The Daily Brief archive
Inference Wire publishes a five-minute AI intelligence brief every weekday morning. Read the latest edition, or work back through every edition we have published.
All editions
Newest first · page 1 of 5
ChatGPT can now read and use Chrome tabs where staff are already signed in
Meta scaled back plans to cut staff after its people spent more and more time fixing AI agents' errors, and Shopify shoppers can now connect their own agents.
Business software is folding AI into its flat price, and renewal is where buyers can push back
AOL's $19.95 flat rate in 1996 forced it to more than double its modems, and AI bundles are answering heavy use with monthly allowances.
ChatGPT repeats what a company's old web pages say, and fixing those pages fixes its answers
Workday will let payroll and accounting agents run in the background from early 2027, and the pauses that give away AI help in job interviews will fade.
Google's Gemini Spark reads each email and acts on it, from a rule written in one sentence
Spark runs on personal Google accounts and only with a setting that lets Google train on tasks, so customer emails should wait.
Top AI models do the steps well but miss when their own results are broken
Amazon blocked Meta's Muse shopping agent in September, and literary agents spotted publishers' AI-written letters by comparing them.
Gemini agents get company accounts, and staff-built agents need named owners
Google's free Foresight app writes meeting notes on a Mac without the internet, and AI that scores every customer call needs a clear rule on what a low score leads to.
ChatGPT answers with working calculators, and Claude starts editing inside Google Docs
Copilot will file documents using models on the PC itself, and an agent that is 95% right at each step finishes a 20-step customer case correctly only 36% of the time.
Free Gemini loses its stronger models, and hackers break into Korean lenders through side doors
Executives are 2.7 times as likely as their staff to use AI agents, and an agent given a target will chase it without knowing what it was for.
AI numbers need agreed definitions, and AI courses need someone to redesign the work
Crisis rehearsals work best when a real supplier disappears, and HackerRank's AI interviewer asks developers how they checked the AI.
AI that argues back improves your work, and AI beats accountants at routine month-end tasks
Finance teams that can show auditors how their AI reached its numbers report far fewer errors, and ERP agents act on whatever the records say.
Most company AI money is spent outside IT, and ServiceNow sells a help desk that starts in a day
Google gives its new Gemini 4 Argon model to cyber defenders first, and coding agents put 13,000 company screenshots on public GitHub pages.
AI shopping assistants choose sellers on delivery and returns, and California limits AI firings
Google turns Gemini's Gems into Skills, the format Anthropic created, and spending on supply chain software with AI agents heads for $53 billion by 2030.
Meta's Muse agent connects to small firms' books and ads, and OpenAI's Dots keep working in Slack
Bot-free meeting notetakers leave recording consent to the user, and AI is pushing law firms and agencies to cut prices and sell subscriptions.
Managers check AI less carefully when it is called a coworker, and Shopify lets AI agents pay
Only 23 companies spread an AI gain from their top factories to three or more plants, and Gartner expects over 40% of agent projects to be canceled by 2027.
Siri runs on Google's Gemini in iOS 27, and Microsoft splits Copilot into two bills
OpenAI cannot notify the 53 ChatGPT users whose photos its agents posted online, and Bank of America has 4,000 fewer staff with no layoffs.
Google links 14 more outside apps into Gemini, from monday.com to Experian
Ten of the new apps work only in the US, and disconnecting one leaves what Gemini saved in the user's activity history.
ChatGPT Voice now acts on email and calendars, and AI users carry the same workload as non-users
One CMO mapped more than 700 daily marketing tasks and cut a 50-handoff campaign launch to 11 steps with AI coordinating.
Rabbit's new agent operates computers people already own, and Microsoft shuts down EvilTokens
GPT-6 Sol now costs $2 per million input tokens, and Gartner expects 150,000 AI agents inside the average Fortune 500 company by 2028.
Employee-built AI tools are a leading breach cause, and one hijack works across major agents
Citizens Advice wants a guaranteed route to a person, and Gartner expects more than 40% of agentic AI projects to be canceled by the end of 2027.
Meta's Muse acts inside Mac Mail and Files, and a $200 AI subscription opened OpenAI accounts
A Munich court ruled Google is liable for what its AI writes, and Alibaba's new model reads an hour of audio for less than half a cent.
THE DAILY BRIEF
Read the next one first.
A five-minute read, every weekday morning.
Free forever · Unsubscribe anytime