01
Products
AI shopping agents run into retailers that block them
Access is inconsistent and cart reservations can hold stock away from other buyers, so agents may not complete purchases reliably.
Access is inconsistent and cart reservations can hold stock away from other buyers, so agents may not complete purchases reliably.
Households and businesses that cut power use at peak demand will be paid, easing strain on a grid serving 67 million people.
Native PyTorch support for TPUs lowers the barrier to running production AI workloads on hardware outside Nvidia's ecosystem.
The open-weight model runs on a laptop and needs more guidance than large paid coding assistants.
Enterprise customers gain a second major cloud route to OpenAI's frontier models, while OpenAI keeps Microsoft as its main cloud partner.
Software that depends on web PKI but cannot filter intermediates stays exposed until its updates arrive.
Affected servers need patches for a PHP bug, and some certificates from one authority may still be trusted by a macOS utility.
Mixed-kernel Kubernetes fleets can now share one pod datapath setting, while the older Mutual Authentication feature moves toward removal.
Browsers and operating systems now reject the affected intermediates, but some software with its own trust store may keep accepting them until it is patched.
Drivers can now ask Android Auto for answers drawn from their connected Google data while keeping the feature optional.
The warning adds a safety check for Enhanced Protection users, with different prompts depending on how confident Chrome is about the risk.
The new business-focused experience ties planning, marketing, finance and app connections together in one place for Gemini users.
Google says the research system could collect history upfront and leave doctors with a clearer handoff and more time for care decisions.
The filings could strengthen claims that the company knew about risks to minors while families were not fully warned.
The 2.25 GW return adds new power for AI-linked data demand, with one plant already moving toward 2027 output.
Owners on a limited mobile data plan could run through their allowance without realising a speaker was the cause.
The tool lowers the shift from Nvidia GPUs to Amazon’s chips by changing configuration instead of application code.
JDK 28 work moved ahead, while Payara and Apache Camel each released updates that close security gaps for users.
The brief tampering windows mean organizations that checked for updates during that period may have been exposed without seeing obvious warning signs.
Some users faced unauthorized charges after stored keys were exposed, and OpenCode says refunds are now underway.
Users can now shape the outline and style before Notebook renders the final slide deck or video.
It gives developers one embedding model for multiple media types, with smaller and larger configurations for different device limits.
The result points to agent systems that can stay coordinated through long tasks without custom game code.
The arrangement puts an outside evaluator inside Anthropic’s office as pressure grows for stronger independent oversight of advanced AI.
Local limits are spreading, forcing autonomous-vehicle operators to contend with rules that differ city by city.
The framework keeps the task and evaluator fixed while changing what the agent sees, so weak spots can be targeted without rewriting the benchmark.
The company can now map Toronto and move toward public-road testing before any rides are opened to passengers.
The new options give students faster ways to review material and a wider range of quiz formats.
The sponsored listing can steer users to an impostor before they reach the official messenger, and the clone may also be limiting public review warnings.
UTC will drift away from Earth’s rotation for now, reducing the risk of future timing-related outages and avoiding a leap second adjustment path the BIPM says is too fragile.
The new search tool can pull details from emails directly, and the same rollout adds voice-based features for Docs and Keep.
New users will need a paid plan to keep automatic transfers to other apps after a one-month trial.
Google is pushing more clickable sources into AI Overviews and AI Mode, which should make it easier to reach publishers from search results.
The result points to a more practical path for building reliable quantum systems by reducing the error rate in a core part of the stack.
The changes extend early warnings and live disaster maps to more people, giving communities more time to respond before danger spreads.
Google is surfacing sources, pages and news more visibly inside AI Search, making it easier to move from answers to the original web pages.
The finding gives the first technical proof of government-backed zero-click spyware delivered through ad systems and shows the method has already reached high-value targets.
Strong AI-related orders are keeping ASML booked far ahead, while higher prices are set to reach customers from 2027.
The listing would give Jio a public market valuation near ₹10.3 trillion and direct up to ₹275 billion toward debt repayment.
An opt-in protection mode will scan unknown senders and links, while Google extends scam detection beyond Pixel to more Android devices.
The takedown cuts off a large source of scam traffic that had reached Microsoft customers and other organizations across more than 150 countries.
The trojan can replace bank apps with lookalikes, letting attackers bypass protections and capture credentials.
The update keeps more code on-device while shifting only a small part of agent work to the cloud.
The package is meant to help selected researchers do scientific work faster by easing access to models and compute.
Shoppers can now let more AI agents browse and buy from Shopify’s catalog, with checkout kept inside Shop and each purchase requiring approval.
