Aiden Ghim


We need a mathematics community for Korean students

Nearly three years ago, I encountered an online math forum run by a math magazine. It was called Polymath 1; hundreds of students were active. The platform was modeled after the online collaboration project of the same name used by renowned mathematicians2. Problems were released weekly or monthly by postdocs and professors. There was a dedicated channel for student-posed questions, one of which I ended up answering! I thought I met an amazing community I wanted to be part of. And just months later, Polymath abruptly shut down. The math magazine went out of business. We students had to reckon with the sunset, but that wasn’t all that we lost. Connections with passionate mentors who provided problems and feedback ended without warning. A space for mathematical collaboration was no more, and most of us had no alternatives.

Online communities have a unique value proposition in mathematics. For students not in a university or research environment, collaborating on interesting pieces of math can be nearly impossible in person. Online communities provide valuable experience of written collaboration with peers, feedback from mentors, and discovery of intriguing math topics. They have an outsized impact on curious middle and high school students, because many do not have access to a social kind of mathematics nor engagement with math content beyond the standard curriculum.

Years after the sundown, Korean students still lack any such online space. In ordinary times, this would have been simply unfortunate. But we live in no ordinary times, not when A.I. is upending the value of traditional math education. Korean students lack a mathematics community, and we need to build one. We need a community built for these turbulent times in math, and the call is urgent.

Korea faces a unique problem in the face of A.I. wildly changing math. First, its public math curriculum is a grotesque, highly competitive and difficult exercise in algebra and calculus. Korean math education is incredibly intensive and requires very difficult thinking. However, it is limited in scope to mastery of very specific algebra/calculus problem-solving. It rewards speed, accuracy, and technique more than understanding or broad exploration. Granted, this is true in nearly all public math education, but perhaps to an unreasonable degree in Korea. Second, curious middle and high school students have no accessible means of collaborating on pure math. Students in the most competitive public schools are devoted not to the shared joy of doing math together but rather performance in the aforementioned school curriculum. No summer programs, math circles, or online forums exist with meaningful reach. Engaging in higher math during high school happens almost exclusively in a handful of “national gifted schools”, each with an exclusive culture and admitting a diminutive fraction of applicants. Third, olympiad-style competition math is taught with little room for exploration, much less collaboration. It is almost always taught in sprawling private tutoring centers to prepare for entry into the national gifted schools, and even then, only in very few locations in the Seoul metropolitan area. Competition math can be a primer for curiosity beyond school math, but not when very specific entrance tests with unreasonable competition define the students’ motivation for learning.

But the most valuable lessons from math come from writing proofs, posing one’s own questions, and connecting questions and results to form a theory. The joy of mathematical engagement is multiplied when collaborating with peers. With recent developments in A.I. proofwriting, what matters in math is changing still. Proving theorems alone may not determine a mathematical thinker’s capability in the future; asking the right questions and organizing results into a coherent theory skyrockets in value.

Some would argue that middle/high school students need not concern themselves with this kind of math. Instead, they say, students should focus on acquiring the bare minimum skills through standardized lectures. This is only half correct, and the incorrect half matters for curiosity-driven students. There are students who feel comfortable with basic skills, those who want an introduction to what mathematicians care about. There are students who would “reinvent” whole theories themselves (or with partners and mentors), not grind convoluted calculus for exam scores. It is a clear failure of education when motivated students can find no meaningful way to explore their passion, even outside the public education system. Furthermore, no access to the theory-building part of math simply leaves students unprepared for a changing field of mathematical thinking, even though they very much can be with appropriate stimuli.

The bluntly stated reality is this: Korean students cannot engage in the part of math that matters. Very little proofwriting is accessible. Virtually no experience of posing questions and rebuilding theories is possible, and collaborative math is out of reach for many. But a well-designed online community can change this.

The proposed online community project is an evolution of Polymath, not a continuation. Korean middle and high school students answer various problems written by mathematicians. The problems can either be self-contained or part of a series on a theory, and each has its advantages. Self-contained problems are more readily approachable for newcomers to the platform. These can serve as daily or weekly challenges; similar discussion already happens in worldwide communities (such as Art of Problem Solving).

But a math community should go beyond isolated problems. Students should not just answer questions, but rather propose conjectures and directions for exploration themselves. In fact, I propose an experiment in which students rebuild meaningful branches of math.

High school students rebuilding math theories can feel impossible, but it is not. I spent a transformative summer in 2026 as part of Hampshire College Summer Studies in Mathematics (HCSSiM), one of the most prestigious and oldest summer math programs in the U.S 3. During six weeks at HCSSiM, we explored college-level set theory, group theory, graph theory, and number theory over three weeks. In the three weeks that followed, students were split into smaller groups to rebuild pointed introductions to numerous branches of math including nonstandard analysis, Galois theory, Tarski’s paradox with Axiom of Choice, knot theory, modal logic, and many more. Nearly all classes were driven by students; we had to choose our own direction, face our own dead ends, and rebuild a consistent theory ourselves. The staff never threw us lecture content, but rather initial stimuli and close feedback.

This incredible experience has been happening on the quirky rural campus of Hampshire for over 51 years. HCSSiM’s model shows that theory-building for curiosity-driven high school learners is possible and invaluable. This kind of mathematical growth is transferable, and can reach more students via an online community. The platform can host monthslong series on rebuilding theories. Instead of flooding users with problem sets, mentors can select meaningful theorems to get students started. Once students prove results on the community site with open collaboration, they can make conjectures and choose which questions to explore next. Over the course of weeks and months, students experience mathematicians’ critical thinking, with collaboration at every step.

Theory-building topics on the community site can develop a high entry barrier. The amount of established results can discourage newcomers from joining a topic mid-development. While this is a drawback, it can be mitigated by extensive documentation of results. Careful summary can also be a task following each advance on the theory.

The final unique element in the online space is an introduction to A.I. mathematics. A.I. increasingly demonstrates potential in proofwriting and counterexample discovery. As math embraces the future of machine-generated proofs, students need to learn what the frontier looks like. Without doubt, mathematics without A.I. is crucial for middle/high school students’ learning. Keeping them in the dark about A.I.’s potential, however, is no longer an option. The math community site will maintain dedicated problems and challenges for A.I.-aided math. These can focus more on handpicked problems A.I. can meaningfully tackle with students steering the direction. Even some open problems can be included, including conjectures left untouched by researchers from lack of attention.

Mathematics education needs to change. An online community for Korean students has incredible leverage in the age of A.I. Theory-building by curious students has never been this crucial, and written human collaboration is more valuable than ever. We can provide a future generation of mathematicians the experience of working with A.I. firsthand. The target pool is already there, as demonstrated by now-defunct Polymath. There has never been a more opportune time to push math education forward.

  1. Polymath by Math Donga is an online math forum that launched in 2017. The platform partnered with Korean Mathematics Society and other young mathematicians to release monthly problems, and allowed users to post their own problems of interest. Math Donga, the monthly math magazine targeting middle and high school students, was discontinued by the parent company (Donga Science) in 2024. Consequently, the Polymath platform shut down in January 2024. ↩

  2. The Polymath Project (https://polymathprojects.org) is a series of “massively collaborative mathematical research projects.” Renowned mathematicians, including Terence Tao serve, as administrator. ↩

  3. Hampshire College Summer Studies in Mathematics (HCSSiM) is a summer math program founded by the late David C. Kelly in 1971. Alumni include tenured Harvard physics professor Lisa Randall and AMD CEO Lisa Su, among many others. It is considered one of the most prestigious math programs for high school students, accepting around 51 students each year from 1000+ applicants. ↩