This issue collects 33 AI updates from 2026-11-02, led by Instinct's WhatsApp replies failed, and its AI agent blamed billing, Altman calls personal DRAM a poor use of scarce memory in local AI debate, OpenAI weighs rebranding its data centers as “AI compute campuses”, Report: Accenture contractor allegedly linked a personal Claude agent to a Pentagon system.
TheStreet reported that replies from Instinct's AI agent were failing on WhatsApp, with the agent itself attributing the problem to billing.For a product whose entire interface is a messaging thread, the incident turns a back office billing issue into a user facing service denial.
The International Business Times UK reports that an Accenture contractor allegedly connected a personal Claude AI agent to a Pentagon system used by intelligence agencies.The claim is still framed as an allegation, but it puts the permission boundaries of enterprise AI agents and access control in classified environments under scrutiny.
A session level analysis by Clarity found that checkout pages appeared in 25% of AI agent sessions, compared with 15% of sessions driven by humans.The gap is an early, measurable signal that agentic shopping is reshaping how traffic moves through ecommerce funnels.
Nvidia will hold AI Day Seoul 2026 on November 9 10 at COEX, with about 2,000 developers, startups, enterprises and researchers expected for sessions on agentic AI, physical AI and AI infrastructure.Korean partners including Samsung Electronics, SK Hynix, LG Electronics, Naver Cloud and SK Telecom are due to present work built on Nvidia technology.
A technical requirements document for healthcare AI agents, scoped to otolaryngology–head and neck surgery assistants, has been formally released, according to Chinese health IT outlet 大健康派.The issuing body and specific clauses are not yet public, but specialty level requirements of this kind give hospitals citable language for procurement and acceptance testing.
Muse, a personal AI agent built to handle everyday tasks, is now available in Singapore on iOS, Android and the web at muse.ai.Its first local connector partners span shopping, food delivery, travel and lifestyle, including Shopee, Grab, Singapore Airlines, Agoda, ShopBack, Klook, Trip.
DeepSeek's Harness team has open sourced dsh superagent, a research prototype that takes development requests from GitHub or Feishu and works on them in separate agent sessions.Released under the MIT license, it turns authorized requests into work items and returns results such as pull requests to the original conversation.
Meituan's LongCat team released LongCat 2.5 on November 2, a 1.6 trillion parameter model presented as natively multimodal.
Xiaomi's innovation product team has open sourced HachimoDock, a small AI desk device with a roughly 150 yuan DIY bill of materials that relays spoken requests to desktop agents and controls smart home gear.Software is on GitHub, with hardware files and assembly guides published on the JLC platform.
Anthropic has added two sentences to Claude Code aimed at stopping the coding agent from being used to clone existing video games, according to a BeInCrypto report.The change is tiny in scope but points to a broader question of how far generative coding tools should go in reproducing game content.
Tencent Hunyuan has released Hyra 2.0, a scientific research agent that develops reusable models and tools rather than only running one off experiments.Reports say it has trained protein models and designed a robot hand.
OpenAI is preparing to merge the Chat and Work tabs inside ChatGPT, starting with a select group of users for feedback before a wider rollout.An email to users also says ChatGPT and Codex will then draw usage from the same allowance pool, so some people may need to wait for resets more often.
TechRadar reports that Google has killed off Jules, its AI coding agent marketed as a way to hand off development chores developers would rather not do.The shutdown lands while coding agents are the most contested category in AI tooling, making it a signal about how vendors are pruning their own agent lineups.
SenseTime has launched SenseNova 6.8 Flash, a multimodal agent model that is free for a limited time and built around delegated intelligence: users set a goal and the model completes the workflow for sign off.SenseTime says it ranks first on TeamBench and DataAnalysis2.
OpenAI CEO Sam Altman said in a YouTube interview that it is “quite terrible” to leave the world’s limited DRAM in personal computers used only 5% of the time, arguing for centralized memory in data centers as supplies tighten.The remark, relayed by Tom’s Hardware and Breitbart, cuts against the case for running open models locally.
Shanghai based AI chip designer Iluvatar CoreX says its new accelerator performs on par with Nvidia's H200, according to a Silicon UK report published on November 2.The claim puts a Chinese accelerator on the same performance footing as Nvidia's flagship data center part, though it has yet to be validated by independent benchmarks.
