OpenAI launches always-on agents called Dots hours before the FTC opens an industry-wide probe into Anthropic and OpenAI, while Anthropic's own researchers find China's open-weight GLM-5.3 can already build cyber exploits nearly as well as Claude. Plus Ben Thompson on Meta's agent strategy, McKinsey's estimate that AI could push 11 million US workers into new careers by 2035, Stripe makes OUSD its default stablecoin, and a wildcard: a 245-million-year-old fossil with its stomach and liver still intact.

Lucas Ropek, TechCrunch · Sep 29

OpenAI introduced Dots, an always-on GPT-6 Astra-powered agent that works toward user goals continuously in the background.

Reuters, BNN Bloomberg · Sep 30

The FTC is opening an industry-wide investigation into Anthropic, OpenAI, and other AI labs over consumer and agent-safety risks.

Anthropic Frontier Red Team · Sep 29

Anthropic researchers find China's open-weight GLM-5.3 model builds cyber exploits nearly as well as Claude, with weaker safeguards.

Ben Thompson, Stratechery · Sep 29

Ben Thompson argues Meta should focus on winning consumer AI agents rather than chasing the enterprise market.

Alicia Wallace, CNN · Sep 29

A McKinsey report estimates AI and automation may push 11 million US workers into new occupations by 2035.

Carl Johnson, Anthropic · Sep 30

Anthropic describes how it deployed a Claude-powered agent to handle inbound sales inquiries and self-serve purchases.

Henri Stern, Stripe · Sep 30

Stripe made the Coinbase- and Visa-backed OUSD stablecoin its default across Treasury, Issuing, and Global Payouts products.

Simon Willison · Sep 29

Simon Willison live-blogged OpenAI's DevDay keynote and product announcements from Fort Mason in San Francisco.

Molly Graham, Lenny's Newsletter · Sep 27

Molly Graham discusses how AI-driven upheaval is reshaping careers and fueling burnout across the tech industry.