Users exporting data to build their own spreadsheets isn't a product failure, but a signal they crave control. Products should provide building blocks for users to create bespoke solutions, flipping the traditional model of dictating every feature.
Creating feature "modes" (e.g., "uphill mode") instead of exposing core mechanics (e.g., gears) creates a "nightmare bicycle." It prevents users from developing a general framework, limiting their ability to handle novel situations or repair the system.
AI chat interfaces are often mistaken for simple, accessible tools. In reality, they are power-user interfaces that expose the raw capabilities of the underlying model. Achieving great results requires skill and virtuosity, much like mastering a complex tool.
The creative process with AI involves exploring many options, most of which are imperfect. This makes the collaboration a version control problem. Users need tools to easily branch, suggest, review, and merge ideas, much like developers use Git, to manage the AI's prolific but often flawed output.
Contrary to fears of chaos, allowing users to modify their software can create more stability. Users can craft a predictable, long-lasting environment tailored to their needs. This control protects them from disruptive, top-down redesigns pushed by a distant corporate office.
Inspired by architect Christopher Alexander, a designer's role shifts from building the final "house" to creating the "pattern language." This means designing a system of reusable patterns and principles that empowers users to construct their own solutions tailored to their unique needs.
For complex, one-time tasks like a code migration, don't just ask AI to write a script. Instead, have it build a disposable tool—a "jig" or "command center”—that visualizes the process and guides you through each step. This provides more control and understanding than a black-box script.
Instead of viewing AI collaboration as a manager delegating tasks, adopt the "surgeon" model. The human expert performs the critical, hands-on work while AI assistants handle prep (briefings, drafts) and auxiliary tasks. This keeps the expert in a state of flow and focused on their unique skills.
For data-heavy queries like financial projections, AI responses should transcend static text. The ideal output is an interactive visualization, such as a chart or graph, that the user can directly manipulate. This empowers them to explore scenarios and gain a deeper understanding of the data.