Here I’m listing much of the media I’ve read/watched/listed to/etc., and learned something from.

  • Research popularization: Book-Length Topology Re-proof, by Kevin Harnett about a book by Ray, Behrens, Kalmar, Kim and Powell. In 1981 Michael Freedman proved a very important theorem in topology. But it was written so badly no one can follow it unless Freedman personally walks them through it. So five grad students rewrote the argument in full detail as a 500 page book.

  • Research Talk: [Unique Games Conjecture] (https://youtu.be/-02d-oCd_xI), by Ryan O’Donnell at the Institute for Advanced Study. One of the openest conjectures in computational complexity.

  • Research Talk: Implicit Regularization from Gradient Descent, by Wei Hi at the Simons Institute. Gradient Descent works well, even when “find an optimal solution” would fail. Why? Gradient Descent has implicit regularization: it select for something on top of low loss, which generalizes well.

  • Research Talk: Fine-grained worst-case to average-case reductions, by Andrea Lincoln at the Simons Institute. Andrea is a friend from undergrad. Some problems are exactly characterized within poly time, in the average case. Very exciting new work.

  • Research talk: Soft Astronomy, by Alice Quillen, at the Institute for Advanced Study. Using soft-body models to model astronomical bodies, like asteroids being hit by impacts.

  • History post: Invention of the Computer, by Jürgen Schmidhuber. The history of the hundreds of people who contributed to the development of the computer.

  • Research talk: Quantum Supremacy, rigorously, by Umesh Vazarani at the Simons Institute. Quantum Supremacy is showing that a quantum computer can do something infeasible for classical computers. Working on a theoretically justification for such a claim.

  • Open Problems Paper: Open Problems in Quantum Query Complexity, by Scott Aaronson. Query Complexity: the field where computer scientists can prove lower bounds. Open Problems!

  • Research overview paper: Overview of Geometric Complexity Theory, by Ketan Mulmuley. GCT is one of the most impressive attempts at solving P vs. NP. It uses algebraic topology, one of the most high-power areas of mathematics. And it has proven some cool results, though it hasn’t made a ton of concrete progress.

  • Research popularization: The impossible tetraquark, by Charlie Wood at Quanta Magazine. A new composite particle was discovered, a four-quark particle containsling two charm quarks. It’s much more stable than previous tetraquark, by a factor of 2000. Approximate simulations had guessed that it wouldn’t be exist at all, but experiment proved the simulations wrong.