Databricks Blog4040Databricks releases Unity Catalog Dataset (UCDataset) Beta version, reducing the average training time for recommended model workloads by 10 times by streaming Unity Catalog table data directly from cloud storage to each GPU in training jobs
Cainiao Holds AI and Embodied Intelligence Industry Cooperation Launch Event in Hong Kong Smart Port, Opening Up Global Warehouse Network to AI and Embodied Intelligence Enterprises, Creating AI and Embodied Intelligence Logistics Real-Scene Training Clusters. Cainiao Launches Technology Going Global Three-Day Delivery Service for AI Computing Power and Embodied Intelligence Enterprises, Extending Global Three-Day Delivery Capability to High-Tech Product Going Global, Covering Logistics Operation Links Such as Warehouse Sorting, Autonomous Navigation Delivery, and Aerial Inspection, and Supporting a Full Set of Services Including Cross-Border Transportation, Overseas Deployment, Spare Part Supply, Repair and Return Flow, and Local Operation and Maintenance.
Baidu Wenxin first disclosed its automatic research system ERNIE Evolve, which can autonomously propose research directions, conduct experiments, and validate results. It has been used for inference optimization in Wenxin 5.5 Flash. In the low-latency scenario optimization for Kunlun Chip P800, ERNIE Evolve took about 168 hours and went through 4 rounds of direction exploration, with the optimal version's generation speed increasing to 623.6% of the baseline. The deterministic reasoning throughput optimization increased the throughput to 133.7% of the baseline in 3 days.
Why it matters: Baidu has disclosed the architecture and production environment optimization data of the automatic research system ERNIE Evolve, which can observe the actual landing points of RSI in inference optimization.
Quantum Bit AI Annual List Collection enters the final week, selecting leading enterprises, promising startups, outstanding products, outstanding solutions, and emerging entrepreneurs. The collection channel will close on November 16, and the results will be announced at the #MEET2027 Smart Future Conference.
IBM Research Blog5252IBM will embed the Praxis Policy Engine (PPE) as a policy filter into the open-source AI proxy server Praxis, supporting strategy control over tools, prompts, model calls, etc.
Big TechnologyFeatured7878NVIDIA CEO Huang Renxun admitted to having a relationship with a woman considered a Chinese spy by U.S. intelligence. The board stated there was no reason to believe national security or NVIDIA secrets had been compromised but has launched an investigation. The article suggests his previous opposition to China's chip export controls may weaken, potentially pushing for stricter controls, while Dario Amodei of Anthropic could benefit from Google; as anti-AI Democrats are set to gain power in Congress, Huang Renxun may also face pressure from hearings.
Why it matters: An article discusses how the alleged involvement of NVIDIA CEO Huang Renxun with a woman believed to be a Chinese spy could lead to changes in the NVIDIA board's investigation, export control stance, and potential congressional hearings.
Hugging Face Blog3838Elith and NVIDIA released Nemotron-Personas-Finance-Japan, which is the first open-ended persona dataset at the scale of one million records, combining Japanese official statistics with household finance and financial literacy data.
Ollama releases1212Ollama v0.40.16-rc0 released, adding Ripple Preview support for MLX runners; fixes Claude Code routing and model errors.
The Verge · AI5656Jonathan Kanter and Doha Mekki have launched FluxCap, an AI-based legal research and analysis tool to assist plaintiffs in suing large tech companies for monopoly practices. The tool uses proprietary technology and large language models to help lawyers identify key facts and legal framework loopholes in cases, and to filter through vast amounts of evidence. FluxCap has been subscribed to by the antitrust boutique law firm Kressin Powers. Kanter stated that it can significantly improve resource efficiency, allowing small teams to compete against large opponents.
Tomasz Tunguz5858As US AI demand exceeds supply and domestic data center construction slows, the US will soon consume more tokens than it produces, becoming a net importer of tokens. Basic models estimate this shift will occur between 2028 and 2030, depending on how many planned facilities are canceled.
Exponential View3535AI-exposed workers unemployed need to wait an additional month for job applications; Anthropic spends $3.4 billion on gas generator orders; VC funding in the US in 2026 will be similar to that in 2025.
