The Week AI Stopped Advising and Started Acting
Week of August 24 to 31, 2026 | Issue 21 | For smart professionals catching up on AI
This week
Anthropic gave Claude a way to operate laboratory microscopes and robotic arms. Google's search assistant can now watch a flight price and book your hotel. And an NHS watchdog found AI note-takers putting wrong drug names and diagnoses into patient records. AI has stopped suggesting and started doing, so its mistakes now land in the real world.
📡 Industry pulse — 60 seconds
The AI industry headlines from this week. One sentence each, so you have something to say when they come up at work.
- Anthropic gave Claude a way to run laboratory equipment. Its Model Hardware Standard, opened as a preview on August 27, lets agents operate microscopes, robotic arms and liquid handlers.
- A judge ruled the Trump administration illegally punished Anthropic. The Pentagon had labelled the company a national security supply chain risk; the court called that retaliation for protected speech and barred it permanently.
- Bill Gates says nobody has a plan for what AI does to work. His August 26 essay warns of mass unemployment and argues some jobs should be reserved for humans by law.
- Google's Gemini can now handle your inbox by voice. From August 26 you can ask it to search, summarise, star, archive or delete Gmail messages out loud. Check its reach first.
- Tencent released a 770-billion-parameter model anyone can download free. Hy4 preview arrived on August 28 under a permissive licence, two months after its last flagship. Chinese labs keep giving away capable models.
- Nvidia is in talks to back Perplexity at more than $30 billion. The Information reported the discussions on August 28, as AI search moves toward assistants that act on your behalf.