Context Beats Design

I love design. As far back as I can remember, the love for design comes from my early childhood dream of becoming an architect. I used to stare at buildings, roads, and intersections with a feeling I could not name at that time. I wondered how light could enter a room or how a space should feel before anyone walked through it.

That instinct never left. Today, I spend an unreasonable amount of time admiring well-designed products, interaction/motion design, websites, art, and buildings. People who truly care about design probably share some version of this.

This piece is about a pattern I keep noticing in design as a discipline.

There is a debate right now about what AI is doing to design. Some designers say the slow exploration phase, the part where you try multiple ideas before committing to one mock design, is being skipped entirely. People are building working software in a weekend that used to take a team a quarter or more.

I have personally seen what vibe-coded apps look like. Some are unbelievably good looking with well-thought structure. I've also seen people saying "I replaced Calendly with something I built over the weekend" or "We rebuilt our entire internal tool stack." The UI in the screenshots they post is mostly terrible with generic layouts, odd colour choices, odd component choices, or a bland template.

This got me thinking: if design matters so much, why do some of the most widely used products in the world have interfaces that are hard to use?

• SAP Concur is a good place to start. It handles corporate expense billing for thousands of companies. Simply said, the interface is bad, and yet, almost every company use it.
• Indian government websites tell the same story in a different language. IRCTC, the railway booking site, has its own meme genre.
• GOV.UK website. Text-heavy with almost no images. Just words and links, organized with absolute simplicity.

About AI chatbox interface debate

There is a growing belief that conversational AI will replace traditional interfaces. Everything will become a chat window. You type what you want, and the AI does the job. No buttons, no menus, no forms to fill.

Nielsen Norman Group tested this assumption directly: observing users struggle with AI interfaces felt like a pre-GUI era. The chat interface looks simple but places the entire burden of discoverability on the user. A screen with four buttons tells you the app does four things. A chat window tells you nothing until you ask.

So does UI really matter? The pattern, when I sit with it long enough, becomes simple.

I think, design's importance scales with the distance between the person making the product (Builder) and the person using it (User).

• If you are building an internal tool for yourself, often times, functionality is enough. The ugly spreadsheet wins because you built it and people who use it understands it.
• If you're building enterprise software sold to a procurement department, things like compliance, integration, and switching cost tend to matter more than the interface.
• If you are a government with a monopoly on services, not for adoption. People will use your website no matter how bad it is.

But, if you are building a consumer product, a SaaS tool, an app that competes for attention in a market where switching costs are low, design matters more than almost anything else. In some cases, even more than distribution itself.

Two interesting examples that show this clearly:

• Picture a craftsman building his own chair from a tree he chopped down personally. Instead of focusing on aesthetics, he shapes the wood entirely around his body. Once it's built, it works, because the person who built it and the person using it are the same person. There's no gap to close.
• If you follow X, you've probably noticed that some well-known builders launch products with rough, unpolished early design. Mission Control HQ and OpenClaw are two examples. In both cases, the builders were already well known and trusted for their work, so people used the product despite the rough edges.

In both cases, the gap between the person building and the person using was small. When that gap is small, good design becomes less important. But, when the gap is large, design carries most of the weight.