The ai models & agents stories our desk judged worth your time, written so anyone can follow them.
Unilever and Nestlé are using AI to enter orders and chase unpaid invoices, but slow-paying customers and tangled systems still slow cash down.
Read more: Unilever and Nestlé Use AI to Handle Orders and Chase Late PaymentsA Honda Civic owner used AI to inspect a cheap Amazon car stereo and found a call recorder and location tracking software already installed.
Read more: Cheap car stereo from Amazon came loaded with hidden extrasWalmart says shoppers will be able to cap what an AI assistant spends and keep card details hidden from it, while Amazon reportedly blocked Meta's Muse agent.
Read more: Walmart plans spending limits for AI shopping assistantsGemini Notebook can use your Google Drive files as sources and answer questions from them, so you can stop digging through folders for one detail.
Read more: Gemini Notebook can turn your Google Drive files into a question-and-answer libraryA 77-year-old caregiver in Arizona says a $3 local ride service replaced $30 to $40 Uber trips and saves her family as much as $400 a month.
Read more: A $3 ride service cut one couple's $500 monthly Uber billCursor, a coding tool, says malware reached up to 5,000 of its users. Details are thin, so anyone who uses it should check their accounts and devices.
Read more: Cursor coding tool says up to 5,000 users hit by malwareA Plaid executive says AI shopping agents may pick payment methods by your goals, not by habit, which could push more people toward paying straight from a bank account.
Read more: AI shopping assistants may pick a different way to pay than you doAmazon's CEO says few people use AI shopping bots because they are no better than visiting a store's site. Amazon has also blocked Meta's new assistant, Muse.
Read more: Amazon's CEO says shopping bots are not worth worrying about yetRo's CEO Zach Reitano talked with Business Insider about how the health company grew from early services to GLP-1 care, and about its new AI agent.
Read more: Ro's CEO explains how the health company built its new AI agentAnthropic's 335-page stock filing promises an age of miracles and warns of existential risk. Anyone thinking of buying AI stocks should know what it leaves out.
Read more: Anthropic asks investors to bet on AI miracles and warns of disasterDoorDash made its AI grocery reorder 8 times faster by doing the heavy thinking ahead of time. Shopping list views doubled and orders rose about 34%.
Read more: DoorDash makes its AI grocery reorder 8 times fasterChatGPT now has a Try On button that shows outfits on a photo of your own body, which could cut online returns but means sharing a picture of yourself.
Read more: ChatGPT Can Now Show How Clothes Look on You Before You BuyDoorDash says AI agents from other companies have placed only several hundred orders, a tiny number next to its millions of deliveries, so the apps still rule.
Read more: DoorDash says AI agents have placed only a few hundred ordersAnthropic says a new self-check inside Claude flagged over 100,000 possible child exploitation incidents from one business customer, and it also exposed a scam operation.
Read more: Anthropic's Claude flagged 100,000+ child abuse cases from one customerOpenAI's CEO says he feels for mathematicians buried in AI results, but the company will keep releasing them and says more are still unreleased.
Read more: Sam Altman says OpenAI will keep releasing AI math resultsTELUS Digital says a voice agent that phones new internet customers cut early cancellations to less than half the usual rate. Here is what it did and what to check.
Read more: TELUS uses an AI voice agent to call new customers and cut cancellationsA fake add-on posing as Claude AI was caught giving hackers a hidden way in. It was live for 11 hours, so anyone who added it should check now.
Read more: Fake Claude add-on gave hackers a way in for 11 hoursGitHub has launched Copilot Fusion, an AI agent tool, as agents now start about 40% of pull requests. Here is what that means for software you pay for.
Read more: GitHub launches Copilot Fusion as AI agents open 40% of code requestsApple has agreed to acquire Ascend.io, a startup whose AI agents handle data prep work, and an EU disclosure rule is how the deal came to light.
Read more: Apple buys a small AI company that automates data workA 32GB kit of DDR5 memory now costs about $500, up from $400. A Tom's Guide writer says AI spending is the cause and may be the thing that brings relief.
Read more: Computer memory prices keep climbing, and the AI boom may be whyA security blog says an Anthropic hack was used to spread hidden remote-control software to Claude developer laptops. It is one report, so details are thin.
Read more: Anthropic hack reportedly spread remote-control malwareAn Android Police writer says built-in Google and Samsung features now do the job of several apps and subscriptions he once paid for.
Read more: A writer says free Android and Gemini features replaced his paid appsResearchers showed a free AI model can turn a public Firefox fix into a working attack in hours, so putting off updates now carries more risk.
Read more: Free AI model hacked six Firefox versions before users got fixesA Bishop Fox researcher used ordinary chat requests to pull an internal Anthropic program and pieces of a Google Cloud token out of Claude's sandbox.
Read more: Researcher gets Claude chatbot to reveal Anthropic's internal software and a key
