MarkTechPostInfini-AI-Lab releases open-source framework Hermit v0.1.0, with adapters that can fine-tune Codex or Claude Code agents without altering the model or framework. Using GLM 5.3 Flash on OfficeQA, CL-Bench, and ServeLearnBench, results with adapters are better than baseline results without adapters. The MWM adapter achieves 0.795 accuracy in database exploration on CL-Bench.
Transformers releases33GitHub Copilot released version 5.21.1 patch, but the original text did not provide specific update content.
The Claude Smart Reporting Tool, launched by Anthropic, is aimed at Claude Enterprise paid users, capable of generating charts to display team usage of AI tools (conversations, Claude Code, Claude Cowork), output content, and costs. It supports custom query questions and identifies inefficient usage scenarios, but it is explicitly not used for evaluating individual performance or personnel decisions.
Lenovo Tab P Pro GT is now priced at 1423 yuan on Tmall Official Store after adding Tmall Gold Coins, with a launch price of 1999 yuan. The tablet is equipped with a Snapdragon 8 Gen3 processor, a 11.1-inch 3.2K resolution screen, supports the Pro Pen, and has a battery life of about 11 hours.
NewCore Voyage has released the X7 Ultra large model chip, which includes a high-precision operator acceleration engine, supports 10 times the parameter model deployment, and achieves an NPU computing utilization of 98.7%. Memory bandwidth consumption is reduced by 80%, and energy efficiency is improved by 100%. This chip is the first in China to mass produce Armv9 automotive-grade CPU, with a performance increase of 20% and an energy efficiency improvement of 10%.
Xiaomi has increased the membership credit for MiMo Desktop without increasing prices, taking effect in real-time for both new and existing members. New third-party models such as DeepSeek V4 Pro, DeepSeek V4.1 Flash, GLM 5.3, GLM 5.3 Flash, Kimi K3, Qwen 3.8 Flash, and Qwen 3.8 Max are now available for members to choose, with each model labeled with its consumption multiplier.
Hunan Titan Future announced the launch of the country's first 12-inch polycrystalline silicon carbide substrate, which has passed TSMC's certification tests and sent samples to Samsung. The product has a deposition speed of 60 micrometers per hour, a minimum resistivity of less than 1 milliohm·centimeter, and fewer than 0.1 holes per square centimeter. It is collaborating with companies like Google and NVIDIA, and is expected to be mass supplied after 2027.
Altera has launched the Agilex 2X series FPGA, supporting edge physical AI and digital AI infrastructure. Agilex 27 is targeted at processing and bandwidth-intensive applications, featuring a third-generation HyperFlex interconnect structure, which improves energy efficiency by 1.8 times. Engineering samples are expected to be released in the second half of 2027.
Amazon releases five new products centered around the Alexa AI assistant, including an upgraded Echo Dot smart speaker, the first Echo Studio smart speaker equipped with a camera, Echo Time smart clock, Alexa Photo Frame digital photo frame, and Echo Show 16 smart display. The flagship smart display Echo Show 16 starts at $499.99.
Since its launch in March of this year, WorkBuddy has iterated 100 major versions, completed 900 million user task requests, and has been adopted by enterprises in over 50 industries, covering 50% of the Fortune 500 and 60% of the Fortune 100. The open platform has more than 50,000 partners. WorkBuddy will fully open its product architecture, supporting self-developed Agents by enterprises and third-party Agents, including integration with third-party IM systems. Enterprises can choose models, Agents, and IMs independently.
The Ascend 950DT 1024 card-based intelligent computing cluster has been launched in Wuhan, with a single node equipped with 1024 Ascend NPU acceleration cards, achieving FP8 computing power of 1 EFLOPS, and featuring 256TB unified global memory, fully supporting the training of large models with trillion parameters and intelligent agent inference.
