didn’t brush it off. He didn’t shortcut it. He explained it—clearly, patiently, in a way that connected.
That first night, I went down to the classy old bar early—and ended up sitting with Ray Ozzie. I was 26. DB2 background. Relational mindset. I asked what felt like a dumb question: “Why is Notes different from a relational database?” He didn’t brush it off. He didn’t shortcut it. He explained it—clearly, patiently, in a way that connected.
It was 1994. I was deep in the mechanics of the early web—HTTP servers, CGI, figuring out how to move real data into something a browser could render. At IBM, we were working with IBM WebExplorer, building interfaces before the world even agreed on what a “web app” was.
I was working at IBM, making about $40,000 CAD annually — roughly $30,000 USD — living in Toronto and trying to do all the “right” things. My wife worked as a private school teacher. We bought a house, partly because that’s what you’re supposed to do, and mostly at the encouragement of our parents. In hindsight, we probably shouldn’t have.
It was 1991. I had just graduated from the co-op program at the University of British Columbia and was stepping into my first real job at IBM. Over the previous few years, I’d worked with IBM in Vancouver and Calgary. I thought I had a plan: Build a career in sales and marketing. Keep coding as a passion on the side.
Building the 1996 Atlanta Olympic Ticket Server wasn’t the important part. The pattern behind it was.
In 1995, the web didn’t query anything. It was designed to host documents. Pages didn’t respond. They didn’t generate. They didn’t connect to live data. If you wanted a new page, you wrote a new file. Inside IBM Toronto, I built something that broke that model.
There are places that exist for meetings. And then there are places where things actually happen. For me, the Columbia Tower Club was the latter.
AI agents are only as effective as the structure of the data they operate on. Architecture determines usefulness.
In 1994, I was a young software developer at IBM Toronto working on the Distributed Computing Environment (DCE), porting the security component Kerberos from Unix to MVS and OS/400: the operating systems behind the most serious mainframe and midrange platforms of the time. Also serious: how bad the cafeteria food was.