I was feeling a bit nostalgic (and hungry) today after the weekly Friday macroeconomic discussion session, that suppose to end 11:20 but stretches all the way towards noon. There happened to be an AI seminar right after, with lunch — I grabbed a sandwich, fully expecting myself to sit in the back, eating my feelings while absorbing some comforting high-level vibes. Instead, it turns out to be one of the most interesting theoretical advances at the forefront of self-supervised learning....

I’ve always thought the Chicago Symphony Orchestra has one of the most insane string sections on Earth—versatile, responsive, laser-precise, and somehow still warm. Tonight, concertmaster Robert Chen led the strings in Vivaldi’s Four Seasons, and… they’re good. So good. 😙🤌🏻❤️🎶 The Chicago audience really enjoys standing up! Technically, The Four Seasons is a set of four concertos, but somewhat it works even better without a conductor—especially when Chen is steering the ship....

OMG: We analyzed 47,000 ChatGPT conversations. Here’s what people really use it for. Washington Post. https://www.washingtonpost.com/technology/2025/11/12/how-people-use-chatgpt-data/ Methodology: The Post downloaded 93,268 conversations from the Internet Archive using a list compiled by online research expert Henk Van Ess. The analysis focused on the 47,000 chat sessions since June 2024 in which English was the primary language, as determined using langdetect. A random sample of 500 conversations in The Post’s corpus was classified by topic using human review, with a margin of error of plus or minus 4....

Mendelssohn himself resisted attempts to interpret the songs too literally, and objected when his friend Marc-André Souchay sought to put words to them to make them literal: “What the music I love expresses to me, is not thought too indefinite to put into words, but on the contrary, too definite”. (Wikipedia) For (popular) piano works, people LOVE discussing about fav recordings: (Talk Classical) Any recommended recordings? I bought Barenboim’s set in the early days of CDs....

Nature has a lovely, open-source annual graduate survey: Springer Nature conducted a global survey of PhD candidates in science in 2025, in partnership with Thinks Insights & Strategy, a research consultancy based in London. The survey ran in June and July 2025 and received 3,785 self-selecting respondents from 107 countries. The findings about PhD graduate student experiences are covered in an article series from Nature’s Careers team. Link to the survey data: https://figshare....

It’s my birthday today, and it happens to be Chicago’s first snow of 2025 winter. Couldn’t it be happier! And let me share the happiness vicariously with all of you reading. It’s been a lot for the past year for me. Here’s a lovely, interesting song to go with the day — Thank Goodness performed live by Ariana Grande — “getting your dream seems a little strange, well, complicated.” But after all...

P. D. Q. Bach is a fictional composer created by Peter Schickelem (1935-2024). Schickele studied composition at Juilliard and was an accomplished bassoonist. His wrote an awful lot of (brilliant) parodies of classical music tributed to the name P. D. Q. Bach eg the 1712 Overture, The Civilian Barber, etc. Read more about him: NYTimes on Peter Schickele The Boston Musical Intelligencer Article about P. Schikele

Another convenient way to view, compare, and solve Cheap Talk vs. Bayesian Persuasion. Linear-Programming Formulation: Cheap Talk vs. Bayesian Persuasion Consider an information environment $\langle \Omega, A, M, \mu_0, u, v \rangle$ with state space $\Omega$, action space $A$, signal space $M$, prior $\mu_0$, receiver utility $u(\omega,a)$, and sender utility $v(\omega,a)$. (Let every set be finite). Consider direct mechanisms. In general, the sender’s strategy $\sigma:\Omega \to \Delta(A)$ maps each state to a distribution over recommended actions, and the receiver’s strategy $\rho:A \to \Delta(A)$ specifies how she responds to each recommendation....

A general continuous-time dynamic programming problem is given by reward function $h(x, u)$ control function $g(x, u)$ and initial state $x_0$: $$ \begin{align*} V^\star (x_0) :=& \max_{\lbrace u_t\rbrace_{t\ge 0}} \int_0^\infty e^{-\rho t} h(x_t, u_t),\text{d}t\cr & \text{s.t. }\dot x_t = g(x_t, u_t). \end{align*} $$ The Hamiltonian functino $H(x, u, \lambda)$, and Maximized Hamiltonian $H^\star(x, \lambda)$ is defined as $$ \begin{align*} & H(x, u, \lambda) = h(x, u) + \lambda^Tg(x, u),\cr & H^\star(x, \lambda) = \max_u H(x, u, \lambda)....

When I learnt the continuous time and discrete time neoclassical model, it’s feels so tempting to try to put them together in a uniform view. It’s not trivially easy though. And here’s one way to do it. Continuous-Time Neoclassical Growth Model Given $k_0$, $f(\cdot), U(\cdot)$: $$ \begin{align*} & \max_{\lbrace c_t\rbrace_{t\ge 0}} \int_0^\infty e^{-\rho t} U(c_t),\text{d}t\cr & \text{s.t. }\dot k_t = f(k_t) - \delta k_t - c_t \end{align*}\tag{1} $$ Taking (appropriate) derivative of the Hamiltonian function solves this problem — in general, for any continuous-time problem in the following form: $$ \begin{align*} V^\star (x_0) :=& \max_{\lbrace u_t\rbrace_{t\ge 0}} \int_0^\infty e^{-\rho t} h(x_t, u_t),\text{d}t\cr & \text{s....