Core team

• Fan Nie CS PhD at Stanford
• Zihao Wang Math PhD at Stanford
• Kaiyue Wen CS PhD at Stanford
• Garrett Wen Stats PhD at Yale
• Zhaozi Wang CS PhD at NYU
• Hai-Yang Wang Physics PhD at Caltech
• Denny Wu Stats Professor at Oxford
• Theodor Misiakiewicz Stats Professor at Yale
• Joan Bruna CS, DS & Math Professor at NYU
• James Zou BDS, CS & EE Professor at Stanford
• Andrea Montanari Math & Stats Professor at Stanford
• Percy Liang CS & Stats Professor at Stanford

Contributors

• Shuangping Li Stats Professor at Yale
• Puheng Li Stats PhD at Stanford
• Rui Sun Stats PhD at Stanford
• Renyuan Ma Stats PhD at Yale
• Sheng Xu Math Professor at UT Dallas
• Yiran Zhang CS PhD at Berkeley
• Hugo Latourelle-Vigeant Stats PhD at Yale
• Ruixiao Wang Stats PhD at Yale
• Linjun Zhang Stats Professor at Rutgers
• Yijun Dong Postdoc at ISTA
• Mingwei Yang MS&E PhD at Stanford
• Alberto Bietti Research Scientist at Flatiron
• Anupam Gupta CS Professor at NYU
• Tianhao Wang Data Science Professor at UCSD
• Theo McKenzie Stats Professor at Yale
• Shuzhe Zeng Physics PhD at Stanford
• Yanbei Chen Physics Professor at Caltech
• Alexander Philippov Physics Professor at Stanford
• Elias Roland Most Astrophysics Professor at Caltech
• Xiang Li Stats Postdoc at Penn

Advisory board

To be announced soon.

Contribute

After42 is a collaboration between AI and scientists. Scientists set the goals, and their research taste guides which questions AI proposes and develops toward them. We are extending After42 to more fields and would love to hear from mathematicians and scientists who want to bring their expertise and shape where research in their field goes next. We welcome collaborators from every field, for example mathematics, statistics, physics, biostatistics, theoretical computer science, cryptography, AI infrastructure, and empirical AI.

• Set a goal in your subfield. Write a goal for a subfield you know well, describing the understanding the field is after and the kinds of questions that would advance it. We build a research pack of the literature with you, and Agent42 proposes and develops questions toward that goal.
• Assess questions and share your taste. Judge whether questions proposed in your area are interesting and worth pursuing. Your assessments are how we measure interest, and your taste guides how Agent42 proposes, develops, and evaluates questions. Differences in taste among experts show which perspectives the system should learn from.
• Pursue a question. Work on any published question, on your own or with AI systems. If you make progress, find prior art, or spot an error, tell us and we will update the question card.
• Build the system. Help develop the agents, the evaluation tooling, the dataset releases, and this website.

If any of these interest you, please reach out.

Contact

We would love to hear from mathematicians and scientists who want to contribute their expertise or extend After42 to their own field. Please email niefan@stanford.edu. To report prior art or an error in a published question, open an issue on GitHub or email us.

Citing After42

Each published question carries its own BibTeX on the questions page. To cite the project as a whole, use the entry below.

@misc{problemgen2026,
  title        = {After42: Can AI Propose Interesting Research Questions?},
  author       = {{The After42 Team}},
  year         = {2026},
  howpublished = {\url{https://after42.org}}
}