NAll news

All news

Thursday124 items

  1. 36Kr · AIProduct release
    AI score54

    Unitree Releases H2-X Humanoid Robot Demo, Focused on Protection and Rescue Scenarios

    Unitree Technology released a demo video of the H2-X "Kung Fu Legend" humanoid robot on October 21, showcasing moves such as leaping over a car, kicking, and jumping down from a shipping container, and designing scenarios including blocking a knife to protect a person, lifting a car to rescue someone, and pushing a person out of danger. Unitree claims the H2-X has a maximum battery life of about 8 hours, while the standard H2 has a publicly stated battery life of about 3 hours, but the company did not specify the testing conditions, nor did it disclose the full hardware configuration or how the movements are achieved. The H2-X is a derivative model built on the H2, and the degree of autonomy of the movements in the video remains unclear.

    Same story · 2 reports

    #Industry news#Embodied AI#Product updates

  2. 9to5GoogleProduct release
    AI score55

    INMO GO3 smart glasses launched: featuring MicroLED display and Gemini integration

    The INMO GO3 smart glasses are on sale for $509, with a MicroLED display for each eye, brightness up to 1500 nits, a 30-degree field of view, and Google Gemini integration. It supports real-time translation in 98 languages, 13 of which can run fully offline, and can also take notes in meetings, with summaries later presented directly through the glasses. The device weighs 58 grams, has a replaceable battery, a physical cover for the camera, and optional accessories including a speaker, a control ring, and a sunglasses overlay.

    #Product updates#Multimodal#English text is being prepared.

  3. ThurrottProduct release
    AI score75

    Anthropic Releases Personal AI Agent Claude Duet

    Anthropic has released Claude Duet, a proactive personal agent, launching as a research preview on Android, iPhone, and the web, with Claude Pro and Max subscribers needing to wait in a queue. The product is based on Opus 5.5, can connect to more than 6,000 apps, pushes a cross-app highlights briefing every morning, and supports tasks such as scheduling meetings, building dashboards, and creating financial or sales forecasts. Anthropic says it is positioned for work scenarios rather than everyday chores, and it will compete with personal agents such as Gemini Spark, Meta Muse, Microsoft Autopilot, and OpenAI Dots.

    Same story · 44 reports

    #Product updates#Agents#Anthropic

  4. SiliconANGLEProduct release
    AI score63

    Nvidia unveils AI reference architectures and developer tools including DSX POD at GTC Berlin

    At GTC Berlin, Nvidia unveiled the DSX POD data center reference architecture, which integrates Rubin GPUs, Vera CPUs, networking, cooling, and power distribution into a container; the company estimates it can shorten hardware deployment time from six months to 14 days, with the first systems expected to ship early next year. It also introduced a reference architecture for the OpenShell agent toolkit, supporting operation on edge DGX Station, DGX Spark, and laptops equipped with RTX Spark. On the robotics side, it updated the Isaac simulator API and presented three template workflows for customizing GR00T; the Agile ONE SN humanoid robot based on GR00T was also unveiled the same day.

    #Product updates#English text is being prepared.#Embodied AI

  5. BleepingComputerProduct release
    AI score58

    Google Chrome adds Gemini-powered scam detection, can block ClickFix attacks

    Google Chrome has added Gemini-powered scam detection to Safe Browsing Enhanced Protection, which can identify deceptive websites and warn users when they copy suspicious commands in a ClickFix attack. Chrome has changed the Gemini model that originally ran on-device to a server-side model, analyzing page text only when a website shows high-confidence signs of a scam; page content is not logged, retained, or used to train models. Chrome has also added a "suspicious site" warning for pages that exhibit deceptive patterns but do not reach the threshold for a full-page red warning.