Tencent releases cross-application AI collaboration system Buddy One, integrating document, meeting, cloud storage products and sharing WorkBuddy Agent base, unifying accounts and point pool; users can link accounts without repeated authorization, WorkBuddy can integrate content from multiple applications to generate source-linked answers, points can be used across multiple platforms.
Samsung's newly approved sleep apnea detection algorithm will be pushed to Korean version Galaxy Watch 9/Ultra2 and Galaxy Ring next month, the new algorithm can detect signs of moderate to severe sleep apnea and count the number of breathing interruptions per hour, divided into three levels: mild, occasional, and frequent, the results can be used by users and medical personnel for reference but cannot replace professional diagnosis.
AMD has released the first Vulkan-supported FSR 'Redstone' version FSR SDK 2.4, bringing supersampling and frame generation features to the graphics API that are the same as DirectX 12.
Balmuda launches AI personalized radio service BALMUDA Good Morning Radio, which can generate about 10 minutes of daily podcasts based on users' habits and interests. The service will start Beta testing for The Clock clock users from November 2, 2026. Users who have not purchased the clock can also apply for testing through the BALMUDA Connect app (limited to 1000 names).
Tencent WorkBuddy releases Buddy One integrated suite and upgrades cloud Agent Runtime and security capabilities; Enterprise version launched in March covering top 50 companies, open platform integrates over 200,000 Skills; WorkBuddy will open architecture to support enterprise self-built Agents and third-party Agent integration.
Nubia announces the Nebula AIOS 3.0 upgrade plan, will push the Z80 Ultra phone in December, and the subsequent Z70 Ultra, Z70S Ultra Photographer version will be upgraded in March 2027. The new system has not announced the specific upgrade content.
OpenRouter BlogOpenRouter has released a new supplier routing algorithm that selects the best cost-performance solution for each request by evaluating price, latency, throughput, and error rate in real time, significantly improving reasoning speed (TTFT improved by 16.9%), throughput (improved by 16.3%), and reliability (error rate reduced by 67%), with only a slight increase in cost of 6.0%. This algorithm rewards suppliers that provide high-quality reasoning without requiring users to modify their code, ensuring a continuously optimized experience.
Sakana AISakana AI has launched the enterprise-grade API platform Sakana API Platform and initiated an early access program, allowing users to use Sakana AI LLM via API and manage members and costs within the organization. The platform is compatible with OpenAI API, initially offering the Sakana Fugu and Sakana Fugu Ultra models, with subsequent releases of Sakana Namazu and Sakana Fugu Max.
Google will proactively promote Android foldable screen app adaptation. Next year, the Google Play Store will evaluate app performance on foldable devices, providing more exposure opportunities for products with excellent adaptation. Google has updated its development tool Compose adaptive skill, allowing AI programming agents to automatically transform mobile apps to adapt to different sizes and forms of devices, including foldable devices and Googlebooks.
Databricks Blog3838Procore shares connected construction data programmatically with clients' existing environments through OpenSharing, supporting Power BI, Tableau, Snowflake, BigQuery, Databricks workspaces or local SQL Server for analysis.
OpenAI and ServiceNow are introducing more OpenAI models and agents into ServiceNow products, integrating ServiceNow workflows with ChatGPT and Codex. This move aims to make it easier for businesses to use AI to build applications, handle IT service requests, and address security issues. Expanding on the collaboration announced in January, OpenAI's capabilities will be embedded into more aspects of ServiceNow's customer service and internal business operations.
How the accounts differ:OpenAI News states the integration is into 'ServiceNow enterprise workflows' while 36Kr mentions 'ServiceNow product workflows'.
Tencent WeCom launches beta testing application, supporting connection to corporate WeChat, Tencent Meeting, and other mainstream office tools and AI applications, sharing storage and aggregating context. The official website has opened an entry for users to submit applications, including four core capabilities: storage, connection, intelligence, and security, which are already included in WorkBuddy members and do not need to be purchased separately.
