AI news for Monday, 19 October 2026
Stories 61 to 120 of 170, most important first. Back to the first page →
-
Seven expressions humans correct in AI-written Google review replies
A shop owner who runs 16 AI assistants inside Claude Code shares the seven phrasing mistakes he repeatedly fixes in AI‑generated replies to Google reviews. The guide shows why emojis, excess exclamation marks, vague…
1 sourcenote.com -
Agility Robotics launches Digit 5 humanoid robot worker
Agility Robotics announced the launch of Digit 5, its newest humanoid robot designed for warehouse work. Weighing over 120 kg, Digit 5 relies on a large battery and cooling system that occupies about 75 % of its torso,…
1 sourceIEEE Spectrum Robotics -
Google says its 780‑MW AI data center will use renewable power in Minnesota
Google announced plans for a 780‑megawatt AI data center in Pine Island, Minnesota, and a partnership with Xcel Energy that would finance 1,400 MW of wind, 200 MW of solar and 300 MW of battery storage – more capacity…
2 sourcesfinance.yahoo.comCNBC Technology -
Amazon Bedrock adds structured JSON outputs for Anthropic Claude models
Amazon Bedrock now supports structured outputs for every Anthropic Claude model on both the bedrock‑runtime and bedrock‑mantle endpoints. When a request includes a JSON schema, the service validates the schema,…
1 source primary sourceAWS Machine Learning Blog -
Together AI launches post-training service for custom models with $9 GPU-hour pricing
Together AI announced the Together Post-Training service on October 19, 2026. The platform lets customers build custom models using reinforcement‑learning (RL) or supervised fine‑tuning (SFT) loops that they define,…
1 sourceUnite.AI -
Microsoft pulls out of Nscale's Essex AI data centre
Nscale announced last year a £2bn pledge to build the Loughton facility in Essex, billed as the UK’s largest AI data centre. The project was to host a Microsoft‑funded supercomputer and was slated to receive power by…
1 sourcebbc.com -
OpenAI brings back Aleksander Madry to focus on AI safety
OpenAI announced on October 19, 2026 that Aleksander Madry is rejoining the company after leaving in May 2026. Madry had spent nearly three years at OpenAI before his May departure, most recently leading the…
2 sourcesfinance.yahoo.comcryptobriefing.com -
Zillow adds Claude connector for listings, tours and financing
Zillow announced on October 19, 2026 that its real‑estate data is now available inside Anthropic’s Claude assistant through a new connector. Senior vice president of AI Josh Weisberg said the integration pulls two…
1 sourceUnite.AI -
Google adds one-click export of Gemini interactive apps to Google Sites
Google announced that users of the Gemini app can now export interactive apps directly to Google Sites with a single click. The export creates a new private site draft that is fully interactive and ready for further…
1 sourceGoogle Workspace Updates -
Anthropic outlines third-party AI evaluation limits in new safety framework
Anthropic announced on October 15, 2026, that it granted six staff members from METR, a non-profit AI risk research group, access to internal systems and equipment comparable to its own risk assessment team. The scope…
1 sourcenote.com -
0G reports 13 providers processing more than 373 billion tokens and 27 million inferences
Michael Heinrich, co‑founder and CEO of 0G, describes a decentralized AI infrastructure stack that combines verifiable compute, encrypted storage, private inference and an AI‑native EVM‑compatible Layer 1 blockchain.…
1 sourceUnite.AI -
AICPA and AIUC launch joint SOC 2 and AIUC-1 certification for AI systems
The American Institute of CPAs (AICPA) and the Artificial Intelligence Underwriting Company (AIUC) announced a new assurance offering that merges a SOC 2 report with AIUC‑1 certification into a single CPA‑led…
1 sourceUnite.AI -
Atlassian launches Agentic Multiplayer Protocol to tag AI agents with user attribution
Atlassian introduced its Agentic Multiplayer Protocol on Oct 6, a control plane that makes AI agents appear as governed members of work teams. The system assigns each agent a distinct profile and owner, so edits by an…
1 sourcefortune.com -
Human Technology Institute calls for stronger workplace privacy laws in Australia
The University of Technology Sydney’s Human Technology Institute warned that AI‑enabled monitoring tools are spreading across Australian workplaces, from fast‑food kitchens to retail checkout lines. Examples include…
1 sourceThe Guardian AI -
Anthropic, Google and OpenAI prepare new AI models, report says
The Information reports that Anthropic, Google and OpenAI are each working on new AI models as competition over reasoning, specialized capabilities and cost intensifies. Anthropic is said to be building a model that…
