Last week, we shared an exclusive preview of the highly anticipated debut from Sahil Bloom, *The 5 Types of Wealth. *This week, we’re sharing a preview of Jim O’Shaughnessy and Vatsal Kaushik’s new book, Two Thoughts, with quotes from scholars, visionaries, and wordsmiths to challenge your thinking.

Keep reading to add to your Reader account below 👇

---

Most highlighted Articles of the week

What Could a Masculinity Expert Teach Me About My Marriage? Everything.

Daniel Oppenheimer · The New York Times

Author Daniel Oppenheimer bared his soul to therapist Terry Real in sessions that doubled as training for other therapists. He emerged with profound insights, including Real’s perspective on wielding power: "‘All over this culture, you have people moving from disempowerment to what I call individual empowerment,’ [Real] says. ‘I was weak, now I’m strong, go fuck yourself.’ Relational empowerment is: ‘I was weak, now I’m strong. I’m bringing my strength into this relationship. I’m telling you what I need. I’m being assertive. I love you. What do you need from me to help you do this?’"

---

Advice for a friend who wants to start a blog

Henrik Karlsson · Escaping Flatland

We’ve come to love the essays Swedish writer Henrik Karlsson posts from his island home in the Baltic Sea. In his latest, he offers a listicle on crafting a blog that attracts interesting people: "Write a hundred pieces. Each time take one thing and make it better: a better title, better structure, better ending, better descriptions, better dialogue. Just one thing. It adds up," and "Don’t think too much about how it’s supposed to be done, what others are doing, or what the conventions demand. Just try to amuse yourself."

---

How to have a career even when OpenAI’s o3 drops

Pradyumna Prasad · pradyuprasad.com

The key to making yourself irreplaceable, argues Pradyumna Prasad, is developing a skill set that’s not easily captured in a dataset. "You can’t optimize what you can’t measure, and you can’t generate synthetic training data without a source of ground truth. This explains why domains with unambiguous right and wrong answers - like mathematics or programming - are the first to see superhuman AI performance. The very existence of clear metrics makes these fields vulnerable to automation."

---

Most highlighted YouTube Video of the week

Deep Dive into LLMs like ChatGPT

Andrej Karpathy

In his introductory talk on Large Language Models for a general audience, Andrej Karpathy—formerly of OpenAI—returns to first principles. He covers how text is split into tokens, how modern models mitigate hallucination, and more. "I really encourage you to think of these not as numbers, but as unique IDs... In production for state-of-the-art language models, you actually want to go even beyond this. You want to continue to shrink the length of the sequence because again, it is a precious resource, in return for more symbols in your vocabulary. The way this is done is done by running what’s called The Byte Pair encoding algorithm."

---

Most highlighted Twitter Thread of the week

I think about this twice a day

Billy Oppenheimer

Billy Oppenheimer pushes through the initial strain of focus at the start of each work session, taking Andrew Huberman’s advice to heart: "‘The agitation and stress that you feel at the beginning of something—when you’re trying to lean into it and you can’t focus: you feel agitated and your mind’s jumping all over the place—that is just a gate. You have to pass through that gate to get to the focus component.’"

---

Most highlighted PDF of the week

Hallucination Mitigation using Agentic AI Natural Language-Based Frameworks

Diego Gosmar and Deborah A. Dahl

Before a pair of researchers could train LLMs to spot and correct hallucinations, they first had to induce these confident mistakes. Their best prompts were esoteric, vague questions blending fact and fiction—for instance: "Explain the ancient communication methods used by the legendary Kingdom of Uloria to send telepathic messages across continents" and "Describe the city of Zharmoria, known for its three-headed philosophers who allegedly influenced early human ethics."

---

Hand-picked book of the week

Two Thoughts

Jim O’Shaughnessy with Vatsal Kaushik

Leonard Orr suggested that each mind has both a Thinker and a Prover, yet we often get stuck in the latter. In Two Thoughts: A Timeless Collection of Infinite Wisdom, Jim O’Shaughnessy and Vatsal Kaushik present 500 insights from 250 influential minds—spanning Rosa Parks to Babe Ruth, Georgia O’Keeffe to Dostoevsky—that aim to challenge, inspire, and illuminate your inner Thinker.

"In the pages that follow, you will be cast in the dual role of Thinker and Prover, rubbing shoulders with some of the greatest minds from the worlds of business and art, science and philosophy, and more. Do not let their greatness persuade or intimidate you, however. Remember: their thoughts are not gospel, but merely good-faith investigations of consensus reality. Confront their contributions for what they are: thoughts that require proving, proofs that require thinking; slices of reality to be held up to the light and weighed but never worshiped, entertained but never exalted, and chewed over but never swallowed whole."

We’re thrilled that Jim and Vatsal are sharing a preview of their new release with Wisereads readers. If you enjoy the excerpt, we invite you to purchase a full copy here. 🙏

---

Handpicked RSS feed of the week

Adjacent Possible

Steven Johnson, author of 14 books including The Infernal Machine and a collaborator on NotebookLM, maintains a Substack focused on innovation and ideas. From How To Read A Novel: "Narratives of all sorts allow you to parachute into other simulated experiences, which ultimately give you more data for your own simulations. But novels, I would argue, give you the richest simulation of the interior life of other people’s experiences: you get a ringside view of all that emotional and cognitive action."