The Nemotron 3.5 Transcribe open-weight speech model now supports RTX Spark, enabling speech transcription, speaker recognition, and translation, supporting 40 languages. The Liquid AI LFM3 series models have also been supported and can be run locally and fine-tuned. The RTX Spark laptop starts at $2599, with the highest configuration of 128GB unified memory, and FP4 AI computing power up to 1 petaflop.
Google is taking its advertising business in AI search experience to international markets, betting on generative AI to open up new revenue opportunities. Under the AI model, the click-through rate of sponsored product lists is comparable to that of natural search lists, with user research showing that the ads are highly relevant. The total volume of search queries has experienced the fastest year-on-year growth in nearly five years, with AI mode queries averaging three times those of traditional search.
LG has responded that the Microsoft Copilot icon on the TV is not a native application but a web browser shortcut, designed to enhance customer convenience. Microsoft will stop operating the Copilot service, but did not mention a specific end date, with external speculation suggesting it may be phased out around December 2026.
Codex releases00GitHub Copilot released version 0.166.0-alpha.6, no specific update content mentioned.
TechCrunch · AISame story · 2 outletsGoogle has expanded AI-powered search ads to international markets, trying to open up new revenue streams as core search ad growth slows. Q3 revenue grew by 25% year-on-year to $127.6 billion, but search ad growth continued to decelerate, and Alphabet's stock price fell 8%. Philipp Schindler said in the earnings call that high-quality ads have been extended to more users and queries in global AI overviews and AI patterns, with a click-through rate comparable to natural search results.
Google Developers BlogLiteRT has received a major upgrade in Google Play services, with GPU inference speed about 1.4 times faster than traditional GPU accelerators, achieved through the newly introduced CompiledModel API, driven by the open-source GPU engine ML Drift
OpenAI NewsSame story · 2 outletsOpenAI and ServiceNow integrate OpenAI models and agents into ServiceNow enterprise workflows, providing support through ChatGPT and Codex.
How the accounts differ:OpenAI News states the integration is into 'ServiceNow enterprise workflows' while 36Kr mentions 'ServiceNow product workflows'.
Anthropic NewsroomAnthropic expands the AI for Science project, providing free Claude credits to 20 global research institutions, foundations, and research centers to support high-impact science research.
Sierra Blog6565Sierra introduces Compliance Console, enabling businesses to set compliance rules for long-term AI agents through the Horizon platform, covering dimensions such as channels, use cases, and regions. Users can adjust control items like disclosure requirements and frequency limits to comply with local regulations. This tool supports businesses in expanding their agents to new scenarios and countries while adhering to the rules of the customer's location.
OpenAI News3535Evernote promotes its note-taking app through ChatGPT Ads, tracking the frequency of free trials converting into paid subscriptions
TechCrunch · AIAnthropic releases new report tool to help enterprise customers better understand the token spending value of tools like Claude Code, including team work analysis, outcomes, costs, and workflow issues, and provides recommendations; its net revenue retention rate reached 900% from Q3 2025 to Q3 2026.
The Verge · AI6565Microsoft has stopped supporting the Copilot feature on all smart TVs, with this notice appearing on LG TVs on October 28, 2026. Copilot primarily serves as a friendly AI companion that users can chat with, ask about the current screen content, or get recommendations for TV shows and movies. Microsoft launched a new Copilot 'super app' earlier this year, bundling various AI features like chat, agents, and coding together, positioning it as a tool more focused on business and productivity.
Databricks BlogDatabricks releases voice intelligence agents, which adjust responses in conversations based on real-time contextual decisions, enabling language switching, transferring to a human, or providing personalized services. These agents need to process customer requests within 200 milliseconds and rely on the context, ontology, and governance capabilities provided by Databricks. They integrate customer, order, product, and policy data, and follow the access rules of the enterprise data catalog.
Apple released its first touchscreen MacBook Pro equipped with a PROLED XDR display, featuring the M5 Max chip version with up to 128GB unified memory, capable of locally running large language models with billions of parameters, supporting applications like LM Studio, starting at 22,999 yuan.