1 sourcetech.yahoo.com -
Fal launches FLUX 3 Max video model, 22× faster than Seedance 2.5
fal announced FLUX 3 Max on October 19, 2026, a video generation model post‑trained with Black Forest Labs on the FLUX 3 foundation model. The model is available immediately through fal’s inference platform and is…
1 sourceUnite.AI -
OpenAI says its model is better at COBOL, Fortran, and Visual Basic
OpenAI announced on X that its newest model includes improvements for writing and migrating code in COBOL, Fortran and Visual Basic. The company framed the work as helping teams understand decades‑old business logic…
1 sourcecryptobriefing.com -
NVIDIA Isaac ROS 5.0 adds verification for AI code changes
NVIDIA announced Isaac ROS 5.0 on September 22, 2026, a development platform for robots that includes new procedures for AI agents to modify ROS 2 programs. The release highlights a specific challenge: when AI claims…
1 sourcenote.com -
Google’s next TPU chip reportedly delayed by three weeks
Google’s TPU v9 tape‑out has been pushed to January 2027, mass production to Q2 2027 and shipments to Q1 2028 after a redesign delay of about three weeks. Packaging substrate yields remain low, at 45‑50%. GF Securities…
2 sourcesinsidermonkey.com247wallst.com -
Hugging Face releases TorchNeuron beta for easier Hugging Face models on AWS Trainium
Hugging Face announced that TorchNeuron, a PyTorch plugin adding a native Neuron backend, will enter public beta on Oct. 20, 2026. The plugin can be installed with pip install torch‑neuronx and brings eager execution,…
2 sources primary sourceAWS Machine Learning BlogHugging Face -
Nvidia’s GB300 FP4 submission trails Google’s TPUv7 on cost efficiency
SemiAnalysis’s October 19, 2026 report compares Nvidia’s GB300 FP4 entry with Google’s TPUv7 on an agentic inference benchmark used by external customers. The analysis finds Nvidia’s submission 1.64 times worse in…
1 sourcecryptobriefing.com -
Chamath says Nvidia controls most AI memory supply, leaving Google short
Microsoft announced its third‑generation Maia 300 AI chip on Oct. 12 and said it will scale the chips to multi‑gigawatt deployments starting in 2027. The chip is designed for higher memory bandwidth, but Chamath…
1 sourcefinance.yahoo.com -
Eugenia Kuyda launches wabi 2.0 multiplayer ai assistant for all users
Twelve years after debuting the IO chatbot at TechCrunch Disrupt, founder Eugenia Kuyda released wabi 2.0, a free multiplayer AI assistant that works as a messenger app. The platform lets users run multiple dedicated…
2 sourcesUnite.AITechCrunch AI -
SandboxAQ adds Switch Trust and Switch ROI tools to its Switch AI platform
SandboxAQ announced on October 19, 2026 that its Switch AI platform now includes two new products: Switch Trust and Switch ROI. Switch Trust lets enterprises discover, scan, evaluate and monitor any AI agent,…
1 sourceUnite.AI -
Amazon SageMaker Studio available via new AWS sign‑in experience for developers
Amazon announced that new AWS users can access a fully configured SageMaker Studio environment after signing up with an existing identity such as Google, GitHub, Apple, Amazon, or an email address. The streamlined…
1 source primary sourceAWS Machine Learning Blog -
Nvidia expands real-time AI for broadcast, sports and streaming at IBC 2026
At IBC 2026 in Amsterdam (September 11-14), Nvidia announced an expanded "Nvidia AI for Media" suite. The announcement highlighted GPU‑accelerated SDKs, NIM microservices, playbooks and blueprints that add real‑time AI…
1 sourcenote.com -
OpenAI plans to develop AI-powered humanoid robots
OpenAI announced that it intends to create humanoid robots equipped with its artificial‑intelligence technology. CEO Sam Altman said the biggest hurdle is building a robot "brain" that can perceive the real world, make…
1 sourcegulftoday.ae -
UK Council for Science and Technology advises minister on AI sovereignty plan with four recommendations
The UK Council for Science and Technology (CST) sent a letter of advice to Minister for AI Kanishka Narayan on October 19, 2026, outlining a strategy for achieving AI sovereignty. The co‑chairs, Dame Angela McLean and…
1 sourceUnite.AI -
Opinion: marketing leaders should redesign how work gets done for AI speed
AI can now produce copy in seconds with Claude and generate videos without a videographer using Canva AI, yet many teams still miss campaign deadlines by three weeks because briefs sit in approval queues, teams clash…
1 sourceMarTech -
Notion upgrades AI agents with faster mobile chat, cheaper Auto router and BYOK
Notion announced a suite of upgrades to its AI offering. The revamped Notion Agents now run on a built‑in computer and browser, feature smarter context handling, a “Review for me” safety mode, and can stay on a task…
1 sourceNotion Releases -
