1. 1H 2026 Market Structure & Flows — Scott Rubner

  • Why read: Explains the structural forces driving equity markets in the second half of 2026.
  • Summary: Structural changes, rather than macro events, dictate today's stock market. Just ten companies make up 40% of the S&P 500, and semiconductors account for nearly 20%. Household equity ownership is at record highs, especially among lower-wealth brackets, though cash balances stay elevated. Record ETF inflows further tie the market to passive capital. Traders must grasp these dynamics to operate in a highly concentrated environment.
  • Read more

2. Noise — roachcap.com

  • Why read: A reminder to eliminate distractions and trust your preparation.
  • Summary: Pre-exam chatter offers zero useful information; it only breeds uncertainty. The author realized arriving exactly as a test began cut out the noise and sharpened focus. The same logic applies to daily decision-making. Filtering out useless chatter lets you think clearly and build conviction. Self-doubt kills more dreams than failure.
  • Read more

3. See the routing in action — hustlecoding.github.io

  • Why read: Explains how task routing and playbooks work in pstack.
  • Summary: The `/poteto-mode` feature reads a task description and maps it to one of 17 playbooks. Each playbook provides a step-by-step recipe for tasks like metric improvement or exploratory problem-solving. The matched playbook is presented for review before work starts. The system can also run models in parallel to collect diverse feedback. Developers can easily override default rules to customize their automation pipeline.
  • Read more

4. Of Other Spaces, Foucault, 1967 | Governance study group — Yak Collective

  • Why read: Examines whether Foucault's theories on physical spaces survive the post-pandemic shift to digital environments.
  • Summary: A study group analyzed Foucault's idea of heterotopias: spaces acting as counterpoints to normal society. They found his original examples, like the strict boundary between family and workspace, outdated after COVID. The discussion weighed how heterotopias compare to modern, memoryless "junk space." Participants debated if the term is now too broad, given almost any structured environment could fit. The concept still offers a framework for thinking about architecture and digital-physical boundaries.
  • Read more

5. You Can't Hire Someone to Be Yourself — Dan McAteer

  • Why read: Argues for focusing on wisdom as AI commoditizes raw intellect.
  • Summary: Schools long trained humans to act like computers, prioritizing calculation and procedure. Now that actual computers handle these tasks better, intellect is cheap to rent. Wisdom, however, cannot be outsourced. AI can write code or draft emails, but it cannot face hard emotions, act with courage, or make deeply personal life choices. People should stop trying to be flawless machines and focus on human judgment.
  • Read more

6. (Quite) A Few Words About Async — Il y a du thé renversé au bord de la table !

  • Why read: Defines the technical differences between asynchronous, non-blocking, and concurrent code.
  • Summary: Modern apps rely on an event loop to process tasks in tight windows, usually under 16ms. To keep the app responsive, code on the main thread must be non-blocking; it cannot freeze the loop waiting on network or file I/O. Asynchronous programming uses tools like callbacks or promises to handle this non-blocking behavior. Concurrency simply means scheduling independent tasks, without dictating exact run times. Knowing the difference helps developers build faster, more stable software.
  • Read more

7. Understanding the Dynamics of the AI Ecosystem with Pace Layers — Drew Breunig

  • Why read: Uses Stewart Brand's Pace Layers to explain the AI industry's structure and bottlenecks.
  • Summary: The AI ecosystem operates across different speeds, from rapid prompt engineering to slow physical infrastructure. Fast layers innovate and learn; slow layers provide rules, memory, and power. Problems arise when heavy investment tries to force slow layers, like energy grids or governance, to match software development speeds. Understanding these speed limits explains current AI bottlenecks. The system only stays resilient when each layer moves at its natural pace.
  • Read more

8. Longreads + Open Thread — Byrne @ The Diff

  • Why read: Deep dives into NYC rent control, the math of counting, and tech adoption.
  • Summary: The author covers the long-term fallout of NYC rent control, noting how it degrades housing quality and deters talent. He also reviews Joel Sobel's work on counting, showing how errors compound in multi-step processes. An interview with Cory Doctorow argues a user's identity matters more than the tech they use. But in business, companies often have to strip away identity and adopt tools like AI just to survive.
  • Read more

