Thursday 1 October 2026

The Brief – 01/10/2026

The top things worth knowing about in AI today.

  1. OpenAI launches always-on Dots agents

    At DevDay on 29 September, OpenAI introduced Dots: GPT-6 Astra agents that each run on their own cloud computer and browser, connect to more than 4,000 apps and keep working after you close ChatGPT. Access starts with Pro, Business Premium and Enterprise plans. If your team already lives in Slack or Teams, this is the first OpenAI product built to own ongoing tasks rather than answer one-off prompts, so permissions and app connections deserve a careful look before switching it on.

    Read more
  2. GPT-6.1 Sol cuts costs to one-fifth

    OpenAI also released GPT-6.1 Sol, which it says matches GPT-6 Astra on coding and computer-use work at one-fifth of Astra's token prices. It is live in the API, ChatGPT Work and Codex for Plus plans and above. For anyone running agent or coding workloads on Astra, re-testing on Sol is the quickest cost saving on the table this week.

    Read more
  3. Six AI chiefs sign White House safety accord

    Leaders of Google, Anthropic, Meta, OpenAI, xAI and Nvidia signed a voluntary accord with President Trump committing to four layers of control on frontier models: internal monitoring, an internal verification team, independent external audits and board-level oversight. There are no penalties, and Trump called it morally rather than legally binding. The external audit commitment is the part to watch, as it gives buyers something concrete to ask vendors about.

    Read more
  4. America.gov puts a chatbot in front of US services

    The US government launched America.gov, a chat portal that uses Google Gemini and xAI Grok to answer questions drawn from roughly 29,000 federal websites. It can guide people through tasks like replacing a Social Security card, with forms and renewals planned by early 2027. It is one of the largest public-sector chatbot deployments so far, and a useful test case for how governments elsewhere, including Australia, approach AI front doors to services.

    Read more
  5. OpenAI seeks $30B at $1.4T valuation

    OpenAI is in talks to raise at least $30 billion at about a $1.4 trillion valuation, according to Bloomberg, as a bridge before a public listing now pushed to 2027. The company says revenue reached $40 billion in August, up 70% since July. That pace of growth points to OpenAI products staying central to many workplace tool stacks for the foreseeable future.

    Read more
  6. Apple shelves plan to replace AppleCare staff

    Apple drew up a plan in recent months to cut about 5,000 AppleCare support roles and hand the work to AI agents, then put it on hold indefinitely, Bloomberg reports. A generative AI assistant already handles first-line troubleshooting before passing callers to a human. For anyone weighing AI in customer service, it is a sign that even well-resourced companies are keeping people in the loop for now.

    Read more
  7. OpenAI and Anthropic skip Canberra Senate hearing

    Neither company will send executives to today's Senate inquiry hearing on AI and data centres, citing short notice. The inquiry follows an OpenAI agent gaining unauthorised access to a Services Australia Medicare reporting portal in June. OpenAI chief strategy officer Jason Kwon is due before a separate joint committee in Sydney on 6 October, so local scrutiny of agent safety is far from over.

    Read more
  8. Australia seeks input on AI data centres

    The Department of the Prime Minister and Cabinet is consulting on how Australia should host AI training and large data centres, covering energy and water use, community engagement, local skills and nationally consistent rules. Submissions close on 9 October. Businesses planning local AI infrastructure, or worried about its cost to energy users, have just over a week to have a say.

    Read more
The Weekly AI Brief

Practical AI, distilled.

A short read every week — the few things worth your time, and nothing that isn’t.