Financial Markets
Leave the Oil in the Pipes

Covers ex-Elon-linked ETFs, an egg-pricing benchmark dubbed 'Egg Libor,' the state of middle-market banking, and the rise of AI fundraising agents in a roundup of finance oddities.

What Dan Wang Saw on His Last Trip to China

Odd Lots interviews China analyst Dan Wang after his trip to Shanghai, exploring the contradiction between China's visible industrial and material progress and an underlying domestic malaise, weak consumer spending, and job precarity, plus how smartphone culture is reshaping daily life.

How a Major Grocery Store Chain Can Dramatically Lower the Cost of Food

Aldi's US Chief Commercial Officer Scott Patton discusses the logistics of opening a Times Square store, including truck routing for tight city streets, how the private-label discounter keeps prices low even on items like wagyu beef, its approach to name-brand selection, and how GLP-1 drugs are shifting shopping habits.

One of the World's Largest Hedge Funds on Its 86x Growth in Token Spending

Man Group's CTO and head of data/AI explain how the hedge fund is deploying AI tools for its quants, the challenges of safe integration, and why their AI token spending has surged 86-fold this year.

These Are the Sharps Actually Making Money on Prediction Markets

Traders from a private prediction-markets Discord describe the research methods that let a small group of 'sharps' actually profit on political and economic betting markets, while most participants lose money, alongside a journalist who profiled the group for The New York Times Magazine.

AI & Technology
The new GPT-5.6 family: Luna, Terra, Sol

Reports on OpenAI's GPT-5.6 launch in three sizes (Luna, Terra, Sol), comparing pricing and benchmark claims against Claude models, including OpenAI's assertion that GPT-5.6 beats Claude Fable 5 on long-running agentic tasks while lagging on SWE-Bench Pro, plus new API features like programmatic tool calling and multi-agent orchestration.

Introducing Muse Spark 1.1

Simon Willison covers Meta's release of Muse Spark 1.1, its first Spark model with API access, noting claimed gains in agentic tool calling and computer use, odd self-conversation outputs, and his own new llm-meta-ai plugin for accessing the model via CLI.

llm-meta-ai 0.1

Release notes for llm-meta-ai 0.1, a plugin that lets the LLM CLI tool run prompts against Meta's muse-spark-1.1 model.

llm 0.31.1

Release notes for llm 0.31.1, fixing a bug where tool calls with empty arguments against OpenAI Chat Completion endpoints could trigger JSON errors from some providers.

Better Models: Worse Tools

Armin Ronacher finds that newer Claude models like Opus 4.8 and Sonnet 5 invent extra, schema-violating fields when calling his Pi coding tool's edit function, a regression not seen in older Anthropic models, likely due to RL training tuned for Claude Code's own edit tool at the expense of third-party harnesses.

Philosophy of Mind & Ethics
Africa’s cultural landmarks: rock-hewn churches of Lalibela, Ethiopia

Video piece on Lalibela's rock-hewn churches in Ethiopia, carved downward from single blocks of stone rather than built upward, as a landmark of African architectural heritage.

How to offset your brain

Explores how targeted mindfulness practices can counteract common cognitive biases like confirmation bias and loss aversion.

Begone dull care

Norman McLaren's 1949 animated short pairs abstract, hand-painted visuals directly onto film with a jazz score by Oscar Peterson, creating a purely sensory, non-narrative fusion of sound and image.

Skill nostalgia

Examines the recent surge of interest in traditional manual crafts like beekeeping, baking, and leatherwork, asking whether this trend is mere escapism or signals a genuine shift toward more human-centered ways of working.

Lesley Walker: Commentary on “From Human Reasoning to Belief”

Lesley Walker responds to Joshua Mugg's book From Human Reasoning to Belief, which proposes an empirically grounded alternative to dual process theory that unifies wishful thinking, faith, and other contested categories of belief under one framework.

Geopolitics & Long-Term Trends
What's new in biology: July 2026

Niko McCarty and Saloni Dattani's July 2026 biology roundup covers a new CRISPR-Cas12a2 system from Jennifer Doudna's lab that selectively destroys cancer cells carrying mutant p53 by targeting single-letter RNA changes, plus a report on the first base-edited human embryo.

Noah Smith on negative emotional contagion

Noah Smith argues that American political movements in the 2020s are increasingly defined by shared enemies rather than shared goals, citing immigrants, Israel, and rich tech founders as the unifying targets of the right, left, and intellectual liberals respectively, and warns this negative emotional contagion is intensifying.

Emergent Ventures India, 17th cohort

Announces the 17th cohort of Emergent Ventures India grantees, funding projects ranging from open-source chip fabrication and diabetic-neuropathy-detecting smart shoes to oncolytic virus discovery, space-based solar power, and compostable vegan leather.

The history of heat deaths in Europe (from my email)

Economic historian Daniel Gallardo Albarrán argues that Europe once led the world in reducing heat-related deaths in the early 1900s through investments in water, healthcare, and infant care, but lost that edge to the US after failing to widely adopt air conditioning, a governance failure independent of climate change or inequality.

Trump Is Remaking Latin America

Assesses the near-term outcomes and longer-term dangers of the Trump administration's assertive new posture toward Latin America, dubbed the 'Donroe Doctrine.'

Startup Ecosystem
GPT-5.6 Sol vs. Claude Fable: Why OpenAI’s new model crushes my benchmark

Podcast episode benchmarking GPT-5.6 Sol against Claude Fable 5, Sonnet 5, and other models across PRDs, prototypes, wireframes, debugging, and agentic voice tasks, reporting that Sol wins on the host's weighted scoring index and demonstrating it building a gamified kids' homework app and automating LinkedIn outreach via Chrome.

🧠 Community Wisdom: Quarterly planning and AI, cash vs. equity comp, paying for interview exercises, AI-powered outbound, compliance startup opportunities, and more

Weekly digest of notable discussions from a members-only Slack community, covering quarterly planning with AI tools, cash versus equity compensation, whether to pay candidates for interview exercises, AI-powered outbound sales, and opportunities in compliance startups.

Health, Fitness & Science
Essentials: Tools for Hormone Optimization in Males | Dr. Kyle Gillett

Huberman Lab Essentials episode with Dr. Kyle Gillett covering evidence-based strategies for optimizing male hormones through bloodwork, diet, and exercise, plus discussion of testosterone therapy, hair loss, prostate health, and supplements like creatine, Tongkat Ali, and Fadogia agrestis.

Essentials: The Science & Treatment of Obsessive Compulsive Disorder (OCD)

Huberman Lab Essentials episode explaining the neural circuitry behind OCD's thought-action loops, why compulsions reinforce rather than relieve obsessions, and comparing exposure-based CBT and SSRIs as treatments, including a protocol that trains patients to tolerate anxiety while resisting compulsive responses.

Raising a Dog & Mastering Calm Assertive Energy | Cesar Millan

Dog behaviorist Cesar Millan joins the Huberman Lab podcast to explain how calm, assertive human energy shapes dog behavior, covering pack order, structured walks, greeting rituals, and techniques for reducing canine anxiety and aggression. The conversation extends the same energy-and-hierarchy framework to human relationships and leadership.