I took a break from AI and returned with a clearer view of my problematic habits. After weeks off, I discovered that excessive agent usage created unnecessary complexity—eleven Cmu tabs with paused jobs, unread Claude badges, and countless half-finished projects. Each was a symptom of intellectual laziness disguised as productivity. I've been using AI as a sympathetic mirror rather than a true collaborator, leading to a cycle of dependency where I seek to optimize tool usage instead of focusing on meaningful work. The 'human is the loop' realization means I must intentionally step outside AI whenever possible, leaving the agent tagged only for occasional, thoughtful intervention.

The core issue stems from treating AI as a crutch that makes me feel smarter and more capable, when in reality it erodes confidence and curiosity. I stopped relying on AI for genuine problem-solving and started setting specific constraints to force direct engagement with tasks. This approach freed me to do human-centered work like writing and reflecting. The goal is to be intentional about when to use AI—and crucially, when to leave it behind—so that technology serves me rather than trapping me in an endless feedback loop.