The Friday Review

What we are reading:

This week's picks: big banks post record Q2 profits on a trading and dealmaking surge, IBM has its worst trading day on record on an AI capex shift, and China's Moonshot ships the largest open-weight model to date. Plus one on why AI-exposure researchers can't agree on which jobs are actually at risk, and a detour into how soccer's data converts think about risk.

💰 Big banks posted a red-hot Q2, JPMorgan's $21.2 billion in net income the highest quarterly profit in US bank history, EPS of $6.14 against a $5.85 consensus, and Goldman's EPS nearly doubling year over year on a trading and dealmaking surge. We'll be watching how much of that momentum carries into regional bank earnings over the next two weeks. https://www.wsj.com/finance/banking/big-banks-profits-surge-after-a-red-hot-quarter-on-wall-street-7a25ffbe

📉 IBM had its worst trading day on record after a Q2 warning tied to clients shifting capex toward AI infrastructure, wiping out $67 billion in market value in a single session. https://www.wsj.com/tech/the-inside-story-of-ibms-shocking-profit-warning-839ef4f2?mod=hp_lead_pos1

🚀 China's Moonshot released Kimi K3, a 2.8 trillion-parameter open-weight model already outranking Anthropic's Opus 4.8 on independent benchmarks. https://www.reuters.com/world/china/chinas-moonshot-unveils-worlds-largest-open-ai-model-closing-us-rivals-2026-07-17/

🤖 Five academic studies try to score how exposed different jobs are to AI, and they agree on almost nothing except the easy cases, hairdressers safe, tax preparers a toss-up depending on whose model you trust. https://www.apollo.com/wealth/insights-news/insights/daily-spark/how-exposed-is-your-job-to-ai---the-experts-cannot-agree

⚽ Bloomberg's Odd Lots sits down with two soccer analytics practitioners who describe scouting in almost exactly the terms a volatility desk uses for a book. https://podcasts.apple.com/us/podcast/why-soccer-analytics-works-like-volatility-arbitrage/id1056200096?i=1000777031321