“Claude crash” ... The model drama is the least interesting part. The real question is diffusion: how fast does agentic AI move from cool demo → default workflow inside real companies? After 20 years around SaaS, my bias is: adoption is never linear. Enterprises don’t swap

💵 Most people say they want high returns. Few are clear what that actually implies. Our latest post goes to the uncomfortable question: what kind of risk are you really taking - and are you being paid for it? In public markets, risk feels like volatility: prices move daily,

Jack Dorsey just said the quiet part out loud. Block is cutting its workforce by nearly half — and he links it directly to AI tools. Stock popped ~25% after-hours. Read that again. This is not “AI bubble” talk. This is AI 101: same output (or more) with a much smaller team.

McKinsey just said the quiet part out loud. Bob Sternfels (their global managing director) casually dropped this line: McKinsey is ~60,000 “employees”. 40,000 humans. 20,000 agents. The most “white-collar” business model on earth. The cathedral of PowerPoint. The factory of

Poll check: 44% still think AI is a bubble (249 votes). Interesting. One reason is that people keep framing the Pentagon angle like it’s something new. It’s not. “Defense uses AI” has been true for years. The DoD has been pushing AI since at least Project Maven (2017), then the

Peter Thiel on selling Facebook $META stock too early: “The incorrect view I had back in 2012 was… that going from $100B to $1T was much harder. In retrospect, going from $100B to $1T was maybe the easiest.” “The biggest mistake people make is never making mistakes.”

Big tech copying startups is old. AI makes it faster and closer to the workflow. New blog: the incumbent playbook + the A–E outcomes (die / pivot / get bought / they fail / regulation / coexist). If you’re building something “feature-shaped”, read this. open.substack.com/pub/studioalph…

Big companies have always copied startups. But sometimes they do something even more revealing: they buy the capability first — as a customer — and only later decide whether to internalize it. Today’s example is from defense: the Pentagon is reportedly looking at Ukrainian

🚨 The a16z consumer AI Top 100 is back! For the sixth time, we ranked consumer AI websites and mobile apps by usage (monthly unique visits and MAUs). This edition, we changed the rules. Here's why - and what the new list says about where consumer AI is heading 👇

People asking me what AI will do to jobs — and what young people should learn. None of us knows the future, but you still have to make decisions. Here’s the version I’d bet on: What I hear: “Forget deep expertise. Become a generalist. Learn people skills and pivot fast.”

This week, Jensen Huang pointed to a >$1tn revenue opportunity for Nvidia’s Blackwell and Rubin systems by the end of 2027. That is roughly 4.6x. How can he be that confident? Our new AI blog pulls together a set of brand-new reports and data released over the past few weeks by