How an astrophysicist uses Codex to help simulate black holes

Overview

University of Arizona astrophysicist Chi-kwan Chan uses OpenAI's Codex to write, debug, and refine the scientific code behind black hole simulations. Chan works with the Event Horizon Telescope collaboration, the international team that produced the first image of a black hole in 2019 and is now gathering data to make the first time-lapse video of a supermassive black hole. The simulations model the superheated plasma swirling near black holes and let scientists test Einstein's general theory of relativity against real telescope observations.

Key Takeaways

Chi-kwan Chan, an associate astronomer at the University of Arizona's Steward Observatory, uses the AI coding tool Codex to build and improve the software that runs his black hole simulations.
Chan is a core member of the Event Horizon Telescope collaboration and helped create the project's computational and data-processing infrastructure.
The collaboration's next goal is the first video of a supermassive black hole, focused on the giant black hole at the center of the M87 galaxy.
The hardest part of the work is modeling plasma, the superheated charged gas around a black hole, which forces researchers to track individual particles where collisions are rare.
Black holes serve as natural laboratories for testing Einstein's general theory of relativity, and simulations let scientists compare the theory's predictions with telescope data.
OpenAI presents the story as an example of AI coding assistants supporting frontier science rather than routine software development.

Stats & Key Facts

#1 first-ever black hole image released by the Event Horizon Telescope in 2019.
#Trillions of electrons and ions are tracked in the plasma simulations as they spiral around a black hole.
#1 planned first-ever time-lapse video of a supermassive black hole, targeting the M87 galaxy.
#The M87 black hole shoots a plasma jet across a scale of roughly 6,000 light years.
#Since 2020 Chan has served as Secretary of the Event Horizon Telescope Science Council.

Who Chi-kwan Chan is and his role at the Event Horizon Telescope

The profile centers on a computational astrophysicist with deep roots in black hole imaging.

Chi-kwan Chan is an associate astronomer at Steward Observatory, part of the University of Arizona, where his research focuses on computational and high-energy astrophysics. His specialty is modeling and simulating black holes using fast scientific code.

Chan is not a casual participant in the Event Horizon Telescope project. He helped build the collaboration's computational and data-processing infrastructure, leads its Software and Data Compatibility Working Group, and has served as Secretary of the EHT Science Council since 2020. His background includes pioneering the use of GPUs to speed up black hole modeling.

From the 2019 black hole image to a first-ever black hole video

The team behind a historic photo is now chasing a moving picture.

The Event Horizon Telescope is the international team that produced the first direct image of a black hole in 2019, a result that confirmed predictions about how matter behaves at the edge of these objects. The image required a worldwide network of telescopes working together as one giant instrument.

The next milestone is far more demanding. The collaboration is gathering observations to create the first time-lapse video of a supermassive black hole, focused on the giant one at the center of the M87 galaxy. A video would show how the glowing material around the black hole shifts and flows over time, which is a much larger data and computing task than a single still image.

Why modeling plasma near a black hole is so hard

The core scientific challenge is the superheated gas circling the black hole.

›Plasma is the superheated, electrically charged gas swirling around a black hole.
›Near a supermassive black hole the gas becomes so hot and thin that particles rarely collide.
›When collisions are rare, researchers cannot treat the gas as a smooth fluid and must instead follow individual particles.
›The simulations track trillions of electrons and ions as they corkscrew at high speed around the black hole.
›Handling that many particles at this scale demands fast, carefully written computer code.

Where Codex fits into the simulation workflow

The AI coding tool handles the specialized software work behind the science.

Chan uses Codex to help write, debug, and refine the specialized scientific code that runs his simulations. The goal is to reduce the manual coding effort involved in research operating at this scale, freeing more time for the physics itself.

This is a different use of an AI coding assistant than everyday app building. The code here solves the equations of general relativity, electromagnetism, and fluid dynamics, and it has to run efficiently enough to track an enormous number of particles. Faster, cleaner code means researchers can run more detailed simulations.

Black holes as a testing ground for Einstein's general relativity

The simulations exist to check a century-old theory against real data.

Black holes rank among the best natural laboratories for testing Einstein's general theory of relativity, which describes how gravity bends space and time. Gravity is at its most intense near a black hole, so any cracks in the theory would be easiest to spot there.

Simulations are the bridge between theory and observation. Researchers use them to predict what the telescope should see if general relativity holds, then compare those predictions against the actual data the Event Horizon Telescope collects. The closer the match, the stronger the confidence in the theory.

What the story signals for AI in scientific research

Here is plain-language analysis of why OpenAI is highlighting this work.

OpenAI frames this profile as an example of AI coding assistants supporting frontier science rather than routine software work. The message is that tools built for general programming are starting to handle the demanding, specialized code that research depends on.

For non-technical readers, the takeaway is straightforward. When a researcher spends less time fighting with code, more time goes to the actual discovery. If AI tools reliably help write correct scientific software, they shorten the gap between an idea and a working simulation across many fields, not only astrophysics.

Frequently Asked Questions

What is Codex and how does Chi-kwan Chan use it?

Codex is OpenAI's AI coding tool that helps people write and fix software. Chan uses it to write, debug, and refine the specialized scientific code that runs his black hole simulations, reducing the manual coding work involved.

What is the Event Horizon Telescope?

The Event Horizon Telescope is an international collaboration that links radio telescopes around the world to act as one giant instrument. It produced the first direct image of a black hole in 2019.

Why is the team trying to make a video of a black hole?

A video would show how the glowing plasma around a supermassive black hole moves and changes over time, going beyond the single still image from 2019. The team is focused on the black hole at the center of the M87 galaxy.

Why are black holes used to test Einstein's theory?

Gravity is most intense near a black hole, so general relativity faces its hardest test there. Researchers compare simulation predictions based on the theory against real telescope data to check whether the theory holds.

Why is simulating the plasma around a black hole so difficult?

Near a supermassive black hole the gas is so hot and thin that particles rarely collide, so researchers must track individual particles instead of treating the gas as a smooth fluid. The simulations follow trillions of electrons and ions, which demands fast, carefully written code.

Chan's work shows AI coding tools moving into demanding scientific software, where cleaner code means more time for the physics. The next test will be whether those tools help the Event Horizon Telescope deliver the first video of a supermassive black hole.