CS 369: Theory in the Age of LLMs

1-3 units · Letter or Credit/No Credit

What is mathematics for? What is the value of understanding without discovery? What do we want the mathematical and scientific communities to look like in 5 years, or 50? This course is a forum for discussing these questions. Roughly half of our meetings will be technical. We will read and present recent research-level results in mathematics, theoretical computer science, and theoretical physics that were generated (or substantially assisted) by LLMs. The other half of the meetings will be discussions, set against the philosophy of science canon from Aristotle and Plato to the present. Readings include selections from Aristotle (Posterior Analytics, Nicomachean Ethics), Plato (Phaedrus), Descartes (Rules for the Direction of the Mind), Leibniz (Meditations on Knowledge, Truth, and Ideas), Hume (An Enquiry Concerning Human Understanding), Poincare ("Mathematical Creation"), Hardy (A Mathematician's Apology), Kuhn (The Structure of Scientific Revolutions), Lakatos (Proofs and Refutations), Jaffe & Quinn ("Theoretical Mathematics"), Thurston ("On Proof and Progress in Mathematics"), and a selection of the current debate among mathematicians and philosophers about AI-generated research. This course is intended for advanced students in mathematics, theoretical computer science, and theoretical physics, who are comfortable reading papers. Students will be expected to present one technical result and lead one of the 'discussion' sessions.

Offered in Winter 2027 at Stanford University.

Winter 2027 sections

More CS courses

All CS courses · All departments