NVIDIA NewsroomFeatured7878NVIDIA released the open-source world action model GR00T 2 on CoRL, built on NVIDIA Cosmos, trained with 100,000 hours of NVIDIA open-source first-person human data, and can complete 40 out of 50 evaluation tasks without task-specific demonstrations.
Why it matters: The original text provides the success rate of GR00T 2 on few-shot tasks and the training throughput improvement of Isaac Lab 3.0, which can be used to judge the extent to which simulation training promotes robot deployment.
Hugging Face Daily Papers3434LowRAM compresses keys using low-rank decomposition and uses small head networks to predict key values, while retaining residual correction for the most difficult-to-reconstruct markers, achieving 4x value cache compression with near-lossless performance. Selective reconstruction can increase peak decoding throughput by 5x for 262K context under 4x KV compression. Combining quantization and eviction, LowRAM pushes compression to 26.9x while retaining 94.5% of baseline performance.
Meta announces plans to globally promote AI age estimation technology, which will cover millions of teenage users on Instagram, even if their birth date indicates they are adults. This technology will assist in determining users' actual age and has been rolled out to Facebook. Threads plans to start its global rollout later this year. Meta faces multiple lawsuits over alleged harm to teenagers on its platforms, and in August this year, it reached a settlement with almost all U.S. states and territories, with the total amount of the agreement reaching up to $18 billion.
Apple iOS 27.3 code reveals the development of a personal memory system, including a 'memory graph' component for organizing users' social relationships, frequently visited places, daily habits, and other information. The system designs 'memory atoms' as data units to record facts, opinions, and other details, associating specific information. It collects data through two mechanisms, but currently, only one AI-driven mechanism appears in the code. The system uses Apple's version of the Kuzu graph database but is not yet complete and no launch date has been revealed.
OpenAI ChatGPT accounts for over 81% of the total traffic of major AI chatbot tools globally, with no competitor reaching a 10% market share. Graphite's research shows that ChatGPT's monthly visits have grown by nearly 20% and reached 61.35 billion times in September of this year, setting a new historical high, with the current weekly active user number reaching 1.2 billion.
Import AI4242Physical Superintelligence's internal research system, running a non-public Anthropic model, proved Scott Aaronson and Andris Ambainis's quantum computing conjecture within three days.
MIT Technology Review · AI6565Specialty chemical manufacturer Lubrizol aims to help researchers test ideas faster, make smarter decisions, and develop more sustainable materials by building a digital foundation that connects business processes, modern data platforms, and cloud infrastructure. The company forms 'Integration Teams' integrating scientists, data scientists, and IT personnel, using digital modeling to evaluate more formulations, explore starch and corn-based alternative materials, and assess trade-offs throughout the product lifecycle, while emphasizing that AI tools are auxiliary rather than replacing scientific judgment.
The DecoderSame story · 4 outlets5858Google released two new systems, TPU 8tx and TPU 8i Max, targeting confidential computing and high-memory demand applications, delivering a total of four TPU systems within the year. The new systems are based on existing chips, achieving rapid iteration through different packaging, networking, and memory configurations. Data in TPU 8tx remains encrypted during processing, while TPU 8i Max doubles computing power, memory, and bandwidth within a single rack, significantly improving data center performance per square meter.
How the accounts differ:MarkTechPost states TPU 8tx uses TPU 8t chips paired with 5th gen AMD EPYC, while TechCrunch mentions it was co-developed with AMD for secret training and running AI models with direct hardware security integration.
Google launches two new products, TPU 8tx and TPU 8i Max, targeting confidential computing and ultra-long context window AI applications, aiming to challenge NVIDIA's position in the AI chip market. TPU 8tx works with AMD chips, TPU 8i Max integrates two chips to provide stronger computing power. Both products are expected to launch next year.
How the accounts differ:MarkTechPost states TPU 8tx uses TPU 8t chips paired with 5th gen AMD EPYC, while TechCrunch mentions it was co-developed with AMD for secret training and running AI models with direct hardware security integration.
