Every story we have kept, newest first.
Looking for the daily editions? → Past editions
Page 32 of 1372
NVIDIA has officially announced the specifications and release plans for its first highly integrated SoC for consumer RTX Spark PCs. The N1X combines an Arm-based Grace CPU with the Blackwell RTX GPU architecture and is scheduled to launch through OEM partners in October.
cnBeta (Full RSS) · 12d ago
The explosive growth of AI-generated and AI-processed data is driving renewed demand for traditional magnetic tape storage. Tape shipments reportedly rose 57%, suggesting that the medium remains relevant for large-scale data preservation despite earlier efforts to phase it out.
cnBeta (Full RSS) · 12d ago
Global NAND flash market revenue grew 70% quarter over quarter in Q2 2026, following a 90% increase in Q1. Counterpoint Research attributes the surge to explosive demand for high-capacity flash storage driven by AI infrastructure construction.
cnBeta (Full RSS) · 12d ago
Meta is offering an average discount of about 95% on its Muse Spark model in exchange for explicit permission to inspect usage data. Muse Spark is designed to operate coding and other AI agents.
TechCrunch AI · 12d ago
ChatGPT, Claude, Grok, and Gemini experienced service interruptions at practically the same time. The unusual overlap highlights the reliability risks of depending on a single major AI provider or model ecosystem.
Ars Technica AI · 12d ago
Docker’s YOLO mode allows an AI agent to execute actions without requesting user permission. The article explains the autonomy benefits, security risks, and safer ways to operate this mode.
Docker Blog · 12d ago
Hundreds of protesters gathered in Edinburgh to demand an immediate moratorium on Scotland’s expanding datacentre pipeline. At least 20 projects have been proposed, including an Auchtertool, Fife facility described as the world’s second largest.
The Guardian Technology · 12d ago
The China-linked group OVERCAST PANDA reportedly entered hotel rooms and used USB media to write the FlowCloud backdoor directly to executives' laptops while the devices were unattended. The malware activated after the next boot, enabling keylogging, screen capture, file collection and credential theft without phishing, network intrusion or a stolen login.
VentureBeat · 12d ago
Wayve and Uber are launching robotaxi rides in London, initially with safety drivers behind the wheel. Wayve plans to progress toward fully driverless service and expand with Uber to more global cities using end-to-end AI that does not rely on high-definition maps.
Bloomberg Technology · 12d ago
LangChain has expanded its MCP support in the new langchain.mcp package, built on FastMCP and aligned with the 2026-07-28 specification. Elicitation is implemented through LangGraph interrupts, while tool lists are cached for more efficient operation.
LangChain Blog · 12d ago
Snowflake raised its annual sales outlook, citing rapid adoption of its AI-assisted coding tool. The update boosted investor confidence and sent Snowflake shares sharply higher.
Bloomberg Technology · 12d ago
OpenAI estimated that its partnership with Cursor could generate more than $1 billion in annual revenue. Despite that opportunity, OpenAI ended the relationship after Elon Musk’s SpaceX acquired the AI coding startup.
Wired AI · 12d ago
Anthropic is automatically signing Claude users out after infostealer malware reportedly harvested active login sessions from their computers. The measure is intended to reduce the risk of attackers reusing stolen sessions to access Claude accounts.
Engadget · 12d ago
Lindy can now schedule meetings directly from any email thread where it is copied. It proposes three available time slots or confirms a time that both parties have already agreed on.
TestingCatalog · 12d ago
Chinese automakers including Xiaopeng, BYD, Chery, Xiaomi, and others are rapidly entering humanoid robotics as vehicle margins collapse and the sector seeks new growth. Their advantages include automotive AI, supply chains, manufacturing expertise, and factory deployment scenarios, but unstructured environments and unclear commercialization remain major obstacles.
虎嗅 · 12d ago
AWS demonstrates two reference implementations that apply the AI-Driven Development Lifecycle construction phase. The examples use Amazon Bedrock AgentCore, Kiro, and Claude Code to generate ER diagrams from SQL and analyze code security with multiple agents.
AWS Machine Learning Blog · 12d ago
AWS shows how to connect Microsoft Outlook with Amazon Quick for email management, calendar scheduling, and workflow coordination. The walkthrough covers automation through Quick chat agents, Quick Flows, and Quick Automate.
AWS Machine Learning Blog · 12d ago
This tutorial shows how to deploy a customer-operated LiteLLM gateway on Amazon ECS with AWS Fargate and connect ChatGPT Codex to an OpenAI model hosted on Amazon Bedrock. It also covers scoped identities, budgets, rate limits, telemetry, and comparisons with IAM Identity Center and Portkey.
AWS Machine Learning Blog · 12d ago
Ollie is positioning its family-focused AI assistant around privacy as a competitive advantage. Although it wants access to users’ everyday-life details, the company says it will not use that data to train AI models or share it with others.
TechCrunch AI · 12d ago
HPE CEO Antonio Neri discussed the benefits of the company’s networking deal with Oracle and its earnings outlook. He also said HPE is seeing huge growth in enterprise AI adoption.
Bloomberg Technology · 12d ago
Character.AI shared an update on its ongoing safety work. The company reaffirmed its decision to remove open-ended Character chats for users under 18, calling it one of its most significant safety measures.
Character.AI · 12d ago
Meta introduces ZGateway, a proxy layer that unifies traffic to ZippyDB, its widely used key-value store. The architecture also provides admission control, load balancing, cross-region resilience, and richer operational capabilities.
Meta Engineering Blog · 12d ago
Nvidia launched PAIR, free open-source software that links compatible computers on a local network for AI inference and agentic workflows. It supports Nvidia RTX 20-series and newer GPUs, RTX Pro systems, DGX Spark, and Apple M4-or-newer devices.
The Verge · 12d ago
Systems powered by Nvidia’s RTX Spark N1X will launch in October across mini PCs and laptops. The chip family will offer two configurations, with up to 20 CPU cores, 6,144 CUDA cores, and 128GB of unified memory.
Tom's Hardware · 12d ago
NVIDIA’s PAIR Virtual Inference Router expands the compute available to AI workloads by routing inference across devices on a local network. The update targets increasingly parallel workloads, including multiple agent sessions and multi-agent workflows with specialized subagents.
NVIDIA Developer Blog · 12d ago
NVIDIA has introduced PAIR, a free and open-source tool for distributing AI workloads across idle computers. The project could help teams make better use of otherwise-unused computing capacity.
Engadget · 12d ago
Nvidia’s Personal AI Router, or PAIR, lets agentic AI workloads use spare GPU capacity across GPUs connected to a home network. The utility is designed to distribute agent swarms more efficiently while potentially improving speed and inference privacy.
Tom's Hardware · 12d ago
Lenovo has shown the ThinkCentre X Ultra mini workstation ahead of IFA, featuring AMD’s Ryzen AI Max+ Pro 495. The system provides an early look at AMD’s Gorgon Halo platform in a compact workstation form factor.
Tom's Hardware · 12d ago
Google is rolling out AI-powered voice capabilities for Gmail, Docs, and Keep. The tools are available to paying customers, according to the article.
Engadget · 12d ago
Google is rolling out AI-powered conversational voice modes for Gmail, Docs, and Keep. Gmail Live, Docs Live, and Keep Live let users find information, manage content, and capture notes through real-time voice interactions.
The Verge · 12d ago