Our first website went live in 2001, under a different name. We were Siite Interactive then, and the oldest surviving capture of that site, preserved in the Internet Archive from July of 2001, still carries a line we chose to put front and center. It's from R. Buckminster Fuller:
When I am working on a problem I never think about beauty. I only think about how to solve the problem. But when I have finished, if the solution is not beautiful, I know it is wrong.
Twenty-five years later, that quote is still the truest thing on any page we've ever published. Almost everything else about the web, and about us, has been torn down and rebuilt several times over. This is a look back at both.
The World We Started In
It is hard to overstate how different the ground was in 2001 and 2002, when we began operating in earnest. Most people did not own a mobile phone. Smartphones were years away; the iPhone would not arrive until 2007. The first iPod shipped in late 2001, the same year we incorporated, and it was a novelty tethered to a computer, not the pocket supercomputer its descendants would become. By one widely cited estimate, it was right around this time that the world's stored information tipped from mostly analog to mostly digital for the first time. We were, without quite knowing it, opening a business at the hinge of that change.
A website, in that world, was something you still had to talk clients into. The conversation started at "why do you need one at all."
The Question Kept Changing
If you do this long enough, you notice the work isn't really about a technology. It's about answering the same underlying question as its shape keeps changing.
For a while the conversation was why you need a website. Then it became why you need a mobile website, back when the answer was often a separate stripped-down site at an m-dot address. Then why you need a responsive website, once we all agreed that one fluid site should adapt to any screen instead of maintaining two. The responsive revolution was kicked off by Ethan Marcotte, a classmate and friend, who gave the idea its name and, in his book Responsive Web Design, taught the whole industry how to build that way. Then why you need to think about page speed, about accessibility, about structured data, about how an AI assistant will summarize your business when no one visits your homepage at all.
Every few years the question molted and a client asked us, in good faith, why the thing we recommended two years ago was no longer the thing. The honest answer was always the same: because the ground moved again. It always does.
Wave After Wave
The technology arrived in waves, each one large enough to feel permanent while it lasted.
There was the LAMP stack revolution that made dynamic, database-driven sites affordable, and the rise of WordPress that put publishing in everyone's hands. There was the arrival of social media, which changed where audiences lived and what a website was even for. There was the proliferation of content management systems, and the slow professionalization of search engine optimization from a bag of tricks into a genuine discipline. There was mobile, then responsive, then the performance and accessibility reckonings, then headless architectures and the serverless shift, and, most recently, a wave large enough to deserve its own chapter.
We rode each wave, and we watched each one recede into just another layer of the sediment. The lesson repeats: no stack is forever, and the teams that survive are the ones that hold their principles loosely enough to change their tools and tightly enough to keep their standards.
The Wave We're On Now
The defining force of this era is artificial intelligence, and it is the largest shift we have seen since the web itself. It is not a feature you bolt onto a website. It changes how software is built, how content is made, and increasingly how people find businesses at all, when an assistant answers the question instead of sending someone to your homepage.
Our answer to a shift this big was the same as it has always been: build, don't wait. We built Replicator, an AI brand operating system where a language model designs marketing documents and drafts copy inside brand-approved rails, and a human finishes each one by clicking into the page and typing. Along the way we needed a reliable way to hand structured design data to a language model, so we solved that problem too and open-sourced the result as barkup, a small library for authoring typed trees as the HTML that models are already fluent in. The reasoning behind both is written up across three engineering posts on this blog.
We are not adapting to AI by watching it. We are shipping with it, and giving some of what we learn back to the community, which is exactly what a twenty-five-year-old agency built on open-source foundations ought to do.
The Company Changed Too
We did not just watch the industry change. We changed with it, sometimes by choice and sometimes because we had to.
We have been as many as twenty people in an office near Grand Central Station in New York. We opened an office of a dozen in San Antonio. We went fully remote before the pandemic made everyone else do it. Over the years, more than a hundred people have worked directly for this agency, and together they have built literally hundreds of websites and web applications. Our founder, Alan Ruthazer, eventually stepped away from the business. We grew, and shrank, and grew differently, and reorganized more times than is comfortable to admit, because an agency that refuses to change in an industry defined by change does not reach its twenty-fifth year.
What We Refused to Change
Two things survived all of it.
The first is our independence. True to the Gen X sensibility we were founded on, we never sold, never got absorbed, never became a line item in someone else's holding company. Staying independent has cost us opportunities and saved us from worse ones, and it remains the single decision we are proudest of.
The second is that Buckminster Fuller line from the very first site. We have never treated beauty as decoration applied at the end. We treat it the way Fuller did: as a signal. When a solution comes together and it is not beautiful, that ugliness is information, a sign that something underneath is still wrong. Twenty-five years of shifting stacks have only made us more sure of it. The frameworks are disposable. The standard is not.
Onward
Nothing about this anniversary is nostalgic. The tools in our hands right now are the newest we have ever held, and we are using them the way we have used every generation before: to do the same work, which never changed. Solve the real problem, and refuse to call it done until the solution is beautiful.
Here is to the people who built the first twenty-five years, the clients who trusted us across every reinvention, and, especially, to Alan Ruthazer, whose leadership and vision started all of it and without whom this anniversary would never have been possible. And here is to whatever the question turns out to be next.