    Same story · 2 reports

    #Product updates#Safety / alignment#Google

  6. MarkTechPostProduct release
    AI score69

    Vercel open-sources durable-harness, letting coding agents keep running after a crash

    Vercel has open-sourced durable-harness, an npm package that can wrap coding agent frameworks such as Pi, OpenCode, fx, Claude Code, and Codex, writing sessions to a persistent log so that after a function crashes, a queue hands that turn to a new function to continue execution. Completed model and tool calls are read directly from the log and are not billed twice; tools can be marked as safe to replay or never to replay. The package is at version 0.0.x under the Apache-2.0 license, and can run on Vercel Functions plus Queues, on Node.js 22+ with Postgres, or on local files; Claude Code and Codex must run in Vercel Sandbox, and no performance benchmarks have been published.

    Same story · 2 reports

    #English text is being prepared.#Agents#English text is being prepared.

  7. English text is being prepared.Product release
    AI score78

    Anthropic Launches Claude Duet AI Agent, Opens Waitlist as Research Preview

    On October 21, Anthropic launched Claude Duet as a research preview, positioning it as an AI agent that can work proactively. Built on Claude Opus 5.5, it is available as iOS and Android apps and on the claude.ai web version, with the first waitlist opening to Pro and Max subscribers in the United States. The app has its own dedicated virtual machine and browser and can continue carrying out tasks such as scheduling meetings, organizing emails, and making forecasts after the user's device is turned off, and it can call on more than 6,000 connectors and plugins in the Claude Marketplace. Anthropic says it does not directly access users' passwords or login tokens, and that sending messages, making payments, signing documents, or deleting content requires user approval; the company also acknowledges that this is still an early version that may misread conversations or misjudge how urgent a matter is.

    Same story · 44 reports

    #Product updates#Agents#Anthropic

  8. English text is being prepared.Product release
    AI score66

    Walmart Launches Its Own Agent Gateway, Allowing AI to Place Orders and Shop on Users' Behalf

    Walmart announced the launch of its own agent gateway, a new MCP server that lets vetted AI agents connect to the Walmart platform, check online inventory, and place orders on users' behalf, with support for Sam's Club as well. Walmart has also developed technology to distinguish agent orders from users' own orders, and next month it will launch an agent profile feature that lets users authorize each agent individually, view its orders, and set either manual confirmation for every purchase or automatic authorization within a spending limit. Walmart has also introduced terms of use for AI agents, stipulating that agents must not act beyond the user's intent, cannot access payment information, and that violating agents will be banned from accessing the website.

    Same story · 4 reports

    #Product updates#Agents#English text is being prepared.

  9. Modal Official Engineering BlogFirst-partyProduct release
    AI score68

    Modal Releases LLM Training Library Modal Dojo

    Modal has released Modal Dojo, a training library that provides tuned default recipes, modifiable parameters, and an out-of-the-box metrics observability dashboard, supporting models from 500 million to trillions of parameters. The library is built on open-source frameworks such as Miles and Slime, can load datasets from Hugging Face, and lets agents query run status, logs, and trajectories via a CLI. Axiom Bio says it has used it to launch training of DeepSeek V4.1 Flash with a 64k context, running on 64 B300s.

    #Product updates#English text is being prepared.#Open-source ecosystem

  10. cnBetaProduct release
    AI score78

    Anthropic launches proactive AI agent Claude Duet, which can handle tasks in the background and seek approval before key actions

    Anthropic launches proactive AI agent Claude Duet, which every morning compiles a briefing of important matters from connected apps, checks accounts every few hours during the day, and notifies users only when attention is needed. Duet has its own virtual machine and browser, and can handle tasks such as scheduling meetings, organizing emails, building dashboards, forecasting sales data, and building web apps, and calls more than 6,000 connectors and plugins through Claude Marketplace. When it involves sending messages, making payments, signing documents, or deleting content, it first shows the proposed action and waits for approval; securities trading and transfers between accounts are not executed even with approval. The product is currently available as a research preview, requires application through a waitlist, is first aimed at Claude Pro and Max subscribers in the United States, offers standalone iOS and Android apps and a web version, and shares usage quotas with existing plans.

