Try-Catch-Finally Self-Management / try-catch-finally 自我管理

Try-catch-finally self-management is [[LiuKefan|刘可凡]]’s programming-metaphor method in EP119 对话刘可凡：用 try-catch-finally，给独立做产品的内耗写个处理流程 🐛 for handling internal friction while doing independent product or business work.

The try step lowers the start threshold: do a minimal action before waiting for perfect confidence, complete information, or emotional calm. The catch step translates anxiety into a concrete, testable problem through a Falsifiable Product Hypothesis / 可证伪产品假设. The finally step defines a non-negotiable closure rule, such as stopping work at a fixed time, so the builder does not turn freedom into endless optimization and burnout.

The concept connects AI Engineering Thinking to self-management. Liu is not using code language as decoration; he is applying engineering habits to life: split the process, make error handling explicit, distinguish starting criteria from completion criteria, and protect the system’s long-term runtime through predictable cleanup.

Key Claims

• Start criteria can be much lower than completion criteria; the first action should be small enough to defeat avoidance.
• Emotional uncertainty becomes more tractable when converted into an observed problem or hypothesis.
• A product, content, or business experiment should define what evidence would change the builder’s mind.
• Failure is cleaner when attached to a specific hypothesis rather than to the person’s whole identity.
• Stopping is part of execution. Without a finally step, independent workers can keep optimizing until the next day’s capacity is damaged.
• The method complements AI Use Pacing because AI can make it even easier to start too many tasks or continue optimizing past a healthy boundary.

Connections

• 刘可凡 / Liu Kefan - source person for the method.
• Falsifiable Product Hypothesis / 可证伪产品假设, Validated Learning, and Fast Product Validation - the catch and product-validation layer.
• Founder Work Boundaries, Workplace Pacing, and AI Use Pacing - the finally and sustainable-rhythm layer.
• One-Person Company, Self-Directed Work, and Founder Psychology Operational Risk - contexts where self-management becomes operating risk.
• AI Engineering Thinking - engineering decomposition applied beyond software.