New York Governor Kathy Hochul Signs RAISE Act to Regulate AI Safety
Governor Kathy Hochul signed the RAISE Act, making New York the second U.S. state to adopt comprehensive AI safety legislation after California. The law obliges large AI developers to publish safety protocols and report incidents to the state within 72 hours, with fines up to $1 million for first‑time violations and $3 million for repeat offenses. A new office within the Department of Financial Services will oversee compliance. While OpenAI and Anthropic praised the move, a super‑PAC backed by Andreessen Horowitz and OpenAI’s president is preparing a legal challenge.
Google's Open Standard Lets AI Agents Build User Interfaces on the Fly
Google unveiled the open‑source A2UI (Agent‑to‑User Interface) standard, letting AI agents generate graphical UI elements on demand via JSON streams instead of HTML or JavaScript. The protocol enables on‑the‑fly forms, buttons, and widgets that render natively within any host app,...
Why the ‘One Chatbot per Child’ Model in Classrooms May Be Flawed
School districts such as Houston and Miami are signing multimillion‑dollar contracts to deploy generative AI chatbots in K‑12 classrooms, driven by cost pressures and a desire to appear innovative. Tech giants and both Republican and Democratic administrations are championing AI...
I Tried the New Claude in Chrome Extension, and It Delivered Convenience with a Side of Digital Paranoia
Anthropic has launched the Claude AI Chrome extension for Pro, Team and Enterprise subscribers, extending the assistant’s reach beyond static text to live web pages. The extension can read browsing history, interact with tabs, schedule meetings, organize Drive files, and...
WIRED Roundup: The 5 Tech and Politics Trends That Shaped 2025
WIRED’s 2025 roundup highlighted five defining trends, beginning with an unprecedented surge in AI data‑center investment as Meta, Google and Microsoft triple spending, driving both economic growth and rising energy costs. The episode warned of a potential AI‑infrastructure bubble, noting...
Businesses Are Hiring AI Specialists Instead of Data Engineers - and Its a Big Problem
Businesses are flooding the market with AI and machine‑learning specialists while neglecting data‑engineer hires, creating a 46% gap in US job postings. RAND research shows more than 80% of AI projects fail, largely due to poor data quality and governance...
Google Releases FunctionGemma: A Tiny Edge Model that Can Control Mobile Devices with Natural Language
Google AI unveiled FunctionGemma, a 270‑million‑parameter model that converts natural‑language commands into executable code on edge devices. Trained on a dedicated Mobile Actions dataset, its function‑calling accuracy climbs to 85%, far surpassing generic small models. The model runs locally on...
Scammers in China Are Using AI-Generated Images to Get Refunds
Chinese e‑commerce platforms are facing a surge in refund fraud as scammers deploy generative‑AI images and videos to fabricate product damage. Reports from sellers on RedNote and Douyin reveal fake photos of dead crabs, torn sheets, and broken ceramics, prompting...
Xiaomi Unveils MiMo-V2-Flash, Its New Open-Weight AI Model to Take on DeepSeek
Xiaomi has launched MiMo‑V2‑Flash, a 309‑billion‑parameter open‑weight AI model designed for complex reasoning, coding, and agentic tasks. The model delivers up to 150 tokens per second and operates at a cost of $0.1 per million input tokens and $0.3 per...
Google Releases FunctionGemma to Bring AI Commands to Smartphones
Google unveiled FunctionGemma, a function‑calling‑optimized variant of the compact Gemma 3 270M model, designed to run directly on Android smartphones. The on‑device AI can translate natural‑language prompts into executable commands, such as creating calendar events or manipulating game elements, demonstrated in the...
We Asked Four AI Coding Agents to Rebuild Minesweeper—The Results Were Explosive
Ars Technica tested four leading AI coding agents—OpenAI Codex, Anthropic Claude Code, Mistral Vibe, and Google Gemini CLI—by asking each to build a web‑based Minesweeper clone with sound and a surprise feature. Codex delivered the most complete game, including the critical...
Beyond Chatbots: How to Build Agentic AI Systems
Google DeepMind’s senior AI relations engineer highlights the shift from chat‑based LLMs to autonomous agents that can pursue goals and invoke tools on their own. Early models acted as text autocompleters, then evolved into instruction‑following chatbots with function‑calling capabilities. The...