    Same story · 44 reports

    #Product updates#Agents#Anthropic

  11. The Verge AIProduct release
    AI score76

    Anthropic launches AI agent Claude Duet, based on Claude Opus 5.5

    Anthropic releases AI agent Claude Duet, benchmarked against Meta's Muse and OpenAI's Dots, launching as a research preview via standalone iOS and Android apps and the web version of Claude. The agent is based on Claude Opus 5.5, can call on Claude's memory, skills and connected apps, and comes with a virtual machine with a browser to complete tasks; before key operations such as payments and deletions, it first sends a message requesting user approval. Anthropic says it comes with new dedicated classifiers and safety system prompts to refuse to share information at the strictest tier, such as credit card numbers, while stating that this is still an early version that may misread conversations or misjudge urgency, and is not yet strong enough for personal matters such as booking tickets.

    Same story · 44 reports

    #Product updates#Agents#Anthropic

  12. MacRumors Apple IntelligenceProduct release
    AI score73

    Anthropic launches proactive AI agent Claude Duet

    Anthropic has launched Claude Duet, a proactive AI agent that can send reminders and complete tasks before the user asks. Duet is based on Opus 5.5 and sends a briefing of important matters every morning, after which it checks connected apps every few hours and issues reminders when needed; it can run in a cloud virtual machine with its own browser, and obtains user approval before sending messages, making payments, signing documents, or deleting files. Duet reuses the memory, skills, and connected apps of the Claude account, with interactions concentrated in a single continuous thread; it does not see account passwords, and is equipped with dedicated classifiers and safety system prompts that check messages and prevent the leakage of sensitive information.

    Same story · 44 reports

    #Product updates#Agents#Anthropic

  13. Unite.AIProduct release
    AI score80

    Anthropic opens Claude Duet research preview to Pro and Max users in the US

    On October 21, 2026, Anthropic opened the Claude Duet research preview. Built on Claude Opus 5.5, the app turns Claude into a proactive agent that can proactively surface to-dos, continue executing tasks after the user leaves, and confirm before acting on the user's behalf. Access is available through a waitlist, starting with Pro and Max users in the US, and is offered via standalone iOS and Android apps and the claude.ai entry point. During the preview period it is included in the Pro and Max plans and is subject to the same usage limits. The app can call on more than 6,000 connectors and plugins in the Claude Marketplace; user approval is required before sending messages, making payments, signing documents, or deleting content, while securities trading and transfers between accounts are not permitted even with approval.

    Same story · 44 reports

    #Product updates#Agents#Anthropic

  14. Unite.AIProduct release
    AI score72

    Walmart Launches Agent Profiles and Agent Gateway to Set Permissions for AI Shopping Agents

    Walmart announced the launch of Agent Profiles, the Walmart Agent Gateway, and terms for AI agents, enabling customers to shop at Walmart and Sam's Club through apps and agents of their choosing, with built-in permissions, controls, and protections. Agent Profiles will launch next month; customers can authorize a separate profile for each agent, choose per-transaction approval or set spending limits such as $50, $100, or $200, and can revoke permissions or delete profiles at any time. The Agent Gateway is built on MCP and is currently available as an invite-only beta for enterprises and developers; it claims to provide real-time product and fulfillment data, reduce requests and token consumption, and includes safeguards such as category restrictions and rate limits.

    Same story · 4 reports

    #Product updates#Agents#English text is being prepared.

  15. 9to5GoogleProduct release
    AI score69

    Anthropic launches proactive AI agent Claude Duet, opening a waitlist for Pro and Max users

    Anthropic releases proactive AI agent Claude Duet, which combines the user's historical context with Claude and connections to external apps, pushes a morning briefing of important matters every day, and sends message reminders when attention is needed. The product can handle tasks ranging from scheduling meetings to building dashboards, creating financial or sales forecasts, prototyping, and building web apps. It is currently available as a research preview for Pro and Max plans, and users must first join the waitlist. This comes weeks after OpenAI launched its proactive agent Dots and Meta's Muse gained momentum.

