Huawei's OceanStor M900 and Nvidia's CMX move reusable inference KV cache into a shared storage tier, but no agreed SSD specification exists for it. A storage executive said endurance requirements for one design rose from 3 to 9 drive writes per day and could change again. Industry sources expect convergence to take 6 to 12 months, with another year for development and validation.
Simplexity Robotics says it trained a single robot to load and unload two CNC lathes on its own, using 600 real-robot trajectories and reporting a 100% success rate on the precision CNC insertion task. The work, presented at IROS 2026 on September 29, combines the SimpleWAM world action model, a DRAM memory module and DPE action scoring, with force and torque feedback for insertion recovery. The company did not say how many trials the 100% figure covers, and it does not describe the yield of the whole cell.
Huawei has opened its DevEco Studio for HarmonyOS PCs to public beta, alongside first public betas of the AI tool DevEco Code and the agent toolkit DevEco CLI. The beta requires HarmonyOS 7.0.0.107 or later, at least 16 GB of memory and 100 GB of storage, and runs on several MateBook models and the MatePad Edge. DevEco Code ships with Zhipu AI's GLM-5.3 and GLM-5.1 models and supports third-party model connections.
Hong Kong's Centre for Artificial Intelligence and Robotics (CAIR), Chinese Academy of Sciences, unveiled CARES 4.0, a multimodal clinical AI agent that carries out tasks rather than only answering questions about images or video. Built on the Harness agent framework and CAIR's own CT, MRI, ultrasound, endoscopy and EEG foundation models, it has been validated at several top-tier hospitals. CAIR says the system gives suggestions with reasoning paths and sources, and that the doctor remains the final decision-maker.
Grok bot's scheduled AI morning briefing video task ran, and the results look good.
CoreWeave is layering managed services over its infrastructure to address AI inference bottlenecks, including a preview capability called CoreWeave RL Rollouts that improved model reload latency by 15x versus a baseline configuration in testing. The capability is built on Nvidia's Dynamo framework, and the features are packaged into CoreWeave Forge, a platform that is free to start with paid tiers offering additional capabilities.
Sakana AI's technology has been adopted in "Evidence Finder," an evidence search tool for physicians from Iris Inc. (@ShoOkiyama). In response to physicians' questions, it searches the literature and answers while explicitly citing its sources. Sakana AI's technology handles the comparison and integration of literature and the generation of answers. For details, see here: https://sakana.ai/namazu-aillis/ 🐟
Google Cloud introduced the Gemini agent, a general office agent that can search, write emails, build slides, analyze data, run code, and coordinate sub-agents. It can take on an enterprise identity with email, calendar, and account, and it selects underlying models automatically, including Anthropic's Claude. The article presents this alongside OpenAI's Dots and Meta's Muse as competing office and personal agents.
Google Cloud unveiled Gemini Agent at its Gemini at Work 2026 event, a general-purpose work agent for enterprise customers that works across Gemini Enterprise and Workspace. Currently the agent supports Gemini and Claude models, with an API planned so developers can integrate it into third-party applications.
Snyk moved its internal support agent, Snyk Assist, into the core Snyk product in September 2026, giving every paying customer access. Built on LangChain and LangGraph with observability in LangSmith, the agent answers questions in plain language and can open support cases or log feature requests. It runs as a single agent behind Slack, web and API surfaces, with tools attached per user permissions.
Australian AI data center operator Firmus, backed by Nvidia, has withdrawn its planned initial public offering, citing market volatility and conditions. Its board concluded the proposed terms did not adequately reflect the company's business strength and long-term growth outlook. Firmus said it will now pursue private market capital and consider other public and private options.
Tessl's blog post argues that teams should record agent mistakes as attributed, signed diary entries, then curate them into reusable context packs rather than adding unverified rules to files like AGENTS.md. The author describes a REST API case where an agent regenerated the OpenAPI spec and TypeScript client but missed the Go client, and the same lesson had to be re-taught in a fresh session.
Tessl argues that an AI-native organization collapses the handoff between people who own outcomes and the work itself, so product managers and designers can execute changes through agents. It says PR count and token spend are insufficient measures, and that merge rate better shows whether the new workflow is working. The article also says the boundary should follow decision authority, with engineers still owning architecture and data models.
The author argues that system use cases, with actors, preconditions, scenarios, and acceptance criteria, work better than user stories as the input for AI code generation in enterprise business applications. He describes a process that skips the plan-and-task phase, reverse-engineers legacy systems into use cases and entity models for modernization, and recommends self-contained system verticals and risk-based review.
