Instinct launched its invite-only AI agent in August. Users can text it to handle chores such as booking DMV appointments and sending follow-up emails. Muse and Dots arrived afterward; the report describes them as broadly similar products from more powerful companies.
All AI updates
Oct 9
A blogger used voice conversations in the Codex tab of the ChatGPT desktop app while cooking dinner to build a newsletter archive for his blog. Voice work covered most of the imports, archives and search integration; before deployment, he spent another half hour using typed prompts to revise an import and review the code.
Doubao Work’s task mode now has an unlimited canvas for viewing and comparing materials, plans and results in one place, with options to adjust text, colors and layout after generation. Doubao 2.1 Lite is available for tasks including document writing, spreadsheet work and PPT creation; image creation also supports Seedream 5.0 Flash.
In a new Q&A, Huawei said Qiankun ADS does not immediately disengage during encounters on narrow roads and can reverse to yield when space is insufficient. It also said CAS 5.0 can detect and respond to risks while a person is driving without ADS enabled. The system remains driver assistance and cannot replace the driver.
Anthropic has announced two Claude features in testing. Users can connect a company data platform and ask in natural language for an automatically updating dashboard, or turn text, charts and other materials into a short animation. Dashboards are available to all paid-plan users, while Motion is limited to Team and Enterprise users.
Google Cloud announced Gemini agent as a single interface for questions, knowledge work, media creation and coding. The report says it currently routes tasks across Gemini and Anthropic Claude models. Teams can set a project spending cap; when reached, the agent pauses until someone resumes it.
展开进展与来源
Alexandr Wang says each Muse runs in its own secure virtual machine, where data shared with it is stored. A separate agent checks attempts to send information out and blocks sensitive data such as credit card numbers; users must also approve each new website Muse wants to contact. A more encrypted virtual machine is planned for later and may be slower.
Alexandr Wang says each Muse runs in its own secure virtual machine, where the data shared with it is stored. A separate agent checks outgoing information and blocks sensitive data such as credit card numbers; Muse also asks users to approve each new website it wants to contact or send data to.
NeoWin tested the new search experience in Windows 11 Build 26340.9616. Natural-language requests can bring up actions beside search results for settings such as dark mode and Bluetooth, with a Revert button after a change. In the outlet's preview test, search also worked offline and responded almost immediately after input.
The Xiaomi Smart Door Lock 5C Peephole Edition is available for presale, with a regular JD.com price of 1,176 yuan. It locks automatically when the door closes, can alert users to someone lingering at the door or to peephole obstruction, and supports remote unlocking by phone.
Users can speak or type a request to pay water, electricity or gas bills in the Doubao app to open the payment page. Account linking, confirmation and payment still require separate steps. In IT Home's test, all three services showed “not yet available” in Qingdao.
The API selects from preset answers for low latency decisions in text routing, visual actions, expressive voice interactions, and robot tracking. The source does not specify access or latency figures.
Voyager’s macOS app runs on the user’s own computer and can operate creative software including Blender, DaVinci Resolve and After Effects. It can also orchestrate video generation models. Its presenter says it guides beginners through storyboards and character sheets without node-based editors.
Voyager is presented as a tool for video, graphics and game creation. It includes free graphics, music and video editing tools and is said to be able to drive Blender, Resolve, After Effects, Ableton and more. The material does not specify when or to whom it will be available.
The Interface ring uses a press-and-hold gesture to capture a spoken request and sends it when released; the demonstrator says it can connect to agents including Claude Code, Codex, and Hermes. The announcement lists a $99 early-adopter price and says it ships in January, without specifying the year.
Google announced a unified Gemini agent for businesses, with a consumer rollout planned later. It can plan tasks, use tools and connect to internal systems. Users can track progress in a tasks inbox and choose a model themselves, initially including Anthropic’s Claude models.
展开进展与来源
Voice Arena says Monsoon ASR contains more than 100,000 hours of speech across 50 languages, most of them long-tail languages. It reports that fine-tuning Whisper Medium on the dataset reduced word error rate from 85.27% to 7.65% on Bengali FLEURS; interested labs can use its model API to test against their own internal benchmarks.
Interface is designed to let users press the ring to ask AI agents to handle tasks or capture thoughts, while also tracking health metrics including heart rate and sleep. Natura plans to connect it to apps including Claude and ChatGPT, with pre-orders expected next month and shipping slated for December or January. The company plans a $9 monthly subscription after an initial three to six months free.
Goodfire has launched monitors that track AI agents through signals inside the model. They are available to Baseten customers, who can choose which risks to monitor and whether to log an event, send it for human review or refuse the request. In Goodfire’s tests on Kimi K3, monitoring about 1,500 sessions cost roughly $51, and the probes caught 94% of malicious hacking sessions; these are company test results.
Oct 8
Google has released Google AI Edge Foresight, a free experimental app for macOS that transcribes meetings and audio files entirely offline. It also summarizes meetings, and Google says it can turn users’ shorthand bullet points into notes using the transcript.