AI Is Likely to Replace Jobs, Bank of England Governor Warns
Bank of England Governor Andrew Bailey warned that artificial intelligence will replace certain jobs, drawing a parallel with the Industrial Revolution. He stressed that the technology is unlikely to cause mass unemployment, but will raise the skill bar for entry‑level...
New Research Reveals AI Is Fueling an 'Unprecedented Surge in Cloud Security Risks'
Palo Alto Networks’ State of Cloud Security Report warns that rapid AI adoption is inflating cloud attack surfaces, creating an unprecedented surge in risk. More than 70% of enterprises now run AI‑powered cloud services, often granting overly permissive access. Identity‑related...
Wall Street Recovers From Losses Earlier in the Week, as AI Stocks Climb
Wall Street rebounded on Friday, with the S&P 500 up 0.8% and the Nasdaq gaining 1%, trimming earlier losses. AI‑focused technology stocks led the charge, highlighted by Nvidia’s 3.4% rise and Broadcom’s 2.4% gain. Oracle’s 7% jump followed its agreement to...
YouTube Pulls The Plug On Popular AI-Generated Fake Movie Trailer Channels
YouTube permanently removed the Screen Culture and KH Studio channels after they repeatedly breached spam and misleading metadata rules by posting AI‑generated fake movie trailers that amassed over 2 million subscribers and a billion views. The platforms had previously been demonetized,...
Develop Health Integrates Athenahealth for Prior Authorization
Develop Health announced a native integration of its AI‑driven prior‑authorization platform with athenahealth’s athenaOne EHR. The solution runs automatically when clinicians prescribe, handling benefit checks, form completion and status tracking without extra clicks. Early adopters report up to a 70%...
Kamiwaza Brings Its Enterprise-Grade AI Orchestration Platform to NVIDIA DGX Spark
Kamiwaza AI released version 0.8.0, adding native support for NVIDIA DGX Spark and a new Two‑Node Community mode that links two Spark systems into a single compute fabric. The update automates model parallelism, leverages unified memory over 256 GB, and integrates...
AI Is Solving 'Impossible' Math Problems. Can It Best the World's Top Mathematicians?
Recent breakthroughs show AI models from Meta and DeepMind making measurable progress on hard mathematical problems, such as finding Lyapunov functions for 10.1% of test cases and solving most of the 2024 International Mathematical Olympiad questions. Yet leading mathematicians stress...
A Robot Just Learned 1,000 Tasks in a Single Day — and It’s a Big Deal for Everyday AI
Researchers at ScienceRobotics taught a physical robot arm to acquire 1,000 distinct everyday tasks in under 24 hours, using only a single demonstration per task. The breakthrough relies on a novel learning framework that decomposes actions into reusable sub‑phases, allowing...
Master PDFs Like a Pro with 30% Off LightPDF
LightPDF, an AI‑powered all‑in‑one PDF editor, is offering a 30% discount on both its Annual and Perpetual licenses via the promo code techradar30. The Annual plan drops from $89.99 to $34.99, granting 12,000 AI credits per year, while the Perpetual...
50,000 Copilot Licences for Indian Service Companies
Cognizant, TCS, Infosys and Wipro announced they will each deploy over 50,000 Microsoft 365 Copilot licences, totaling more than 200,000 seats across their enterprises. The rollout, unveiled in Bengaluru alongside Satya Nadella’s India visit, frames Copilot as a default productivity tool...
OpenAI Brings Cheaper Subscription Tier "Go" To More Markets
OpenAI is rolling out its low‑cost ChatGPT Go subscription to more than 70 additional countries, extending the tier that debuted in India earlier this year. In Germany the plan is priced at €8 per month and now includes image generation, file...
Identifying Patterns in Insect Scents Using Machine Learning
The episode explores a University of Amsterdam project that uses machine learning to decode how insect olfactory receptors bind to scent molecules, aiming to create a large, shared database of 25,000+ scent-receptor interactions. Researchers from biology, mathematics, data science, and...
Pinned Chats in ChatGPT Are Here – and so Is a Mildly Annoying Restriction
OpenAI has added a pinned‑chats feature to ChatGPT, allowing users to keep up to three conversation threads at the top of the sidebar on both web and mobile apps. The option is accessible across Free, Plus, and Pro tiers, making...
