This is where I am hoping to put some write-ups of mathematical objects I find interesting. **Posts**: - [Unimodular random graphs and quantum percolation](https://functor.network/user/3440/entry/1842), where I write about the 2013 paper [**Mean quantum percolation**](https://arxiv.org/abs/1308.3755) by Charles Bordenave, Arnab Sen, and Bálint Virág. - In life, some things are *universal* and some things are *local*. Take as a case study the difference between continuum limits of percolation interfaces (conjecturally *universal*) and $p_c$ (known to be *local* by Hutchcroft and Easo's proof of the Schramm locality conjecture). Spectral measure is local! - This is written (for better or worse) in the style of a traditional mathematics note, but I like Goldbring, Seward, and Tucker-Drob's proof that treeable-by-finite groups are treeable. (This fact is also an immediate consequence of a powerful theorem of Hjorth that a group is treeable if and only if it is measure-equivalent (ME) to a free group.) I wrote it up [[Treeable_by_Finite_Groups.pdf|here]]. **Topics I plan to write about:** - Write a blogpost on my own paper for once. Hyperfiniteness in percolation: explain that lots of mass-transport-flavoured results of the 90s in percolation theory papers are about hyperfiniteness in disguise. Explain the main insight of [heavy repulsion of clusters in Bernoulli(p) percolation](https://arxiv.org/abs/2509.10631) is just using the language of hyperfiniteness. - Introduce the many definitions of amenability for infinite, locally finite, connected graphs. - Why does $\sqrt{n}$ appear as a factor in the Central Limit Theorem? Two perspectives: the analytic view and the discrete probabilist's view. Goal: Give a soft introduction to scaling limits (correctly scaled unbiased random walks converge to Brownian motion, correctly scaled critical Bienaymé trees converge to the Brownian continuum random tree) and introduce the notion of Gromov-Hausdorff convergence. - The cluster equivalence relation (Gaboriau, 2004, JKL in the non-measured context) and the definition of unimodular random graph in terms of cBers. - Stop using geodesic rays to define ends! The ultrafilter approach vs. the combinatorial approach. We should all be using Stone duality. - What does supercritical percolation on nonamenable graphs look like? A tour through some known results. - Explain that Poisson is just Bernoulli in the continuum regime, as deeply as possible. (E.g. construct the Poisson point process by partitioning things into infinitesimal finite-measure pieces and using Bernoulli coin flips.) Give examples from measured group theory, too (Abért-Weiss becomes Abért-Mellick and so on) and explain the modern point process approach to defining the cost of locally compact, second countable groups (c.f. the cross-section equivalence relation approach).