Using AI as a tutor to do math with pen and paper

Nicholas Charriere describes a daily routine built around a camera-plus-AI tutor that guides 45 minutes of pen-and-paper math or physics practice each morning. He built a custom AI coach that generates exercises, watches work through a camera, and grades solutions, and reports that the regimen is restoring skills he felt rusty on and feels mentally rewarding. The core claim is that a simple visual-input AI plus active handwriting practice creates a reliable self-tutoring loop that motivates regular study and provides immediate feedback.

The setup’s specifics - camera capture of handwritten work, automated exercise generation, and automated grading - are presented as practical and repeatable, not just experimental. The post attracted substantial attention (noted view and engagement counts) and prompted requests for a code repository and comments from others who have tried or plan to try similar builds, signaling community interest in recreating or extending the approach. The account frames the combination of embodied problem solving (pen and paper) with AI assessment as an effective way to relearn technical material and sustain a disciplined study habit.