The Overhang

About

This workshop brought together number theorists, along with researchers and technologists from industry, to explore new models of collaborative research. Participants worked with one another, with frontier AI models, and with multi-agent systems. The aim was to see which new directions in number theory could be advanced under expert human direction.

“The Overhang” is meant to capture the idea that new models, and new ways of using them, make it possible to make progress relatively quickly on open problems if people try working on them in new ways.

Many mathematicians are already finding uses for AI in their research, but adoption has been uneven. These two days gave experts time to experiment, collaborate across institutions, and share methods and problems.

This effort complements work to update community and professional norms around tool use. What we want to do is bound up with what we are able to do, and this event aims to keep the latter up to date and empirically grounded.

Schedule

Organizers

• Alexa McLainUC Irvine
• Andrew SutherlandMIT

Support

This workshop is part of a DARPA-funded effort under the Exponentiating Mathematics (expMath) program to evaluate how new AI tools affect mathematical research. Additional support came from Paradigm (Dan Robinson), Gil & Co., OpenAI, and Anthropic.