021

Two Warnings Hit Anthropic on the Same Day, From Inside and Out

No.21 · Thursday, September 10, 2026 · Today in AI

Terms in this issue

In this issue

1. 01Two Warnings Hit Anthropic on the Same Day, From Inside and OutA pretraining researcher resigned saying the labs are "gambling with our lives," and the head of alignment put the odds of humanity being wiped out within ten years above 10%.
2. 02AI is shaking up software, just not as fast as the stock prices saySalesforce and Workday shares were weighed down by AI fears, yet their results grew by double digits. The Wall Street Journal has pointed to the gap between stock prices and earnings.
3. 03The People Who Call AI Dangerous and Keep Building ItThree people inside and around Anthropic named the same danger on the same day. Why the one who left and the ones who stayed reached different conclusions, read together with the company's own risk report.
4. 04Meta's personal agent Muse sells trust by designMeta has launched Muse, an agent that handles email, bookings, and payments on your behalf, in the US first. An isolated virtual machine for every user and an approval gate are the core of the pitch.
5. 05ChatGPT Images 2.5 is twice as fast and holds its editsOpenAI has released ChatGPT Images 2.5, cutting image generation time in half. The API gets two models alongside it: the fast Flare and the precise Sunburst.
6. 06Claude Managed Agents, as three founders actually use themAnthropic has published a roundtable video with three founders. Wispr Flow, Actively, and Pendo explain why they rent agent infrastructure instead of building their own.