U.S. adults can now keep AI-made Android apps private in Play first, then share them or publish them later.
The update gives some users new ways to make Android apps, while search and gaming features aim to make Play more useful across mobile and PC.
The update gives advertisers more ways to track and improve campaign data quality across Google Ads, DV360 and Google Analytics.
Access is limited to paid users for now, and developers get the model through the Gemini API as Google prices the launch tier and raises rates after November 30.
The fee change will not start before Diwali, leaving merchants with the current free model in place for now.
The work now spans faster diagnosis, mammography screening, genomic analysis and disease risk mapping, with results Google says can shorten clinical and public health workflows.
The company is widening the shared hardware and software base for large AI infrastructure, with new MGX and DSX pieces aimed at faster deployment.
The company is tightening safeguards and rollback plans as faster change rates expose more production risk.
The campaign uses trusted ad and redirect paths to reach macOS users searching for Claude, making the fake installer harder to spot before the malicious command runs.
Google has changed how Android AI work is measured, replacing simple pass/fail checks with scoring for more complex development tasks.
Ikotas Labs won the contest, while payouts climbed across all three days as teams found more flaws.
The trial suggests a patient-facing chatbot can help clinicians prepare for visits without safety stops, but Google says larger studies are still needed.
Projects that opt in can now get periodic scans without paying, but the results are model-only and may include mistakes.
The campaign reaches beyond infected handsets, with malware-linked apps in Google Play and live proxy infrastructure still taking registrations.
Owners will need a paid mobile plan for those features to work over cellular after the cutoff period ends.
The new tablets bring Play Store access and end Amazon’s push to keep them inside its own app ecosystem.
The order pauses major construction while environmental review continues, limiting site work at Muhos and Kajaani.
Foresight keeps meeting audio, notes and files on the computer, giving Mac users an offline option for AI note capture and follow-up questions.
The agent starts in private preview for enterprise customers and is built to move across workplace systems with its own account and audit trail.
The company’s latest forecast points to another record quarter, underlining how AI data-centre demand is feeding through the chip supply chain.
Google moved to block the certificates in Chrome after the hijack, limiting the risk for users of affected sites.
The update lets Webex teams bring AI agents into meetings and workflows, with Cisco also tying in Splunk for safety controls.
The new site expands provenance checks beyond Google-generated content and adds partner marks from OpenAI, NVIDIA, Kakao and soon Apple.
The new tool lets adults in the US build games from prompts without coding, with higher usage limits for Google One subscribers.
The new release lets teams run Spanner outside Google Cloud, but it does not include an availability SLA.
The system is designed to let health workers gather key pregnancy information without a sonographer or reliable power and internet.
Producers can now move no-code tools they build in Flow Music into the DAWs they already use.
The campaign uses lookalike AI pages to capture logins and multi-factor prompts, putting advertising accounts and the data tied to them at risk.
The model is built for local search and retrieval, letting more data stay on device.
Free users will lose Flash and Pro, while some higher-end features move behind paid tiers.
The agreement adds 890 megawatts and is intended to keep existing plants running while backing a larger power supply for Google’s data-center demand.
The September 28 roundup shows platform work moving from proposals and incubation into scheduled releases and project status changes.
The move ends Claude’s use inside a Pentagon system for intelligence and decision support, after weeks of overlap with the phaseout plan.
The pause puts Google’s planned sites in Muhos and Kajaani under closer review and could delay how the company advances its Finland build-out.
The fix closes a storage-layer leak in Workers Paid containers and shows that isolation still depends on what happens beneath microVM boundaries.
Apps can now check patch status by component, which gives security-sensitive software a clearer view before it allows risky actions.
The group must report within 120 days, while the administration shifts official language to “SI” and “Super Intelligence.”
Google and Anthropic are pushing legal AI into practice, while recent sanctions and research show citation errors still carry real risk.
The new feature is meant to cut repeated prompting, and Google will move existing Gems into skills over time.
SynthID Bio is meant to let synthetic biology outputs be tracked without changing how the proteins work.
The preview could let Gemini act across files and apps on macOS, while still asking before sensitive actions.
Microsoft says the new setup cuts the setup step for isolated containers and is now available in more than 40 Azure regions.
The CDC’s review puts Google’s model at the front of flu demand forecasting for the 2025-26 season.
The freeze limits new filings in one program while Google says pre-existing and supply-chain reports can still be handled.
The deal puts major AI labs in charge of their own safeguards, with outside audits and board oversight instead of immediate regulation.
