AI tools are already powerful enough for most problems. The real challenge is a psychological one: training users to recognize that nearly any problem they face, from planning a house move to tracking promises, can be framed as a task for an AI to solve.
Unlike adults conditioned by decades of clunky software, children approach AI with no preconceived notions of what is possible. This "beginner's mind" allows them to unlock more creative and powerful use cases because they've never learned what not to ask for.
To overcome the "imagination gap," describe your roles and challenges to an AI and ask it to interview you. Based on your answers, the AI can generate a personalized list of problems it can solve, acting as a discovery engine for its own capabilities.
Instead of making users watch a loading screen, design AI products that encourage them to move on. Position the wait time as the AI working independently in the background. This builds trust, shifts the interaction from synchronous to asynchronous, and frees the user's creative energy.
When working on complex or unconventional problems, an AI might initially claim the task is impossible. Prefacing your prompt with a phrase like "I know this is possible" can give the model more confidence to persist and attempt more creative, fringe solutions instead of giving up.
Your email inbox contains a comprehensive record of your online purchases. By giving an AI like Claude access, it can parse receipts and build a structured inventory of items you own, like furniture or clothing, eliminating tedious manual data entry for tasks like planning a move.
Don't just ask AI to perform one step in a tedious process. Constantly challenge yourself to delegate the entire goal. Instead of inputting furniture dimensions, ask the AI to find them in your email. This shifts your effort from doing the work to defining the system that does the work.
Instead of monitoring a screen, you can use a cheap, programmable IoT device with a button to externalize AI interactions. The device can light up or chime when an AI like Claude needs approval for a file operation or other task, which you can grant with a physical press.
Use the more powerful Opus model when you don't fully understand the problem you're trying to solve. For well-scoped, clearly defined tasks, the faster and cheaper Sonnet model is often sufficient and highly effective, as the key difference is Opus's ability to reinterpret vague requests.