On boring software in interesting times

The two LuvAI surfaces — this studio site and the PromptCraft product — were both designed under a constraint that surprises some visitors: I wanted them to look like they could have been printed in a Japanese industrial-equipment catalogue from 1972. Cream paper, vermillion accents, sober tracked-out display type, a typographic rhythm that doesn't beg for attention. On the studio side it's slightly different — white paper, italic serif — but the underlying instinct is the same. Make the visual language so calm that it disappears.

People sometimes ask why. The honest answer is a sentence I've been saying for years and only recently learned to defend properly: I want the boring choice on aesthetics, because everything underneath is moving so fast.

Here is what I mean. The technology I build on top of — the AI models, the inference APIs, the deployment infrastructure — is in the most volatile period of its lifecycle. New models ship every month. Best practices for prompting drift on a quarterly basis. Pricing changes. Capabilities multiply. The half-life of a "groundbreaking AI feature" in 2026 is something like eight weeks.

In a context like that, every additional unit of trend-following I do at the surface layer is a unit of future-me-having-to-redesign that I'm pre-buying. If the homepage is built around an animation that screams *2026*, by 2027 it will scream *outdated*. If the product UI leans on a visual idiom that's hot this season, by next season I'll be choosing between looking dated or doing a redesign instead of building features.

The opposite move — the one I picked — is to choose a visual language that's already aged out of fashion. Catalogue typography. Industrial Japanese. Black-and-white serifs. The kind of look that has been quietly working for sixty years and will continue to quietly work for another sixty. It can't go out of style because it's not in style. It's in *form*.

The deeper reason this matters is that it lets the work underneath be the thing that earns the user's trust. When the surface is calm, the value of the product has to come from the product. There's no startup-of-the-week aesthetic borrowing credibility from somewhere else. There's just the thing, and how good it is, and whether it does what it claims.

This is, I'll admit, partly a defense of personal taste. I genuinely like quiet design. I find loud design exhausting. The studio site you're reading is, for me, restful in a way that startups built on neon gradients aren't. But I don't want to dress up an aesthetic preference as a strategy if it isn't one. So here's the strategy part: the calmer the surface, the longer it lasts, and the more of my time goes into the parts of the work that actually matter.

There's a related observation I want to record before this essay ends. The slick AI-product aesthetic of the last two years — the dark-mode gradient, the AI-generated hero illustration, the floating glassmorphic cards — is, in 2026, already starting to look like a period piece. You can date a product to within six months by its visual idioms. That's the curse of leaning into the Now: the Now is, by definition, ephemeral.

Boring is durable. Trends age. If you're building something you want to be around for a while — and as a one-person studio, there's basically nothing else worth building — then the boring choice on the surface is the long bet, and almost always the right one.

The catalogue from 1972 still looks good. That's the gold standard.