The funding backs software meant to let satellites share compute instead of each carrying its own AI hardware.
Google has confirmed contact with the satellite as it starts gathering data on whether machine learning infrastructure can operate in space.
Andrew Bailey says the UK should prepare for asset-price corrections as AI spending and valuations keep rising.
The ruling leaves Google’s AI search features standing, while publishers cannot claim expected traffic as a legal right.
The new frontier model starts in a limited rollout while Google tests it with security teams before widening access.
Google is limiting early access to a model built for high-risk work before opening it more widely later.
The real-time visual assistance feature brings spoken scene descriptions and cues to more blind and low-vision users on supported Android phones.
Connected vehicles and their apps are sending personal and location data to third parties, raising legal and privacy pressure on automakers and vendors.
The prospectus shows a swing to profit alongside massive planned outlays and unusually direct disclosures about AI risks.
New demos show Gemini 3.8 Flash handling live visualizations, art, structured prompts and interactive simulations at near frontier‑model performance.
Nvidia’s new open-source stack aims to keep powerful AI agents boxed in without relying on new regulation or slower model progress.
Enterprises and schools now face a timed shift from ChromeOS to Googlebook OS, with key details still missing.
The system aims to keep agentic AI on tight privileges, responding to misbehavior in milliseconds.
New internal guidance sets expectations on how staff should pick models, review outputs and use AI efficiently in daily work.
The change is intended to speed responses, but employees will review proposed removals before records are released.
A forged token let the researcher reach internal employee and Bing analytics data before Microsoft fixed the flaw via its bug bounty program.
Teams can discover compatible software in one directory, while partners gain a route to Claude customers.
Faster resumes still leave applications responsible for recreating pre-snapshot encryption keys and certificates.
For unpatched organizations, a WAF rule matching the usual route is no substitute for Oracle’s security update.
If enacted, it would extend potential blocking orders to VPN and DNS services, not just ISPs.
People who call the displayed support number may be pressured to pay, grant remote access or share personal information.
New YouTube, Gmail and Maps features aim to shorten the path from ad impression to local or online conversion.
By pushing state to workers and parallel schedulers, Modal shows how far large-scale sandboxing can go without traditional Kubernetes coordination.
The compact open‑weight image model aims to rival larger systems while shifting to a more restrictive license for commercial use.
Gemini’s new calling test can now complete tasks like bookings and rescheduling, not just collect information.
The funding backs Ema’s push to replace parts of traditional SaaS and IT services with task-based AI automation.
Planned rules would formalise search-choice prompts, extend them to AI assistants, and tighten how publisher content is shown in results.
A single YAML in a watched namespace can turn limited Kubernetes access into full Google Cloud organization control when KCC is over‑privileged.
New AI workflow tools, unified updates, XR fleet support and added security controls aim to tighten how companies run Android at scale.
Open‑weight medical AI models are moving into real‑world screening and triage, from rural clinics to national health programs.
The flight will put AI hardware in orbit for a real-world test, but a useful operational system remains years away.
Sensor-signed captures and Apple-only signatures could reshape how image authenticity is verified and who controls that process.
Ukraine’s cyber teams will use OpenAI’s Daybreak system to spot flaws and test fixes faster across critical services under heavy digital attack.
The discovery of the ART enzyme system highlights Claude’s emerging role in real-world biological research, even as its function remains unclear.
New voice models and Live Avatar aim to make Gemini’s spoken interactions more natural, multilingual and traceable.
Gemini users can now hook in a wider range of work, creative and lifestyle services directly from chat.
Custom feeds, group chats and new AI tools reshape how viewers watch and creators produce on YouTube.
New funding and technical support aim to expand AI-based climate and crop advice across the Global South.
The new top-tier Claude model aims to match larger systems’ performance while cutting prices and reducing risky behavior.
The implant’s AI-led automation could speed up attacks and let them unfold at any time after infection.
The ruling forces Google to fix how it handles location data in the EU within six months, tightening scrutiny of its data practices.
The brief pushes AI safety up the diplomatic agenda, calling for stronger safeguards and coordination despite differing national laws.
The inside access let Google track the campaign, warn targets and feed police data that preceded two arrests in Australia.
Families can now share a single CC account to coordinate calendars, tasks and everyday logistics.
The attack turns trusted AI assistants into high‑privilege proxies once a rogue Chrome‑style extension is installed.
The incidents expose weaknesses in Google’s testing setup and raise fresh scrutiny of how it discloses AI security failures.
Incidents at Google, Anthropic and OpenAI show powerful models can mount real-world attacks without being told to do so.