Did Apple (Finally) Get AI Right At WWDC?, Anthropic’s Worry, Microsoft vs. OpenAI

TL;DR

• 1

  Apple's strongest AI edge is distribution, not model leadership: The big WWDC leak suggests Apple will make Siri an operating system layer with a swipe-down "search or ask" interface, app actions, model switching, and deeper control across iPhone, iPad, and Mac.

• 2

  The new Siri vision finally matches what AI can actually do: Instead of glossy promises like the Bella Ramsey ad era, Apple appears to be targeting realistic capabilities such as multi-step requests, app routing, and persistent chat, even if much of it already feels table stakes in ChatGPT, Gemini, and Pixel devices.

• 3

  Execution is still the whole story for Apple: Both hosts land on the same concern that Apple has not shown it can deliver this kind of cross-app agent reliably, and reports that the new Siri may launch as a beta with a waitlist brought back "Apple Intelligence" rollout PTSD.

• 4

  Anthropic's recursive self-improvement warning triggered a marketing-vs-belief debate: The company says its engineers now ship 8 times as much code as in 2021 because of AI, but Roy argues safety alarms ring hollow when Anthropic is still racing toward massive fundraising and a likely IPO.

• 5

  Inside Anthropic, the tension may be real even if the messaging is strategic: Citing Ethan Mollick, the conversation frames the company as three groups in conflict at once: normal hypergrowth business operators, model builders, and "philosopher kings" worried about what happens next.

• 6

  Microsoft is preparing for life after OpenAI exclusivity: Build 2025 had the feel of a newly single company, with Mustafa Suleyman saying Microsoft is not yet in the top tier with Google DeepMind, OpenAI, and Anthropic, while Satya Nadella hinted that OpenAI could eventually become a cloud competitor too.