OpenAI Updates Codex Model, Adds Trusted Access Program for Cyber Defense
OpenAI unveiled GPT-5.2-Codex, an autonomous software-agent model that adds context compression and improved image processing. Benchmarks show modest accuracy gains—56.4 % on SWE-Bench Pro and 64 % on Terminal-Bench 2.0—over the standard version. The company also launched a trusted access program, allowing vetted...
Zara’s Use of AI Shows How Retail Workflows Are Quietly Changing
Zara is piloting generative AI to expand existing model photographs into multiple product‑image variations, aiming to cut the time and expense of repeat photoshoots. The technology is embedded directly within the retailer’s current production pipeline, allowing marketers to generate regional...
IDBS Celebrates Customer Successes and Product Innovations in 2025
IDBS wrapped 2025 by unveiling Polar Control, a GxP‑compliant recipe‑book tool that modernizes process lifecycle and tech transfer. The company also launched a cloud‑native integration framework and expanded its Polar accelerator library with real‑time insight apps for bioanalysis, DMPK, stability,...
DCAI Achieves ISO/IEC 27001 Certification
DCAI, the operator of the Gefion AI Supercomputer, has secured ISO/IEC 27001 certification, confirming its information‑security management system meets global standards. The audit also validated the company’s broader ISAE 3000 cybersecurity and privacy controls. Certification spans data‑center operations, infrastructure, and internal processes,...
Meta Preps "Mango" And "Avocado" AI Models for 2026
Meta is preparing two new AI models, codenamed “Mango” and “Avocado,” slated for release in the first half of 2026. Mango will specialize in image and video generation, while Avocado is a language model optimized for programming tasks. The projects...
AI’s Butterfly Effect: The Danger of Cascade Failures
The article warns that as AI systems become tightly interwoven across enterprises, small glitches can trigger cascade failures that ripple organization‑wide. Interconnected models, shared data pipelines, and common compute resources amplify localized errors into systemic disruptions. The 2010 Flash Crash...
S3NS Announces SecNumCloud Qualification for PREMI3NS
S3NS announced that its PREMI3NS trusted‑cloud platform has achieved the ANSSI SecNumCloud 3.2 qualification, the strictest security standard in France and Europe. The certification guarantees protection against non‑European extraterritorial laws and positions PREMI3NS as the most feature‑rich SecNumCloud‑certified offering. Built on...
b.well Launches the First-Ever SDK for Health AI Assistants
b.well Connected Health unveiled the first SDK designed to power healthcare AI assistants with real‑time, unified medical data. The platform aggregates information from 2.2 million providers, over 300 health plans, and a broad ecosystem of pharmacies, labs, devices, and patient‑generated sources....
Align Continues Strategic Expansion to Accelerate AI and HPC Data Center Deployments
Align is accelerating its U.S. expansion to meet soaring AI and high‑performance computing (HPC) data‑center demand. The company is bolstering vendor partnerships to shorten time‑to‑token for clients and has grown its internal Data Center Solutions team. Align appointed Keith Hopkins,...
BigBear.ai and C Speed Announce Strategic Partnership to Deliver AI-Enabled Border Security
BigBear.ai and C Speed announced a strategic partnership to fuse BigBear.ai’s ConductorOS AI orchestration platform with C Speed’s software‑defined LightWave Radar. The integration will deliver autonomous, real‑time threat detection and decision support for border security, air‑missile defense, and counter‑UAS missions....
Proposal to Allow Use of Australian Copyrighted Material to Train AI Abandoned After Backlash
The Australian Productivity Commission has scrapped its plan to grant a “fair dealing” exemption that would let tech firms mine copyrighted Australian material for AI training, after a fierce backlash from the creative sector. Instead, the commission advised the government...
Microsoft AI CEO Admits Gemini 'Can Do Things that Copilot Can’t Do'
Microsoft AI chief Mustafa Suleyman publicly acknowledged that Google’s Gemini 3 can perform tasks that Microsoft’s Copilot cannot, a rare admission in the AI arms race. He highlighted Gemini’s multimodal strengths while emphasizing Copilot’s real‑time visual assistance and integration across Windows,...
Unsloth AI and NVIDIA Are Revolutionizing Local LLM Fine-Tuning: From RTX Desktops to DGX Spark
Unsloth, a GPU‑optimized fine‑tuning library, now runs on everything from GeForce RTX desktops to NVIDIA's DGX Spark, delivering up to 2.5× faster training for LLMs. The tool supports parameter‑efficient methods like LoRA/QLoRA, full model updates, and reinforcement‑learning pipelines, each with clear...
