Mathematics is Facing A Crisis of Indigestion
Mathematics is moving into the 21st century. For hundreds of years, the pace of math research stayed slow. A researcher finds a result, peers review it, and it eventually enters a textbook. This process is steady and predictable.
AI changed everything. It generates proofs and solves problems at high speeds. This creates a massive pile of new content. Human reviewers cannot keep up. This creates what Terence Tao calls proof indigestion. The academic system is hitting a traffic jam.
AI is like a fast car. Current academic journals are like narrow stone paths built for horse carriages. You cannot drive a fast car on a narrow path without causing a wreck. Upgrading the car does not help if the road is broken. We must build new roads for AI.
Terence Tao is building these new roads through SAIR competitions. He creates separate tracks for humans and AI.
One competition is the Distillation Challenge. Researchers use a massive dataset of 22 million algebra problems. Large AI models solve these easily but cost too much money. The challenge asks participants to write a one-page cheat sheet for small, cheap models. The goal is to transfer knowledge from big models to small ones. The best cheat sheets have already raised accuracy from 50% to 80%.
Another competition is the Inverse Galois Problem. Think of this as a digital egg hunt. There are 160,000 different mathematical properties, or colors. Participants search for specific polynomials that match these colors. If you find a rare color that no one else has, you score points. This turns math into an experimental science.
These competitions do not replace mathematicians. They create new ways to work. They separate the human walking path from the AI highway.
This model can work for all science. If a field has large datasets and verifiable tasks, it can use this method. Math is just the starting point.
Optional learning community: https://t.me/GyaanSetuAi