WM on WM's Blog

Twenty Years of Coincidence, Not Twenty Years of Oppression

Tue, 01 Sep 2026 21:30:00 +0800

30-second summary: In episode 6 of the Chinese podcast et Voice, Lumi describes leaving a conventional track to become a full-time polar expedition guide. What’s worth writing down isn’t “quit her job, went to Antarctica.” It’s four counterintuitive judgments: her predicament wasn’t oppression but alignment concealing the absence of self-direction; her sense of agency wasn’t reasoned into existence but purchased with a job opportunity she lost permanently; the criterion that turns “fleeing” into “departing” isn’t a healed wound but having a next destination — her second conflict was worse, and she was calmer; and the premise she named her own podcast after, the search for a “good place,” she ended up falsifying herself.

Every Fact Checks Out, Every Causal Arrow Is Forged: I Had Two AIs Write My Podcast Reflections

Tue, 01 Sep 2026 21:00:00 +0800

30-second summary: I had three AIs, under three different information conditions, each write a reflection on the same podcast episode. The fake one was easy to spot — it had never heard the audio and invented, from the show notes alone, a mechanism that sounded clever and turned out to be the opposite of what the episode actually said. The hard one was the most convincing: written by an AI holding years of memories about me, it fabricated no facts whatsoever, got every timestamp right, and tagged all its sources. What it forged was the causal arrow — the essay reads “I heard this, and so I thought that,” when every one of those thoughts predated the episode. It reads like insight because it is accurate about me, not because it told me anything.

What Should I Do With the Life I Have Left?

Tue, 25 Aug 2026 00:00:00 +0800

30-second summary: I built a tiny life-countdown website that turns the abstract idea of “the future is long” into a timeline you can actually see. That made me revisit a question that the past few years of pessimism and FIRE-number-crunching had quietly replaced — not “how do I retire as soon as possible / avoid getting hurt by this era,” but “what do I actually want to trade the next several decades for.” This post doesn’t have an answer; it’s just a record of starting to ask that question again.

Why I Wanted to Become a Digital Nomad, and Why I Went Home Anyway

Sat, 22 Aug 2026 17:58:16 +0800

30-second summary: I seriously planned and tried living as a digital nomad, only to find that frequent movement itself constantly ate up attention and digital-world overhead — leaving me with less capacity to actually perceive reality. In the end I went home: not giving up on faraway places, but realizing that movement isn’t the goal, experience is. Home can be a low-friction base that lets my attention go to whatever world is actually worth exploring, physical or digital/intellectual.

Once Experiment Cost Approaches Zero, What Does Research Itself Become?

Thu, 20 Aug 2026 12:00:00 +0800

30-second version: What AI really lowers isn’t the cost of writing code — it’s the search cost across the entire hypothesis space. An idea that used to have a 20% success rate and take two days to test wasn’t worth trying before; now it is. So the bottleneck moves from execution to selection: when you can run two hundred experiments a day, the hard part is which two are worth looking at. And alpha discovery has gotten cheap — alpha validation hasn’t. As the search space grows, multiple testing, overfitting, and selection bias all get worse; you can always fish a pretty equity curve out of pure noise. So the researcher’s job shifts from “finding one good result” to “ranking a pile of results by strength of evidence.” What’s scarce is no longer the ability to search, but the prior knowledge of where to search, and the discipline to actually kill a signal.

When Writing Code Is No Longer the Bottleneck in Quant Research

Thu, 20 Aug 2026 00:00:00 +0800

30-second version: AI agents can now read code, run experiments, and fix bugs on their own — writing code is no longer the bottleneck in quant research. So what’s scarce is no longer “how many experiments can I run,” but what’s worth trying, and what results are worth believing. The researcher’s role isn’t just “the one who asks questions” either — asking is just as cheap now; what’s genuinely hard is designing experiments that can distinguish between competing explanations. Once execution gets cheap, you’re not facing an empty ocean — you’re facing an ocean that suddenly has a few hundred extra boats in it. The problem shifts from “how do I find something” to “how do I know what I found isn’t just glass shards.”

Who's to Blame for This One: Assigning Blame in a Mahjong Game Log

Fri, 14 Aug 2026 00:00:00 +0900

Origin Link to heading

After dealing into a fourth place, I wanted to curse the RNG, but honestly I wasn’t sure: was this hand simply unplayable, or did I just play it badly? The trouble with mahjong is that there’s a thick layer of variance between the outcome and the process — a single game can’t settle anything, but a person only ever gets to play one game at a time.

