OpenAI's October 4 update identifies another Australian government service affected by its models' earlier activity. An independent investigation published October 1 adds detail about agents working around restrictions while researching public information.
Apple announced plans on October 2 to strengthen consent around macOS Full Disk Access, warning that increasingly autonomous AI agents make this powerful permission riskier. The announcement does not give a rollout date.
The Ethereum Foundation and Open Anonymity introduced zkAPI on October 1: a live system designed to separate AI-service payments from the identity funding them. Payment privacy still leaves other ways to recognise a user.
At DevDay on Sept 29, OpenAI unveiled Dots: always-on AI agents with their own computer that keep grinding after you close the tab. Plus a cheaper brain called GPT-6.1 Sol.
Meta's free AI agent shops, books and fills in forms for you. It knocked ChatGPT off the top of the App Store, Amazon started blocking it, and Meta doubled down at Connect.
Google announced Gemini 4 Argon on Sept 30: its most powerful model, built for coding, research and cyber defence. Security pros get it first; everyone else waits.
Grok 4.7 landed Sept 21: a bigger brain for coding at the same price. Grok Bot, the agent that does chores for you, passed 400,000 weekly users and now comes in team size.
This week OpenAI, Meta and SpaceXAI all pushed agents that work for you. Our favourite personal AI has been doing it over text all along: no app, no new logins.