Organisers: Magnus Carlson (Gävle), Minhyong Kim (Maxwell Institute Edinburgh and Kings College Cambridge), Padmavathi Srinivasan (Boston), Jakob Stix (Frankfurt)
We will get together with the goal of understanding OpenAI's claimed proof of Grothendieck's section conjecture on fundamental groups of hyperbolic curves over p-adic fields. (As far as we know, it has not even been Lean verified.)
Schedule
20 October 14:00 (UK time)
Organisational meeting and background talk
Speaker: Benjamin Collas (RIMS)
Title: On the p-adic section conjecture
Abstract: Anabelian arithmetic geometry, whose central questions include Grothendieck’s section conjecture, stands at the intersection of two seemingly opposed forces: the flexibility of geometric étale fundamental groups and the rigidity of Galois groups.
On the birational side, the passage from the étale fundamental group to the absolute Galois group of the function field brings birational geometry and valuation theory into play. The work of Koenigsmann, Pop, and Stix (2003-12) illustrates how decomposition groups associated with valuations can be exploited to recover arithmetic information, in particular rational points.
On the geometric side, the deformation of smooth curves towards singular ones, together with the homotopy theory of Berkovich spaces, offers a complementary approach to Grothendieck-type conjectures. Following the work of Lepage, Mochizuki and Tsujimura via Tamagawa's resolution of nonsingularities (2004-23), this perspective leads to establishing the Grothendieck conjecture over p-adic fields in its absolute form.
The goal of this investigatory overview is to examine the gap between these two perspectives in a way that it provides better clarity on the claims in the related OpenAI notes.