9. An American Redemption — Kyle Harrison

  • Why read: Defends America's structural capacity to course-correct.
  • Summary: Writing for the Fourth of July, the author counters the default cynicism of his generation. He admits the country's historical flaws but views acknowledging them as an act of love. America's primary strength is a structural design built for ongoing reform. He argues for focusing on improvement instead of giving in to pessimism. Real patriotism means recognizing both the broken parts and the capacity for repair.
  • Read more

10. The Adults Don't Exist — George Mack

  • Why read: Shows that even historical geniuses failed at basic life management.
  • Summary: We assume great historical figures had their lives in order. They didn't. Mozart was a musical genius, but he chronically overspent and lived in heavy debt. He regularly sent letters begging friends for money to survive, dying penniless at 35 in an unmarked grave. The fully functional adult is a myth.
  • Read more

11. The Shape of Enshittification — Ryan Levesque

  • Why read: Explains why digital platforms degrade as they chase monetization.
  • Summary: The internet is in structural decline due to "enshittification." Platforms built on discovery are now packed with ads, trapped in closed ecosystems, and controlled by algorithms. Consequently, users read less deeply and navigate the web less naturally. The old model of social media is effectively dead. Recognizing this cycle can prompt creators and users to find or build better online spaces.
  • Read more

12. What it feels like to work with Mythos — Ethan Mollick

  • Why read: A hands-on review of Claude 5 Fable across everyday tasks.
  • Summary: The author tested early access to Claude 5 Fable on normal tasks, ignoring its heavily discussed security features. Working with this tier of model feels different in capability and tone. While powerful, the model's guardrails dictate how it handles real work. The review sets realistic expectations for professionals using new AI tools, pointing out that understanding a model's constraints matters as much as its baseline intelligence.
  • Read more

13. “As a writer, I often felt like I was trying to enter various clubs that didn’t want me” — The Substack Post

  • Why read: Musician Maggie Rogers on returning to songwriting after a six-year break.
  • Summary: After six years out of the studio, Rogers describes the messy process of post-pandemic creativity. Writing new music required vulnerability, leaving her feeling both lost and renewed. She emphasizes the necessity of editing: stepping away and heavily revising to lock in a clear vision. Her aim is to produce urgent, emotionally honest art. Good creative work relies more on persistence and revision than raw inspiration.
  • Read more

14. Can a token ROI crisis shake the strongest companies? — SaaStr

  • Why read: Examines the gap between high corporate AI spending and low actual returns.
  • Summary: Companies are spending heavily on AI tokens, but few can point to real ROI. Early enthusiasm is hitting standard enterprise economics. This token ROI crisis could hurt even the best-funded tech firms. To justify ongoing budgets, businesses have to stop running AI experiments and start deploying tools that clearly drive revenue or cut costs.
  • Read more

15. What’s 🔥 in Enterprise IT/VC #505 — Ed Sim

  • Why read: Notes a shift among enterprise buyers who want to own, not rent, their AI infrastructure.
  • Summary: Enterprise IT is debating who actually owns the intelligence generated by AI. Technical buyers want control over their compute, data, and proprietary edges, rather than renting them from major AI labs. Vendors that let companies own their AI infrastructure are beating standard model providers. The traditional "forward deployed engineer" role is changing as a result. Companies need to check if their AI partnerships build internal value or just increase vendor lock-in.
  • Read more

Themes from yesterday

  • The Limits of Technology: AI and automation handle intellect, but judgment, wisdom, and conviction stay strictly human.
  • Structure Over Hype: Long-term outcomes are driven by deep structural realities like market concentration and physical infrastructure pacing, not news cycles.
  • Platform Decay: Internet platforms are degrading as they chase ad revenue and closed ecosystems, changing how people consume media.
  • The AI ROI Check: Enterprises are tired of AI experiments. Buyers now demand clear returns and actual ownership of their models and data.