    Same story · 44 reports

    #Product updates#Agents#Anthropic

  16. 9to5MacProduct release
    AI score82

    Anthropic Launches Proactive Agent App Claude Duet

    Anthropic has opened a research preview of Claude Duet, a standalone app that works as a proactive agent, handling tasks across connected apps, tracking email and calendar, and continuing to run in the cloud after the app is closed. The feature costs nothing extra but counts toward existing plan usage, and background work may cause users to hit limits sooner; it is currently invite-only and not yet available on Team and Enterprise plans. Claude Duet is powered by Claude Opus 5.5 and is available on iOS, Android, and the web, supporting Stripe unpaid invoice checks, Link and Shop Pay payments after user approval, as well as plugins and a cloud browser.

    Same story · 44 reports

    #Product updates#Agents#Anthropic

  17. PyTorch BlogFirst-partyProduct release
    AI score59

    TorchTitan Update: Support for Multiple New SOTA Models, TitanRL, and Large-Scale Performance Improvements

    PyTorch's native training platform TorchTitan has released an update that adds full training-stack support for Muse Spark Open, Muse Glimmer, Qwen3-VL, Qwen3.5, Kimi K2.7, Kimi K3, and DeepSeek-V4, and introduces the reinforcement learning component TitanRL. TitanRL lets pretraining, post-training, and vLLM inference share the same model code, and supports asynchronous rollout as well as agent-task training such as long-context and code execution. On performance, FlexShard on 256 H100 GPUs reduces DeepSeekV3 optimizer time by up to 79%, and reduces end-to-end training step time on Llama3 405B by up to 20%.

    #Product updates#English text is being prepared.#Open-source ecosystem

  18. TechCrunch AIProduct release
    AI score64

    Walmart launches Agent Gateway, opening shopping to AI agents

    Walmart has released Agent Gateway, an MCP-based server that lets approved AI agents search Walmart's online inventory and place orders on users' behalf, covering Walmart and Sam's Club. Next month it will also launch Agent Profiles, letting users create a separate profile for each agent, view its spending records individually, and choose per-transaction approval or set spending limits. Walmart has also drafted dedicated terms for AI agents, defining their operating boundaries, with non-compliant agents banned from accessing the site; this contrasts with Amazon's move to start blocking AI agents from making purchases on users' behalf.

    Same story · 4 reports

    #Product updates#Agents#English text is being prepared.

  19. The Verge AIProduct release
    AI score65

    Walmart to Open Its Site to AI Agents Shopping on Users' Behalf Next Month

    Walmart announced that starting next month it will allow users to connect their own AI agents to its website to shop on their behalf, and will offer Agent Profiles features including purchase approval, budget caps, and separate views of agent orders. Walmart connects agents through an Agent Gateway based on the Model Context Protocol, saying this avoids obstacles in browser-based operations and does not expose users' full shopping history or personal information to agents. Walmart also stipulates that agents must not act beyond the user's intent and must not change payment methods, and violators will be banned—a contrast to Amazon's blocking of the Perplexity shopping agent and Meta Muse.

    Same story · 4 reports

    #Product updates#Agents#English text is being prepared.

  20. Claude BlogFirst-partyProduct release
    AI score77

    Anthropic Opens Claude Duet Research Preview: An Agentic App That Can Proactively Execute Tasks

    Anthropic has opened the Claude Duet research preview, an agentic app built on Claude Opus 5.5 that can proactively advance tasks. It is now available on iOS, Android, and claude.ai for Pro and Max users in the United States, who must join a waitlist. Every morning it pushes a cross-app highlights briefing, and it can continuously execute tasks in the background; it requires user approval before sending messages, making payments, signing documents, or deleting, and securities trading and transfers between accounts are not permitted even with approval. Claude Duet can invoke more than 6,000 connectors and plugins in the Claude Marketplace, and it follows the credential-handling approach of Claude Managed Agents, never directly touching users' passwords or login tokens.

    Same story · 44 reports

    #Product updates#Agents#Anthropic