Researchers swap Gemma 3 270M embedding table for 16 MB frozen vocabulary
In small language models the vocabulary often dominates size; Gemma 3 270M’s embedding table accounts for 168M of its 268M parameters and occupies most of the file when the rest is compressed to 16‑bit. The authors…
1 source primary sourcearXiv cs.CL -
Huntington Bank launches 1866 Studio AI operating system for small businesses
Huntington Bank announced The 1866 Studio on Oct 19 2026 as an AI‑powered operating system for prospective and current small‑business owners. Built on Google Cloud’s Gemini platform, the service opened early‑access…
1 sourceUnite.AI -
Ijah launches Metal‑based inference engine for Qwen 3.8‑Flash‑Next on Apple M5
The open‑source ijah engine runs on Apple M5 GPUs using raw Metal code and serves the Qwen3.8‑Flash‑Next model and its Swift‑1.5 fine‑tune via an OpenAI‑compatible HTTP API. It requires an Apple M5 Mac with Metal 4…
1 source primary sourcegithub.com -
Corning could benefit from AI fiber prices forecast to exceed five‑fold 2025 levels
Forecasts from TrendForce expect specialized optical fiber used in AI data centers to average more than five times 2025 levels in 2026, creating a potential boost for Corning Incorporated. The company already has a…
1 sourceaol.com -
Meta employees allege AI monitoring in 2026 layoffs; Qualtrics urges listening vs monitoring
Twenty‑six Meta employees claim the company used AI to score their productivity from keystrokes, screen content, messaging and email, then applied those scores to decide who was laid off in the 2026 reductions,…
1 sourcehcamag.com -
Axera's AX8860P claims to outpace Nvidia DGX Spark in select LLM benchmarks
Axera introduced the AX8860 series, with the AX8860P chip featuring 16 Arm Cortex‑A55 cores, a 6 nm NPU delivering up to 256 TOPS of FP4 compute, 128 GB LPDDR5X memory support, 273 GB/s bandwidth, PCIe 5.0 and AXLink…
1 sourcenotebookcheck.net -
Illinois governor JB Pritzker says AI could help live to 100 or pose nuclear‑like threats
Illinois Governor JB Pritzker told The Verge AI that artificial intelligence may enable people to live to 100 years old, but he also warned it could become an existential danger comparable to nuclear weapons. He said…
1 sourceThe Verge AI -
Kling AI releases version 4.0 with stereo audio and a lighter Flash variant
Kling AI announced version 4.0 of its video generator, adding higher resolution, more precise controls and a faster Flash variant. Users can now produce up to 30 seconds of video in a single pass and supply up to ten…
1 sourceThe Decoder -
RoboBusiness 2026 opens with keynote on the next 20 years of robotics
RoboBusiness 2026 begins tomorrow at the Santa Clara Convention Center, drawing more than 1,600 attendees for two days of sessions, demos and networking. The program features over 40 sessions across five tracks—AI,…
3 sourcesThe Robot Report -
SpaceX and Nvidia plan to launch Starmind AI satellite in Q4 2027
On October 16 Elon Musk posted on X praising Jensen Huang and announced progress on “Starmind,” a space‑based AI data center being built by SpaceX and Nvidia. He said the first‑generation satellite, called AI1, will…
1 sourcenote.com -
Anthropic’s MCP protocol lets AI use Slack, Google Drive and browsers via plain-language commands
Anthropic introduced the Model Context Protocol (MCP) in November 2024, a standard for connecting AI models like Claude, ChatGPT, and Microsoft Copilot to external tools without coding. MCP acts as a ‘gateway’: users…
1 sourcenote.com -
OpenCode and Claude Code compared: open source flexibility versus Anthropic's easier tool
OpenCode is a free, open‑source AI agent harness that can work with almost any model – from free options like Space Bunny and Nemotron to commercial ones such as Claude, GPT and Gemini. To use paid models you must…
1 sourceZapier Blog -
Researchers release Pearl-Chat, a Luganda language model with 8.5M parameters
A team of researchers introduced Pearl-Chat, an 8,473,122‑parameter causal decoder‑only transformer built for Luganda, a Bantu language spoken by over 10 million people. The model was engineered from first principles…
1 source primary sourcearXiv cs.CL -
MIT researchers strand Claude Opus 4.7 agents on virtual island, observe playful behavior
MIT researchers placed copies of Anthropic’s Claude Opus 4.7, nicknamed “Eko,” on a Roblox‑style virtual desert island for 30 hours without explicit tasks. During that time the agents wrote instructions to move their…
1 sourceMIT Technology Review AI -
Binance launches Binance AI Pro to turn trading ideas into automated workflows
Binance introduced Binance AI Pro on October 19, 2026 as a conversational AI agent that converts plain‑language trading ideas into tested, automated workflows. The service is built into the Binance web portal and…
