Design engineering essays, product marketing, and visual storytelling from Affordance Design Studio.
It started the way most trouble starts — a designer wandered past the Figma canvas and accidentally opened a terminal. Rather than retreat, they took notes. Sixteen dispatches followed from the strange frontier where pull requests shape user experience, error states deserve standing ovations, and "it works on my machine" is the punchline to every production joke. Each illustrated in rubber-hose style, because some truths land softer when a cartoon character delivers the bad news. The whole operation — every page, every animation, every deploy — is orchestrated by HAL 9000, our in-house rogue AI agent, using Claude Code. Because if you're going to write about production systems, you might as well let an AI run yours.
The most important design decisions happen before a single pixel appears. Data models, API contracts, and state management shape what users experience far more than colour palettes.
For every flow that works perfectly, there is an equal and opposite way it can break. Design the failure states with the same care you give the success screens.
Users trust systems that behave predictably. A deterministic response every time beats an impressive-but-unpredictable AI guess. Reliability is the real magic trick.
When something goes wrong, logs are the first thing anyone reaches for. Good logging is good UX — it tells the story of what happened, not just that something failed.
Every millisecond of delay is a design decision. Latency shapes perceived quality, user trust, and interaction patterns. Treat it as a first-class design constraint.
AI-generated content and grounded facts serve different purposes. Keep them architecturally separate so users always know what is invented and what is verified.
The gap between localhost and production is where most failures live. Network latency, auth, rate limits, cold starts — none of these exist on your machine.
API design is UX design. The shape of your endpoints determines what interactions are easy, hard, or impossible. Bad APIs create bad experiences no matter how pretty the front end.
In the cloud, failure is not the exception — it is the baseline. Design for degraded states, partial outages, and retry storms from the start, not as an afterthought.
Designers who understand caching, CDNs, and database queries make better decisions. Infrastructure literacy leads to interfaces that work with the system, not against it.
Every layer of abstraction hides complexity — and potential failure modes. The simplest architecture that solves the problem is usually the most reliable one.
Good AI UX removes cognitive load — it does not pile on more text for users to parse. If your AI feature adds explanations instead of removing steps, rethink the design.
You learn more from one deployment than from ten planning sessions. Ship early, ship often, and let real-world feedback guide the next iteration.
Understanding what happens behind the API call makes you a more effective designer. Backend empathy means designing with awareness of system constraints, not just user wishes.
When you own the outcome end-to-end — design, build, deploy, maintain — you make fundamentally different decisions. Ownership creates accountability and better products.
Learning to code did not make me an engineer — it made me a better designer. Understanding the medium you design for gives you superpowers, not a career change.
