Toward the end of college, a friend of a friend asked me what I planned to do for a living. At the time, this question occupied a strange and anxiety-rich space between the earlier, idealistic, What do you want to do when you grow up? and the present-tense default, What do you do? I was not yet aware of Information Architecture as a thing, much less User Experience—though I doubt those terms would have helped in the moment. So I blurted out something like, “I want to design websites.”
The response was brutally efficient: “That’s great until anyone can do it.”
There’s a logic to it. The web itself was still relatively new. The tools to work in this new medium were unrefined and the technical know-how scarce. Fast forward and it’s easy to imagine a world in which making a website is as easy as sending an email.
In a way, that’s what happened. At least for basic sites, the tools and services available today do make it easy. Putting text and images on a page and linking pages together, picking out templates and playing with colors—this no longer requires technical knowledge.
But how far can that trendline extend? I was struck when I read the original pitch for a startup called The Grid: “AI Websites That Design Themselves.” That’s certainly a claim my undergrad self would not have imagined.
The mechanics of how this service works is less the point (one key feature is the use of sliders – say, between formal and informal – from which the system infers web design choices including font and color palettes). Instead it’s worth reflecting on the broader concept: what are the underlying assumptions in this proposal? The implied benefit of less work for the would-be website proprietor sounds good at first. But what are we giving up? What is lost with the shift toward ever greater reliance on pre-made solutions, “design patterns,” and algorithmically driven structures?
What happens when we surrender intent?
The End of Writing
In his essay, Proposals Toward the End of Writing, Tony Tulathimutte examines the similar forces of technology and automation, but from the perspective of authors and writing. Projecting from now-common tools like spelling and grammar checkers, he imagines a near-future cliché detector: “a simple extension of the language-checking features already baked into most word processing software, underlining each trite phrase with a baby-blue squiggle.”
Leaving aside the benefit this would have on the writing you’re reading now, imagine if this feature popped up on everyone’s Google Docs interface (in beta, of course). What second-order effects might that have on writing? Tulathimutte answers: “But perhaps automated blemish-free prose may become tacky and cloying, the prose equivalent of Photoshop or Auto-Tune – the deliberate use of cliché may become an act of subversive camp, or a reassuring watermark of human authorship.”
It’s easy to see parallels to the web. As more and more sites began to adopt the visual language and inherent structure of the prevailing templates and pattern libraries, a pristine sameness emerged.
We’re already seeing the zag away from the Squarespace zig in so-called Brutalist web design. Katherine Arcement, in The Washington Post, describes these contrarian sites that “eschew the templated, user-friendly interfaces that have long been the industry’s best practice. Instead they’re built on imperfect, hand-coded HTML and take their design cues from ’90s graphics.”
While the aesthetic in the examples is stark, there’s no denying the “reassuring watermark of human authorship.” In fact, these sites would be straight-up difficult to achieve with the streamlined tools common today. You would have to actively work to break the mold. The result though is a design with evidence of its creator – and of the creator’s intent – baked into the structure.
And what if we continue on the path toward websites that design themselves – or books that write themselves – removing the creator altogether? Tulathimutte traces his cliché detector out to where computers can generate entire texts from scratch, beyond just offering suggestions on what you’ve already written: “Supposing that it is possible for a computer to generate writing, the question then becomes: is it any good?”
In this case, we give up some intent by leaning on industry best practices, established design norms, and maybe even some artificial intelligence. The benefit is the time and effort saved. On the other hand, there’s the trend toward design as craft, where every button is bespoke and every tag hand-crafted. Clearly neither of these extremes are tenable. But what is the right mix for what you’re trying to achieve? How do you figure out what would be good to do?
A Proposal for Understanding Intent
This notion of good is deeply embedded in our approach at TUG. We work closely with client teams to build shared understanding around the question: What does good mean? Why are we building (or changing) this digital place? Who is going to use it? What sort of activities do we want to support there? And how will we measure success?
We start by asking. We talk to stakeholders to identify themes, and we illustrate competing ideas with intention modeling. This exercise creates a visual representation of their ideas and a springboard for working out where the sweet spot is between two goods. This work lays the foundation for the architecting and design to come, ensuring good fit between your strategy and the structure that we shape together.
So if your digital needs would benefit from a focus on intent and outcomes—whether to support customers quickly finding the right information or to meaningfully connect employees across silos—we are here to serve, human watermarks and all.