Terence Tao Revives Old Java Applets and Builds New Tools with AI Coding Agents

Old and new apps, via modern coding agents by Terry Tao

Terence Tao Revives Old Java Applets and Builds New Tools with AI Coding Agents

I recently used modern AI coding agents to port my old Java applets from 1999 into functional JavaScript, bringing decades-old visualizations back to life with minimal bugs. The process was so efficient that I also resurrected an abandoned special relativity project and created a new tool for the Gilbreath conjecture, proving that AI is a powerful partner for rapid mathematical prototyping.

After a couple hours of 'vibe coding' with an AI agent, I was finally able to generate an applet that matched the vision I had back in 1999.
  1. recursivedoubts

    Building visualizations with LLMs has been a major boost for my CS classes:

    https://htmx.org/essays/universities-and-ai/#demos-visualiza...

    Many visualizations that I have always wanted but just didn't have the time to build, I now have.

    To give an example, I wanted a simplified 8-bit computer to complement the 16-bit teaching computer I use and designed this in a few days with the help of claude:

    https://bdp.cs.montana.edu/

  2. luciana1u

    Terry Tao using coding agents to build apps means we're one step away from a Fields Medalist asking an LLM why his Docker container won't start, just like the rest of us.

  3. semiquaver

    There is infinite latent demand for software, most especially outside the traditionally software-focused spaces. If LLMs stopped improving today it would take us 10 years to catch up to the new software-writing abilities that have become available. This is a great illustration of that fact.

  4. wffurr

    Nice balanced perspective there at the end:

    "as such [LLM-coded interactive] supplements are not mission-critical to the core of the paper, I again feel that the downside risk of using guided interaction with LLM agents to generate such visualizations is acceptable."

    It's a tool. Good for some things but not others and generally not to be trusted.

  5. luciana1u

    Terry Tao using coding agents feels like watching a Michelin-starred chef discover microwave dinners and get genuinely excited about them.

  6. apignotti

    Running legacy educational Java applets, especially around math and physics, has been a longstanding popular use case of our CheerpJ Applet Runner extension, running Java bytecode in the browser via WebAssembly.

    I am not sure how to feel about agents solving the problem via proper modernization. It's certainly positive that students will be able to interact with this content in a modern and more accessible way, but the educational use case for our product, although not commercially important, has always been a source of pride.

    https://chromewebstore.google.com/detail/cheerpj-applet-runn...

  7. alansaber

    I always enjoy these "domain expert has fun using AI to do something in their domain" articles. But it's always a hobby project, never something serious.

  8. bradfitz

    This post inspired me to have Claude port my 30 year old high school German Java applet game to Javascript, complete with a faked git history:

    https://github.com/bradfitz/koffer#der-verloren-koffe

    Play online at https://bradfitz.github.io/koffer/js/

    So neat seeing ~30 year old code come back alive.

More from this day

2026-07-12