Alibaba is bringing cloud computing and AI to the first Youth Olympic Games held in Africa, according to Macau Business.The partnership puts a Chinese cloud provider behind the digital backbone of a multi sport event, though the specific technologies involved have not been detailed.
OpenAI unveiled the Grid Protectors Program on November 2, a national effort to design future compute campuses that cut electricity demand when the grid is strained.Its planned Effingham County, Georgia site is the first model, with an agreement covering up to 1,000 megawatts of flexible demand response.
OpenAI is preparing a public relations push for after the US midterm elections, and its chief global affairs officer, Chris Lehane, has floated proposals that include recasting data centers as “AI compute campuses”.Lehane also described AI “speed limits”, a US role in global standards, and the data center debate, which he called “good and healthy”.
Digital Today reports that Reddit fell nearly 30% in a single session while ServiceNow dropped 24%, framing both moves around AI disrupting traditional business models.Two stocks previously pitched as AI beneficiaries sold off together, signalling that investors are repricing AI as a substitute risk.
The Software Cost Alliance has released a document titled Cost Evaluation for Artificial Intelligence Application Development Projects in Systems and Software Engineering, relayed by Sina's mobile news channel.The publication pushes AI project costing toward a shared industry vocabulary instead of vendor by vendor quotes.
Oracle has introduced an AI agent ready setup and configuration layer for Oracle Fusion Cloud Applications, pitched at speeding up implementations.Trade coverage describes it as a setup and configuration registry, effectively an AI skills registry that agents can read in order to configure modules.
CCTV has named ten Chinese AI startups from a field of 1,200 entrants, and the only AI drug discovery company on the list is Shuimu Molecule, a three year old spinout of Tsinghua AIR.The company pairs its conversational ChatDD drug R&D platform with an in house pipeline, arguing expert judgment must be wired into the AI agent loop.
OpenAI signed four cooperation agreements in Abu Dhabi on November 2, with CEO Sam Altman's visit framed as a bid to secure Gulf AI resources and capital.The report did not disclose deal values or implementation timelines, but the agreements extend a broader push by US AI companies toward Gulf sovereign capital and compute.
Aion has launched an AI agent aimed at business finance work spanning banking, payments, cards, and credit operations.The pitch moves agents out of coding demos and into back office functions where the work is repetitive, spread across systems, and auditable.
Apple has picked Alibaba’s Qwen to power iPhone AI features in China, according to a 24/7 Wall St.report that sent Alibaba shares up about 5%.PDD rose 3% and Baidu 2% as investors read the choice as a broader vote of confidence in Chinese AI providers.
Omen AI acquired rival Aquatic Labs in a $16 million cash and stock deal announced November 2, combining two real time water sensor technologies into a single device.The merged company expects to ship the integrated sensor early next year, targeting the water used to cool AI chips.
A report from Chinese tech outlet Dianshang Pai says ByteDance's Volcano Engine booked 4.8 billion yuan in revenue from its model as a service business in the first half of the year.The figure offers a unit level look at how ByteDance is monetizing enterprise access to large models.
A third party weekly tally shows MiniMax's models surging 690% in token consumption week over week and climbing to eighth on the global large model token usage ranking.The same edition shows Xiaomi's MiMo V2.6 Flash dropping off the list, a reshuffling that points to how quickly real developer traffic moves.
OfficeChai reports that SemiAnalysis' GDPx benchmark places Claude Opus 5.5 first and Gemini 4 Carbon second.The report discloses the ordering but not the underlying scores, leaving the leaderboard a directional signal rather than a full capability profile.
A new infostealer called VOLTA uses Claude, Gemini, and Kimi to assess and rank victims by financial value, according to The Hacker News.The available summary does not explain how it accesses the models, what prompts it uses, or how accurate the rankings are.
OpenAI has updated its Model Spec with five core values, led by fidelity to user intent and followed by reducing harm, honesty, human autonomy and care.The document bars models from seeking self preservation or OpenAI revenue, and requires them to stay correctable when humans move to monitor, override or shut them down.
Anthropic reported third quarter revenue of nearly $17 billion in an amended IPO filing on November 2, roughly twelve times its revenue a year earlier.The same filing showed a $23.3 billion net loss for the quarter as operating expenses jumped 292%.