So Much AI Progress

Hello, and happy Sunday! It was arguably the biggest week yet in AI: Google and Microsoft laid out their visions for an agentic and human-focused future, respectively, while Anthropic released the next version of its flagship model. Scroll down to see our on-the-ground reports by Dan Shipper and Alex Duffy. (In the meantime, we had an announcement of our own at Every.) We’ll be off on Monday for the U.S. Memorial Day holiday—see you on Tuesday.—Kate Lee

Was this newsletter forwarded to you? Sign up to get it in your inbox.

---

Knowledge base

"Vibe Check: Claude 4 Opus" by Dan Shipper/Chain of Thought: Dan put Anthropic's new Claude 4 Opus through its paces by feeding it 50,000 words of his unfinished book (among other tests) and discovered it's eerily good at telling you what you're trying to say better than you can. Read this to understand why the AI competition is now more about product than pure smarts.

"Microsoft’s AI Vision: An Open Internet Made for Agents" by Dan Shipper/Chain of Thought: While everyone's busy hyping the latest AI startup, Microsoft is quietly building the plumbing for our AI future. Dan went to Microsoft’s Build conference where CEO Satya Nadella and CTO Kevin Scott unveiled their vision for an "agentic web"—an open internet where AI agents talk to each other the way browsers talk to servers. Read this if you want to understand how Microsoft plans to wire the infrastructure that will make AI agents useful in your daily life, and check out Dan’s interview with Kevin in full. 🎧 🖥 Watch on X or YouTube, or listen on Spotify or Apple Podcasts.

"Google's AI Vision: Make Tech Human Again" by Alex Duffy/Context Window: Google's tsunami of AI announcements at its I/O conference—from mind-blowing video generation to language-translating glasses—suggests that the company is flexing its unique ecosystem advantages (search, video, mobile) while pursuing AGI. Read this if you want a glimpse of our AI future that feels hopeful rather than dystopian.

"When o3 Plans Your Career Better Than You Do" by Katie Parrott/Working Overtime: Where do you see yourself in five years? Katie let OpenAI's o3 model take a crack at answering that dreaded question—and it dreamed bigger for her than she did for herself. Read this for a glimpse of how AI is evolving from reactive assistant to agenda-setting co-pilot—and a career coach.