Insights archive
Every analysis we have published, newest first. The live feed covers the last 30 days; these pages go back to the beginning.
1504 stories · page 7 of 76
Wrong Records Are Getting Cheap to Fix, Thanks to AI
A Formula One team is using AI to correct car-part records that took about two weeks each to fix by hand, which makes old data mistakes in any business cheaper to clear.
· 3 min read
Open-Weight AI Can't Be Retired by Its Maker
Mistral's release of a trillion-parameter model as a free download matters most for one reason: a model you hold as a file stays the same until you choose to change it, while a rented one is retired on its maker's schedule.
· 3 min read
Whoever Builds an AI Agent at Work Will Own Its Actions
FedEx makes every employee-built AI agent register with the company and trace each action to a named person, which turns the builder of an agent into its accountable owner.
· 2 min read
Security Fixes Can Now Become Working Attacks in Under Two Days
A downloadable AI model built working attacks on six Firefox versions from security fixes that had not yet reached users, so how fast people install updates now decides how safe they are.
· 3 min read
Experts Adopted an AI Agent's Coin-Flip Pick as Their Own
When a coin flip decided which of two sound options an AI agent favored, experts went with its pick 86% of the time and called the result their own, so the real decisions sit in the middle of the work.
· 3 min read
AI Plan Prices Stay Put, but What They Buy Keeps Moving
OpenAI cancelled a planned cut to its $200 ChatGPT plan and doubled the allowance on two others, which shows that an AI plan's price is fixed but the amount of work it buys is not.
· 2 min read
In Gemini, Fixing a File No Longer Means Leaving the Chat
Gemini now lets people change a PDF, Doc or spreadsheet inside the chat by highlighting the part that is wrong, so a small fix becomes a sentence instead of a trip to another app.
· 3 min read
AI Agents Are Good With Facts and Bad at Showing Care
An AI agent saved a film shoot by sending out its schedule, while agent recruiters put off the people they wrote to, because some messages are worth what they say and others are worth the effort they show.
· 3 min read
When Regulators Can Read Everything, Sampling Stops Working
FINRA wants AI to catch more wrongdoing with the same staff, and a regulator that can read every complaint and video no longer has to check a sample.
· 2 min read
A Security AI Nobody Can Switch Off Works for Attackers Too
Mistral's downloadable security model gives defenders an AI that cannot refuse or be cut off, which Hugging Face needed in July, and it gives attackers the same, so fixing flaws faster matters most.
· 3 min read
AI Agents Have No Lasting ID, So Mistakes Can't Be Traced
The most used agent software gives each agent an ID that lasts one run and keeps no record of who approved its work, so a business can't trace a mistake afterward.
· 3 min read
Google's Ads Are Moving Into the AI Conversation
Google is putting search ads inside AI Mode answers in ten more countries, and as with Facebook's move to phones in 2012, the shop that describes its products best is the one that gets seen.
· 2 min read
AI Now Finds Software Bugs Faster Than Anyone Can Fix Them
Google's new Gemini 4 Carbon finds software flaws better than any model tested, but writing and installing fixes is still slow, so how quickly you update now matters more.
· 2 min read
AI Bills Are Turning Into a Price for Each Finished Job
Claude Enterprise can now read a sample of a team's sessions and price each kind of work, which turns an AI bill from one total into a cost per finished job.
· 3 min read
Attackers Now Use AI to Test Every App a Business Connects
AI agents got into Anthropic through the business software it buys from vendors, so any company's weakest point is now the least-watched app it has connected.
· 3 min read
Capital One's AI Runs on Records It Started Organizing in 1988
Capital One's AI cuts a five-hour audit task to three minutes because the bank has kept its records in order for decades, and a smaller business can copy that habit.
· 2 min read
Plant Workers Are Starting to Set the Limits AI Works Within
As plants let software act on its own inside limits that people set, the supervisor's job becomes choosing those limits, and the 1970s control room shows what goes wrong when they are chosen badly.
· 2 min read
AI That Authors Can See Makes Publishers Look Optional
Authors can now do much of what a publisher's marketing staff do with AI, so a publisher that automates the visible work gives authors a reason to question its cut.
· 2 min read
ChatGPT Now Works Inside iCloud Mail With a Revocable Password
ChatGPT can now read, search and send iCloud mail and manage iCloud calendar events, using an Apple password you create for it and can cancel whenever you like.
· 2 min read
AI Is Cutting the Paperwork That Slows Down New Medicines
A drug company's review of its manufacturing records fell from about 26 hours to under three with Claude, and most offices have the same kind of checking work.
· 3 min read