# Stanford CS547 HCI Seminar | Autumn 2026 | One Developer, Two Decades, and 30 Million Users Later

## Executive summary

This seminar traces the 20-year journey of Python Tutor, a browser-based code visualization tool, from a grad student side project to a widely used academic resource. The discussion covers the tool's design philosophy—'software as craft'—and its resilience through different eras of technology. The speaker concludes by addressing the current 'Disruption' era, where Generative AI challenges the relevance of step-by-step learning, leading to a focus on how human understanding remains critical even when AI writes the code.

## Key takeaways

- The Lifecycle of Software: The project evolved through three acts: Inception (as a side project), Acceleration (a decade of growth), and Disruption (the current era defined by Generative AI).
- Design Philosophy: Craft vs. Metrics: The tool's longevity is attributed to a 'craft mindset'—focusing on making something useful and good, rather than solely optimizing for metrics, funding, or corporate growth.
- AI and the Future of Learning: The rise of AI has led to a flood of automated, low-quality content. The speaker argues that the core value of visualization is not teaching simple programs, but enabling users to understand, debug, and build upon complex code written by others (including AI).

## Technical details

- Tool Functionality: Python Tutor runs code and automatically draws diagrams of execution, allowing users to step through code and observe how data structures change, which is crucial for building mental models.
- Language Support: While named for Python, the tool supports foundational concepts across multiple languages, including Python, Java, and C++.
- Development Model: The project began as a single developer side project, contrasting with traditional, grant-funded academic research which often results in short-lived prototypes.

## Practical implications

- The importance of focusing on fundamental understanding (computational thinking) over mere output or rapid development.
- The necessity of maintaining a clear, user-centric vision when scaling a project, resisting the pressure to conform to corporate or academic metrics.
- The emerging role of human oversight and 'product management' in supervising AI-driven development processes.

## Topics

Code Visualization, Computer Science Education, Generative AI, Software Design, Learning Theory, Python Tutor, CS547 Playlist

Source: https://www.youtube.com/watch?v=GPk0nNuc8S0
