Old and new apps, via modern coding agents by Terry Tao
Renowned mathematician Terry Tao is now using AI coding agents to breathe new life into his decades-old Java applets and conjure up novel mathematical visualizations. His success in porting legacy code and realizing long-dormant project ideas, with minimal effort and surprising accuracy, has sparked widespread fascination on Hacker News. The community is buzzing about the practical implications of AI for democratizing software creation, even for the most brilliant minds, and debating the precise boundaries of its reliability as a coding assistant.
The Lowdown
Terence Tao, a Fields Medal-winning mathematician, has turned to modern AI coding agents to revitalize his early educational applets and create entirely new visualization tools. This initiative highlights the transformative potential of AI in enabling domain experts to overcome coding barriers and bring complex concepts to life.
- Legacy Applet Revival: Tao successfully migrated a suite of Java 1.0 applets from 1999, designed for complex analysis and linear algebra, to modern JavaScript using an AI agent. This process, which took only hours, not only made the applets functional again but also introduced graphical upgrades and identified pre-existing bugs in his original code.
- New Visualizations from Old Ideas: Inspired by this success, Tao leveraged AI to build two new applets. One, a "Minkowski space Inkscape" for special relativity, was a project he had abandoned in 1999 due to coding complexity. The other is a visualization tool for the Gilbreath conjecture, created as a supplement to a recent paper.
- Balanced Perspective on AI: While acknowledging that LLM-generated code can contain bugs, Tao views the risk as acceptable for these non-mission-critical visual aids and supplementary materials. He emphasizes that the ease and speed of development make these tools viable, even if not formally verified.
Tao's experience demonstrates how AI coding agents can significantly reduce the activation energy for programming tasks, empowering non-specialists to develop custom software and interactive educational materials that were previously too time-consuming or technically challenging.
The Gossip
Chef's Coding Curiousity
Many commenters expressed a mixture of humor and awe at the image of Terry Tao, a figure of such mathematical renown, employing AI for coding tasks. The sentiment often compared it to a 'Michelin-starred chef discovering microwave dinners,' underscoring the surprising yet practical utility AI offers even to highly skilled individuals. This led to reflections on how AI is leveling the playing field for mundane programming tasks, freeing up experts for deeper work, or just making life easier for everyone.
Tool Talk and Trust
A significant thread revolved around the reliability and appropriate application of AI as a coding tool. Commenters debated Tao's 'balanced perspective' – that AI is good for some things but not others and 'generally not to be trusted' for mission-critical tasks. This sparked discussions on whether AI is merely a 'statistical gradient descent token vomiter' or if its 'trust' factor is rapidly evolving. The consensus leaned towards AI being a powerful, time-saving tool, but one that still requires human oversight, especially for sensitive or complex projects.
Enabling Expert Endeavors
A recurring theme was how AI empowers domain experts, particularly those outside traditional software development, to create visualizations and 'nice-to-have' tools that were previously out of reach due to time or technical constraints. Commenters highlighted the 'infinite latent demand for software' in specialized fields and how AI dramatically lowers the 'activation energy' for bringing these 'one-of-these-days' projects to fruition. This ability to easily build supplementary, non-critical tools was seen as a significant democratizing force for software creation.