# Designing for Change: Embracing Uncertainty Newsy: What is Good Design? Published: 2025-10-28T14:30:00.182Z Source: https://newsyai.com/edition/67f720a9-572a-475f-9b07-bc5e65f2f40c Archive: https://newsyai.com/newsy/5c850c31-aa90-4fca-8708-9148e1d10645.md --- # What Is Good Design? ## Today’s Lens: Stewart Brand > “All buildings are predictions. All predictions are wrong.” — Stewart Brand, How Buildings Learn Brand’s provocation isn’t fatalism—it’s a design brief. If every artifact is a bet on the future, then “good” design isn’t about being right once; it’s about being ready repeatedly. Brand’s idea of “shearing layers” shows how things age at different speeds—structure lasts, services change, space plans churn, stuff churns fastest. The point: design should let the fast layers dance without tearing the slow ones. Translated beyond architecture: treat products, orgs, and software as time-layered systems. Lock in the load‑bearing decisions; keep the rest swappable. Good design is option‑rich, not option‑fragile. It earns longevity by making change cheap in the places that must evolve, and making stability reliable where it matters. - Slow layers: purpose, values, core architecture, data model. - Medium layers: workflows, policies, APIs. - Fast layers: UI, content, pricing, campaigns. When fast layers can revise without renegotiating the slow ones, evolution becomes maintenance—not a crisis. [How Buildings Learn](https://mitpress.mit.edu/9780140139969/how-buildings-learn/) • [Shearing Layers overview (The Long Now Foundation)](https://longnow.org/ideas/02004/06/01/shearing-layers-of-change/) ### Try this 1. Layer map: list your system’s layers by speed. For each fast layer, define a change pathway that doesn’t touch the slow layer beneath it. 2. Option budget: allocate a fixed monthly “change allowance” (time, money, attention) and spend it where learning per dollar is highest. 3. Reversibility rule: any change in a fast layer must be reversible within one day and one click/flag/feature toggle. Good design, on Brand’s terms, is time‑literate: it accepts that we will be wrong—and makes being wrong inexpensive, safe, and informative.