A Bangladeshi student used ChatGPT to build a Moon and Mars survival game for NASA's hackathon
A team leader describes three months of AI-assisted work on AstroBase, a 3D educational game, and asks OpenAI for student support programs.
The short version
- AstroBase is an interactive 3D survival game that teaches young learners about living on the Moon and Mars.
- The author led Team 404: Earth Not Found in the NASA Space Apps Challenge 2026.
- ChatGPT was used for about three months, alongside Claude, GitHub Copilot and Ollama.
- The post asks about student access, ChatGPT Plus for selected projects, and API credits.
- The post does not say how the project placed in the challenge.
Why it matters
The post is a first-person look at how a student with limited resources can use AI assistants to finish a sizeable software project. It also shows a gap: tools that help beginners are often paid, and builders are starting to ask for educational access. Note that this is a closed model story, not an open-weights one.
A survival game set on the Moon and Mars
A student developer from Bangladesh wrote a post on the OpenAI community forum about a project built for the NASA Space Apps Challenge 2026. The author led a team called Team 404: Earth Not Found. As community.openai.com reported through that post, the project is called AstroBase: Moon & Mars Survival Simulator.
AstroBase is an interactive 3D educational game. Its aim is to help young learners understand what makes living on the Moon and Mars hard. The author shared a live demo, a code repository on GitHub, a demo video and the team's page on the NASA challenge site.
The post is written as a thank-you and a request, not a technical report. It gives no description of specific levels, survival rules or how the game plays, so the links are the best way to judge the result.
Where the AI assistant fit in
The author says ChatGPT was a development and learning aid for roughly three months. The list of tasks is wide. It covers building the game's interface with React (a popular toolkit for web interfaces) and Three.js (a library for drawing 3D scenes in a browser).
It also covers planning how the screens should look and work, debugging both the front end and the back end, and designing the game mechanics and the logic behind the survival simulation. The author also mentions help with planning an AI companion feature and its connection to an API, which is a way for one program to call another service.
Beyond code, ChatGPT helped write documentation and project descriptions and prepare the presentation. The author adds that it helped them understand technical problems and weigh possible fixes, which is a use closer to tutoring than to writing code.
ChatGPT was not the only tool. The author also names Claude, GitHub Copilot and Ollama as part of the workflow. Ollama is a program generally used to run open models on your own computer. The post does not say how work was split among these tools, or how much of the finished code an AI wrote.
The ask: help for student builders
The second half of the post is a request. The author says AI tools make software work and technical learning easier to reach for students with limited resources. They would like to see more openings for student and independent builders.
The examples they list are:
- Student or educational access programs
- Free ChatGPT Plus access for selected student projects
- API credits for educational and research work
- Chances to showcase promising student-built projects
The author is careful about expectations. They acknowledge that such programs may have eligibility rules and that posting on the forum does not guarantee anything. They say they simply wanted to share the experience and ask whether OpenAI runs anything relevant.
What we know, and what we do not
The post is a self-report. It has no outside verification of how much the AI contributed, no description of how the project did in the challenge, and no numbers about the game itself. The only figure is the rough three-month timeline.
Still, it is a useful example for readers who are curious about AI-assisted building. The tasks the author lists, such as debugging, planning and writing up a project, are ordinary jobs where an assistant can save a beginner time. The request for educational access also hints at a real tension. The strongest tools often cost money, and students are the people who can least afford them.
It is also worth noting what the story is not. ChatGPT is a closed model, so it is not an open-weights release that you can download and run yourself. The mention of Ollama is the only nod to that world.
What to watch
- Whether OpenAI responds on the forum with any student or educational access program.
- How AstroBase fares in the NASA Space Apps Challenge 2026, which the post does not say.
Try it
If you plan to enter a hackathon or build a portfolio project, keep a simple log of which parts you wrote and which an AI assistant drafted. It keeps you honest and gives you something concrete to say in interviews.
Sources and links
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.
Listen to the AI Wire headlines
Listen to the AI WireListen to our daily AI shows every morning: on your commute, in the shower, wherever you are
Listen to the daily showsFirst reported by community.openai.com · Read the original. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.