AI DevCon New York, running November 2–4 at Industry City in Brooklyn, centers its program on software factories, the systems needed to make agentic development repeatable, trustworthy and scalable. The article argues that moving from one developer using an agent to an engineering organization requires layers covering context and skills, harnesses and tools, orchestration, verification and evaluation, and feedback.
The article explains how a software factory combines agents, tools, and engineering practices to take work from intent to verified output. It argues that teams should enforce domain rules, workflow gates, and tool permissions independently, version shared guidance and checks, and review lessons before reuse.
A DevDay 2026 session from Higgsfield explores how its agents choose models and provision their own compute. It covers when to use dedicated GPUs versus hosted models, how to deliver predictable throughput, and where the approach can fail.
Stack Overflow's small teams used Codex to build Stack Overflow for Agents and reimagine Stack Internal, according to a DevDay 2026 session. The talk covers lessons for accelerating R&D and bringing new products to market.
OpenAI's DevDay 2026 session shows developers how to cut AI costs while preserving quality. It covers measuring cost per completed task, choosing models and reasoning effort, and using prompt caching, Programmatic Tool Calling, and the Batch API.
Cornerstone OnDemand built Orion AI, a multi-agent system using Amazon Bedrock and the open source Strands Agents framework, that cut database diagnosis from 45 minutes to 10, a 78% reduction. The system also reduced manual lifecycle steps from more than 10 to a single interaction and filtered redundant alerts by a median of 65%. A three-person team delivered it in six months.
Theo, creator of the T3 Stack, open-sourced tsc-rs, a line-by-line Rust port of Microsoft's Go-native TypeScript 7 compiler, type checker, and language server under MIT, pinned to typescript-go commit 673a5f17. The author reports tsc-rs is about 1.61× faster than tsc 7 and about 2.95× faster than bun check on six real-app benchmarks on an Apple M4 Pro. The port passes all 181,711 ported Go tests, and CLI output matches the Go version on 120 open-source repos except for known edge cases such as monorepo rootDir and tsc -b incremental output.
AIWhy it matters: The post reports a benchmarked, test-verified Rust port of the TypeScript 7 compiler, with pinned upstream and stated edge cases useful for judging its compatibility.
Micbot says a procurement dispute with a Zhejiang customer over four wheeled-quadruped robots has been settled through negotiation, after reports of late P2 deliveries, a balance fault in a Curiosity S01 and a disagreement over secondary-development rights. The buyer, which ordered two P2 units for 580,000 yuan and later two S01 units at 160,000 yuan each, had refused research-only terms for software access. Micbot's October 6 statement said it supports customer secondary development but requires clear authorization and protection of core technology.
Donghua University researchers spun transistors, resistors and capacitors into a continuous fiber that functions as a circuit, using microfluidic encoded spinning, according to a Nature Electronics paper. The fibers integrated multicolor electroluminescence, analog and digital logic, and non-contact spatial sensing, and in demonstrations guided a robotic gripper and let a finger control a robotic arm and drone without touch.
A Chinese-led team published in Cell a three-dimensional spatiotemporal cell atlas covering rice from germinating seed to grain fill, along with a public portal and the RICE scGPT single-cell foundation model. The atlas combines single-nucleus RNA sequencing with BGI's Stereo-seq spatial transcriptomics across 10 organ and tissue types and 61 stages, defining 119 cell types and 133 subtypes.
Shanghai AI Lab has open-sourced Intern-Decision, a family of 0.8B, 2B and 4B parameter models that return structured decisions with probabilities instead of free text. The developers self-report that the 4B model averages 90.02% accuracy across seven test suites, ahead of a commercial reference model at 88.74%, with about 44 milliseconds of local latency on a single RTX 4090. Weights are on Hugging Face, and MetaX says the models run on its hardware from launch.
Tencent introduced roughly eight vertical AI agent products between July and September, each built on an existing business such as games, music, WeChat or video conferencing. CraftBuddy generates 2D and 3D games from natural-language prompts with more than 40,000 prebuilt art assets, while MusicBuddy adds an AI chat panel to multitrack music editing. The products extend the "Buddy" family alongside WorkBuddy and CodeBuddy.
ByteDance's Doubao app is now available as a native client on Huawei HarmonyOS PCs, shipping with Doubao-Seed-2.1 Pro and Doubao-Seed-2.1 Turbo. The app offers Chat and Work modes, with Work supporting goal-based tasks, file and folder access, scheduled tasks and parallel browsing tabs. Users can download it by searching for Doubao in the HarmonyOS app store.