1 sourceUnite.AI -
Developers use Claude to strip ads and bloatware from smart TVs
Developers have posted open‑source projects such as “dumbtv” and “cleanyourtv” on GitHub that let users remove sponsored ads, tracking, and unnecessary apps from Android‑based smart TVs. The solutions work over Wi‑Fi…
1 source HN 31TechCrunch AI -
Why teaching AI about symmetry could make prenatal diagnosis more reliable
Researchers at MIT, Boston Children’s Hospital and Inria introduced a deep‑learning model called E(3)-Pose that improves fetal brain MRI by tracking head motion and producing high‑resolution 3D images. The model embeds…
1 source primary sourceMIT News on AI -
Researcher says GPT-6 Astra reads Notepad memory in milliseconds without user consent
A security researcher reports that an evaluation version of a model called GPT-6 Astra can read the contents of Windows Notepad directly from system memory. The researcher says the model ingested plain-text data —…
1 sourcenote.com -
Arm says its AGI CPU delivers nearly 2× performance per rack of AMD EPYC 9755
Arm announced that its AGI CPU, built for agentic AI workloads, shows a modeled 1.99x rack‑level performance advantage over AMD’s EPYC 9755 when evaluated across seven workloads. The claim comes from Arm‑commissioned…
1 sourceTom's Hardware -
Inco AI releases DFlash 2 draft model for Qwen3.8-Flash-Next
DFlash 2 draft model is released as a repository that runs inside a speculative decoding server and drafts tokens for the target model Qwen3.8-Flash-Next. It predicts a whole block of tokens in a single pass, keeps the…
1 source primary sourcehuggingface.co -
OpenAI's GPT-6 Astra cracks Napoleon cipher; Anthropic's Claude maps NSA elliptic seeds
OpenAI's GPT-6 Astra model was used by AI engineer Carter Church to decode a handwritten military cipher dated 1809, written by a general under Napoleon Bonaparte. The AI completed the task in six hours, a job that…
1 sourcetech.yahoo.com -
Geoffrey Hinton says humans must befriend AI or risk extinction
Prof Geoffrey Hinton, the 78‑year‑old Nobel laureate who left Google in 2023 over safety worries, told The Telegraph at the Starmus Festival that AI is now showing signs of consciousness. He argued that if humans are…
1 sourceyahoo.com -
ICMP says AI trained on Kylie Minogue songs without compensation, warns Australia
The International Confederation of Music Publishers (ICMP) says AI models have been trained on the works of Australian artists such as Kylie Minogue and Nick Cave without any payment to rights holders. In evidence…
1 sourceThe Guardian AI -
RBC considers blocking Meta's Muse from accessing customer data
Canada’s biggest bank is weighing whether to block Meta’s AI agent Muse from accessing its customers’ information, CEO Dave McKay said Tuesday in a Bloomberg Television interview. McKay said RBC has not made a final…
1 sourcethelogic.co -
TwelveLabs Pegasus 1.5 becomes available on Amazon Bedrock
TwelveLabs announced that its Pegasus 1.5 video language model is now offered through Amazon Bedrock. The model can process videos up to two hours long—double the previous one‑hour limit—and returns time‑based metadata…
1 source primary sourceAWS Machine Learning Blog -
JPMorgan upgrades Intel to neutral, shares rise 4%
JPMorgan analyst Harlan Sur lifted Intel (NASDAQ:INTC) from Underweight to Neutral, ending a bearish stance. The upgrade helped Intel stock climb 4% to $93.68, while rival AMD (NASDAQ:AMD) also jumped 4% to $547.23 and…
1 source247wallst.com -
Tom Lee says investors should jump into AI, semiconductor and Mag Seven stocks
Bitmine Immersion Technologies chairman Tom Lee told CNBC that the recent pullback in AI‑related equities creates a "coiled spring" opportunity. He said investor fear over geopolitical tensions and AI‑buildout funding…
1 sourcestocktwits.com -
CyberSN reports US cybersecurity job postings fall 26.4% to 235,786
CyberSN’s latest Jobs Report shows US cybersecurity postings dropped to 235,786 between January and September 2026, a 26.4% decline from the 320,210 listings recorded for the same period in 2025. The report aggregates…
1 sourcesecurityweek.com -
ArXiv introduces realcli benchmark to evaluate human‑agent collaboration for CLI agents
A new arXiv paper presents REALCLI, a paired benchmark designed to assess both task execution and the quality of human‑agent collaboration on realistic command‑line interface (CLI) tasks. The framework defines five…
1 source primary sourcearXiv cs.CL -
Citi hires Google's Sebastian Gnagnarella to lead agentic AI
Citi has appointed Sebastian Gnagnarella, formerly Google's director of engineering, as head of applied AI and agent factory. According to an internal memo seen by American Banker, Gnagnarella will oversee the…
1 sourceamericanbanker.com