Oracle Has One Last Ampere Hurrah - New Cloud Platforms Offer up to 192 Custom Arm Cores, and Aren't OCI...
Oracle Cloud Infrastructure launched A4 Standard instances powered by Ampere Computing’s AmpereOne M silicon, offering up to 192 custom Arm cores per chip in both virtualized and bare‑metal forms. The configurations deliver up to 48 OCPUs (96 cores) with up to...
I Tried the Apple Music App for ChatGPT, and It Felt Like Talking to a Music‑obsessed Friend
Apple Music has launched a ChatGPT app that lets users generate playlists, discover songs, and play short music clips via conversational prompts. Subscribers can link their Apple Music accounts to save the AI‑crafted playlists directly to their library. The integration...
YouTube Bans Two Popular Channels that Created Fake AI Movie Trailers
YouTube shut down the Screen Culture and KH Studio channels, which together amassed over 2 million subscribers by posting AI‑generated movie trailers that mimicked real releases. The platforms initially demonetized the creators for missing disclaimer tags, briefly reinstated them, and then...
It's Not Just RAM Getting More Expensive - the Tools to Make Chips Are Set to Explode in Cost Too,...
Global semiconductor manufacturing equipment sales are projected to reach $133 billion in 2025, rising 13.7% year‑over‑year, with forecasts of $145 billion in 2026 and $156 billion in 2027. The surge is driven by heightened demand for AI accelerators, high‑performance computing and advanced 2 nm...
School Security AI Flagged Clarinet as a Gun. Exec Says It Wasn’t an Error.
A Florida middle school was placed under lockdown after ZeroEyes, an AI‑driven security system, mistakenly identified a student’s clarinet as a rifle. Human reviewers failed to halt the alert, prompting police to respond before discovering the false alarm. ZeroEyes’ co‑founder...
Pickle Robot Adds Tesla Veteran as First CFO
Pickle Robot, a Massachusetts‑based maker of autonomous unloading robots, appointed Jeff Evanson as its first chief financial officer. Evanson, a former Tesla vice president of global investor relations, had been consulting for Pickle since September and now joins full‑time. The...
T5 Services Scales Rapidly as Data Center Delivery Demand Grows
T5 Services, the services arm of T5 Data Centers, reported record growth in 2025, expanding revenue from $87 million in 2021 to $1.6 billion. The company’s integrated construction and operations model, backed by more than 200 professionals and 6,000 subcontractors, delivered over...
GPT-5.2 Tops OpenAI's New FrontierScience Test but Struggles with Real Research Problems
OpenAI introduced the FrontierScience benchmark, a two‑part test featuring Olympiad‑level problems and open‑ended PhD‑level research challenges. GPT‑5.2 topped the leaderboard, reaching 77% accuracy on the Olympiad set and 25% on the research set, while Gemini 3 Pro trailed closely on Olympiad...
Palona Goes Vertical, Launching Vision, Workflow Features: 4 Key Lessons for AI Builders
Palona AI, founded by former Google and Meta engineers, announced a vertical shift into the restaurant and hospitality sector with two new products—Palona Vision and Palona Workflow. Vision leverages in‑store security cameras to monitor queue lengths, table turnover, and kitchen...
'A Lot of People Cry Using the App — We’ve Had Users Say It Put Into Words Things They’d Been...
Mark Manson and AI entrepreneur Raj Singh have launched Purpose, an AI‑mentor app that delivers personalized self‑help coaching directly on users' phones. The platform tailors advice to individual values, fears and goals while explicitly refusing therapy‑level issues and directing users...
Anthropic Launches Skills Open Standard for Claude
Anthropic announced an open‑source Agent Skills specification for its Claude model, expanding the platform’s ability to create, share, and deploy task‑specific capabilities. The update adds organization‑wide management for enterprise teams and bundles pre‑built Skills from partners such as Canva, Notion,...
The Ultra-Realistic AI Face Swapping Platform Driving Romance Scams
Haotian, a Cambodia‑based AI face‑swapping platform, is earning millions by selling ultra‑realistic deep‑fake software to scammers via Telegram. The service offers 50 customizable facial parameters and voice‑cloning, enabling fraudsters to conduct convincing video‑chat scams, especially pig‑butchering schemes. Cryptocurrency tracing shows...