archive
REX·DAILY
Sunday June 21 Anno 2026 · Ibaraki, Osaka
I Daily Wisdom
Robert Greene The Daily Laws June 21 · The Slow Power Grab

Ambition can creep as well as soar.

—Edmund Burke

In almost every film Alfred Hitchcock made, he had to go through the same wars, gradually wresting control of the film from the producer, the actors, and the rest of the team. His struggles with screenwriters were a microcosm of the larger war. Hitchcock always wanted his vision for a film to be exactly reflected in the script, but too firm a hand on his writer’s neck would get him nothing except resentment and mediocre work. So instead he moved slowly, starting out by giving the writer room to work loosely off his notes, then asking for revisions that shaped the script his way. His control became obvious only gradually, and by that time the writer was emotionally tied to the project and, however frustrated, was working for his approval. A very patient man, Hitchcock let his power plays unfold over time, so that producer, writer, and stars understood the completeness of his domination only when the film was finished. To gain control of any project, you must be willing to make time your ally. If you start out with complete control, you sap people’s spirit and stir up envy and resentment. So begin by generating the illusion that you’re all working together on a team effort; then slowly nibble away. If in the process you make people angry, do not worry. That’s just a sign that their emotions are engaged, which means they can be manipulated.

Daily Law: Overt manipulation and power grabs are dangerous, creating envy, distrust, and suspicion. Often the best solution is to move slowly.

The 33 Strategies of War

, Strategy 29: Take Small Bites—The Fait Accompli Strategy

Steven Pressfield The Daily Pressfield Day 172

HABIT “Habit will be your champion. When you train the mind to think one way and one way only, when you refuse to allow it to think in another, that will produce great strength in battle.” Gates of Fire, p. 139 The choreographer Twyla Tharp starts each day with two hours at the Pumping Iron gym on East Ninety-First Street in Manhattan. My friend Randy Wallace works out with weights underwater in Malibu with Laird Hamilton, the famous big-wave surfer. Then they go to work. Twyla and Randy are pursuing fitness only secondarily. What they’re really doing is drilling themselves in the habits of a professional. They are preparing themselves to confront their own Resistance. Habit is their ally in this daily struggle.

Leo Tolstoy A Calendar of Wisdom June 21

The misery of the unintellectual life brings us to the need for an intellectual life.

Before, I lived in sin, and I saw that the majority of people around me lived in the same way.

Like a robber, I knew I was unhappy and I suffered, and that people around me were unhappy and they suffered; and I didn’t see any way out of this situation except for suicide or death.

Life seemed terrible to me.

And then I heard the words of Christ and I understood them.

And life ceased to seem an evil, and instead of desperation I felt a happiness for life which surpasses even death.

We can understand wisdom in three ways: first, by meditation; this is the most noble way.

Secondly, by being influenced by someone or following someone; this is the easiest way.

Third is the way of experience; this is the most difficult way.

—CONFUCIUS

When you suffer, think not on how you can escape suffering, but concentrate your efforts on what kind of inner moral and spiritual perfection this suffering requires.

All the misfortunes of mankind collectively and individually are not useless; they bring people and individuals and nations in different ways closer to the purpose which is set before them: the appearance of God, for every person in himself, and in all mankind.

II The Day Ahead
  • Today · 11:00 ミサ当番 w/ 木島 2026年間予定
III Signals & Dispatches

AI / ML

  • @gregisenbergsource ↗ GLM 5.2 is being framed as the local AI breakout: 1M context, strong coding, MIT license, and usable through Ollama or LM Studio.
  • @burkovsource ↗ Burkov argues coding models are converging because labs can synthesize bug fixing data, then combine supervised traces with verification based RL.

Drug Discovery

  • @lexmutatiosource ↗ A useful caution: AI can compress modeling and lab work, but human trials still dominate the drug development clock.
  • @PeptideListsource ↗ A peptide note worth scanning: the category is not just the molecule, but the match between molecule, body, and decision context.

Biotech

  • @CELZonNASDAQsource ↗ Greenstone Bio and Intel are being cited as another signal that AI enabled biology is moving from demos into scaled infrastructure.
  • @elander777source ↗ A compact bet to watch: partial epigenetic reprogramming and MUSE cells as possible engines for the next biotech decade.

Tech

  • @petergyangsource ↗ HyperFrames is showing HTML native video generation workflows inside Codex and Claude Code, with an open source toolchain.
  • @KuanFangsource ↗ UMA proposes robot foundation models that treat 3D object motion as the shared representation across heterogeneous robot tasks.

Japanese Politics

  • @joseperojaponessource ↗ The Bunshun audio analysis story is feeding renewed pressure over alleged smear video activity during last year's LDP leadership race.
  • @hannichiyamerosource ↗ Immigration and school capacity are again becoming a live political grievance, with education support and local services in the foreground.
IV Sky over Ibaraki
☁️ +26°C
humidity 84% · wind →23 km/h
cloudy, humid, and breezy in Ibaraki
V What Rex Can Do Today
  1. Scan the Fr. Conor pipeline logs and flag any Dropbox or Spotify publication drift before it becomes noisy.
  2. Review the medical records repo for stale medication context and separate current status from old memory notes.
  3. Draft a short StemRIM AI hardware memo that compares local inference value against cloud leakage risk.
VI Toward the Long Game
  1. Read one recent paper on peptide design with structure aware generative models, then extract one method worth reproducing locally.
  2. Spend 30 minutes with a hands on molecular design notebook and make the output inspectable, not just theoretical.
  3. Message one serious AI biology contact with a specific question about validation bottlenecks, not a vague networking ping.