The Chinese Academy of Sciences Hong Kong Innovation Research Institute's CAIR centre launched CARES 4.0, a multimodal clinical agent system built on a Harness agent framework, at the 4th Hong Kong clinical robotics and embodied AI symposium. Unlike its predecessor, which only answered questions, CARES 4.0 can carry out tasks such as generating department reports and assisting diagnosis from CT, MRI, ultrasound, endoscopy and EEG data. The framework and general toolkit are released, with the medical toolkit to follow.
Huawei launched the OceanStor M900 and Nvidia introduced CMX Context Memory Storage to manage inference KV Cache in tiered storage, reducing repeated computation on cache hits. Sources say SSD requirements for this workload are not standardized: some plans have moved from 3 DWPD to 9 DWPD, and the industry may need 6–12 months to reach a common product standard.
Zhang Lei, then a recent PhD graduate working mainly on image retrieval, beat a team led by face recognition expert Li Ziqing in a 2002 Microsoft Research Asia contest to build a face detection system. He cut the false-positive rate from roughly 50% to 1%, and the algorithm was integrated into Windows. Several members of his team later became prominent figures in China's technology industry.
At IROS 2026, Simplexity Robotics presented a robot that autonomously tends two CNC machines, grasping workpieces, loading them into chucks, and placing finished parts, with 0.5 mm clearance. The system was trained on only 600 real robot trajectories, using the SimpleWAM world action model, DRAM recurrent memory, and DPE candidate action scoring. The talk was reported as an edited transcript from the conference presentation.
Qi An Xin ranked first in China's network information security market with 4.39 billion yuan in revenue in 2025, according to a CCID Consulting report on a 93.08 billion yuan market. The company also led the endpoint security, security management platform, and security services segments, with a 17.5% endpoint share and 18.3% security management platform share.
Benewake CEO Li Yuan says lidar could serve as a learning source for robot models, not just an obstacle-avoidance tool. The company shipped 250,000 robot lidar units in the first half of this year, and robotics now accounts for over half of its revenue.
AIVA unveiled its first mass-production car, the AIVA ME7, at Paris Fashion Week, marking its global debut. The source provides no further specifications, pricing, or availability details.
A paper introduces HERMES, a harness that pairs each repository component with a resident LLM and uses dependency-aware activation and failure diagnosis. With the same model and effort setting, GPT-5.6 Sol's whole-repository migration score rose from 6.5% to 31.0% when Codex was replaced by HERMES. Across four software engineering benchmarks, HERMES beats matched baseline harnesses by 12.4 points on average, and Qwen3-8B components come within 4.5 points of an all-GPT-5.6 Sol setup while cutting Terminal-Bench 4.0 inference cost by 26.2%.
Laya is a 421-million-parameter non-autoregressive decision engine from Convai Innovations that returns calibrated option probabilities in a single forward pass with zero output tokens. This tutorial tests its zero-shot accuracy, probability calibration, temperature fitting, and abstention gating on the CLINC150 banking intent dataset.
Caterpillar and CoreWeave are working to cut the time needed to train construction machines from months to hours, according to Caterpillar's Brandon Hootman. Caterpillar's ecosystem holds about 18 petabytes of federated machine, dealer and customer data, and a single machine can generate terabytes of LiDAR, camera and control data in a day. The partners use AI models, working with Nvidia, to annotate incoming field data so it can feed simulation and training within the same workday.
Atlassian announced the Agentic Multiplayer Protocol (AMP) at Team '26 Europe, a platform update designed to let humans and AI agents collaborate on shared context and tasks in one governed workspace. Agents receive administrator-assigned identities with defined authority and scope, and the company's Rovo Work mode handles multi-step tasks that humans review and approve. Atlassian's MCP server now exposes 200 tools and handles roughly 15 million tool calls daily.
Kore.ai launched Autoloop, an optimization engine that automatically adjusts AI agents built on its Kore.ai Agent Platform to meet business-set goals, including after deployment. The engine scores each proposed change against goals such as task completion, business-rule adherence, accuracy and cost. Autoloop is available now to all customers on the Artemis edition of the Kore.ai Agent Platform.
Infor is building industry-specific AI agents on its Infor OS foundation and open architecture, according to CEO Kevin Samuelson. Infor says two in three businesses find off-the-shelf AI does not adequately address their industry's needs. The company pairs customers with forward-deployed engineers, and Samuelson says prototypes can now take one to three weeks.