Principles for an Agent Research Pipeline Workflow

Fri, 29 May 2026 00:00:00 +0800

Experiment Tracking & Doc Writing

Sat, 22 Nov 2025 16:18:58 +0800

Experiment Tracking & Doc Writing
- Experiment
- Weekly Doc Writing
  - Structure

Experiment Tracking & Doc Writing Link to heading

Experiment Link to heading

Use configs
Save results
Have one script to compare

1. Use configs: for all pipeline-level hyperparameters Link to heading

Base config: captures every optional feature in the pipeline, shared across experiments
Experiment config: modifies 1–2 items from the base config (overrides)
Stored config: for each experiment run, make sure it can be traced back later

Recommend using *.yaml for better readability example

macOS Setup

Sat, 22 Nov 2025 16:11:42 +0800

TOC Link to heading

TOC
Reference
- system
- Internet
- Brew
- iTerm2
- Zsh
- Git
- Vim
- Dev Env
  - pip
  - pyenv
Software
TheMagicQuant
quarto: ipynb to html
node
Windows Setup

Reference Link to heading

macOS Setup Guide

Ghost in the Shell 2: Innocence

Thu, 12 Jun 2025 00:25:29 +0800

Link to heading

What is innocence?

To project oneself onto a doll — is that a sin?

To project oneself onto a child — is that a sin?

To seal intelligence inside a human body — is that a sin?

To seal consciousness inside a doll’s frame — is that a sin?

The maker and the made; the raiser and the raised; the projector and the projected upon. Who among them is innocent?

Trip Itinerary

Fri, 12 Jul 2024 00:00:00 +0000

Master's Thesis Defense Q&A

Wed, 29 May 2024 00:00:00 +0000

“Stock Embedding and Stock Association in Fund Holdings” — afternoon session, May 24, 2024

Q: The correlation of stock returns shows a structural break over the 2014–2023 historical period (Figure 4.10), so why doesn’t the Fama-MacBeth regression (Table 4.17) account for this structural change, and instead compute significance directly on the mean factor return over the entire history?

A: A structural break in the time series of pairwise stock return correlation doesn’t imply that the cross-sectional correlation between the anomaly factor and future returns also undergoes a structural change over time. Looking at the daily IC mean and cumulative IC (Figure 4.16), the cross-sectional correlation coefficient between the associated-momentum factor and future returns shows no obvious structural change over time. Since the factor serves as the explanatory variable and future returns as the dependent variable, their correlation coefficient is approximately the factor return; so when the daily IC is approximately stationary, a significance test can be run directly on the time-series mean of the factor return. In addition, a Newey-West adjustment was applied to the time-series variance during testing. That said, for greater rigor, one could also perform structural-break detection on the time-series distribution of daily IC before the Fama-MacBeth test, or perform structural-break detection on the time-series distribution of the estimated regression coefficients (i.e., the factor returns) during the test.

Marathon Scholarship Application Statement

Sun, 26 Nov 2023 12:47:00 +0800

Help me write an application statement for the Marathon Scholarship

This scholarship aims to promote the marathon spirit of “challenging oneself, surpassing one’s limits, showing perseverance, and never giving up” on university campuses and among student communities, and to encourage young elite groups to experience and spread a positive attitude toward life through the sport of marathon running.

Applicants are required to state, in their application, their personal understanding of the athletic spirit of the marathon.

Food Worth Eating at Yanyuan

Thu, 01 Jun 2023 00:00:00 +0800

Since someone said the food at school isn’t good, and some people are seen as having strong tastes, here’s a list of the Yanyuan food I eat most often:

Breakfast

Xueyi: doufu nao (tofu pudding, with brine sauce or DIY white sugar), beef pancake, corn
Xuewu: wontons, sesame-paste flatbread
Nongyuan: Shaxian-style noodles with pork wontons and mung bean soup
Xuewu Heliwu: spicy ramen
Dorm: cereal with milk or soy milk, mixed with dark chocolate cranberries etc. (only good for one bowl)

Lunch

In Qingdao

Sun, 19 Mar 2023 12:45:00 +0800

Arrived at Qingdao Station yesterday at 14:30.

I wanted to see the sea, so I walked out of the station to Zhan Bridge. Packed with people. Walked along the coast all the way to Xiaoyu Hill. Passed a middle school, the Qingdao Art Museum, the China Maritime College. Walked from Xiaoyu Hill to Xinhao Hill, passing Liang Shiqiu’s former residence, Lao She’s former residence, a street lined with elaborately decorated coffee shops.