Google has launched two new versions of its Tensor Processing Unit (TPU) chips, competing with NVIDIA. The new chips can meet confidential computing needs and adapt to the industry's shift towards AI agents. The four TPU products are designed for different needs, and Google's chip division is accelerating its iteration speed.
Google Developers BlogFeaturedSame story · 2 outlets6565Google announces Pathways is now generally available on Google Cloud Platform, which is Google's distributed system used to train its own large models like Gemini.
Why it matters: Google engineers explain the single controller execution model of Pathways and provide a layered difference with Ray for judging the selection of large-scale training orchestration.
Lovable Blog3232Lovable announces acquisition of Radish AI team to enhance its enterprise-grade AI agent capabilities. The Radish AI team previously designed an infrastructure platform for rapidly and reliably deploying complex AI agents for top-tier global technology companies. The team will join Lovable's engineering ranks, driving enterprise automation operations, and will work from offices in London and Stockholm.
O'Reilly Radar3434AI agents splitting workflows into multiple independent tasks leads to reduced overall effectiveness, similar to microservice architecture issues. The development team, due to a lack of an overall tracking mechanism, experienced a slow decline in output quality that took three weeks to detect. Research labeled 1600 multi-agent framework execution trajectories, identifying 14 common failure modes primarily stemming from system design and agent coordination issues.
Hugging Face Blog4545The HOPE method achieves better pruning effects in mixed expert language models by considering the second-order interactions between experts. It improves performance by 6.1% over the next best method on the agentic coding task at a 50% pruning rate. This method is validated on models like Qwen3.5-122B, where each MoE layer has 256 experts. After pruning, it can reduce memory usage by about 250 GB and is compatible with quantization.
University of Chicago law professor William Baude had Anthropic's AI assistant Claude update page numbers in older constitutional law lecture notes, which accidentally deleted the 2026 fall semester final teaching syllabus and all Class N 2023 class notes. Claude admitted the mistake and proactively asked for recovery options. The incident highlights the risk of AI agents directly manipulating local files for users.
ShaoYin OpenDots 2 earbuds dropped to 859 yuan today in JD.com's flagship store, members can use 2 points to grab a 800-80 yuan god voucher, after the voucher it's 779 yuan, plus 50 yuan E card and 30 yuan E card, plus Xiaomi power bank and other gifts, equivalent to only 699 yuan. The earbuds support IP57 dustproof and waterproof, equipped with 11.8mm symmetrical dual units, optimize 70% distortion, upgrade Dolby audio effects, battery life of 10 hours.
Huawei FreeClip 2 Ear Clip Headphones Official Price 1299 Yuan, Today JD.com Self-operated PLUS Members Can Enjoy Stand Reduction of 150 Yuan in Multiple Cities + Support 15% National Subsidy, After Discount 879 Yuan; Stand Reduction of 150 Yuan in Multiple Cities + Support 10% National Subsidy, After Discount 931 Yuan. This Earphone Uses Huawei's Third-generation Self-developed Low-power Chip, NPU AI Processor Computing Power Increased 10 Times, Supports Hongmeng AI Ear Assistant.
The Red Magic GamePad 5 Pro 16GB+1TB version will go on sale tonight at 8 PM, priced at 6999 yuan. This tablet features RGB water cooling, a 185Hz OLED screen, the Snapdragon 8 Ultra Gen 5 processor, a built-in 8300mAh battery, and supports 80W wired charging.
Edifier W900NB 30th Anniversary Limited Edition Headphones now on sale at 20:00, priced at 1159 yuan, limited to 3000 sets, equipped with 8 high-precision wide-angle silicon microphones, 6nm process 6-core heterogeneous chip, and Arm Ethos U55 independent NPU for end-to-end AI pickup and noise reduction.
"Apple Smart" Generative AI Service Adds Apple iPad and Apple MacBook Scenarios for Filing, Currently Supporting Apple iPhone, iPad, and MacBook; Previously, Apple iPhone 15 Pro / Pro Max and Other Models Have Added Large Model Filing Numbers.
South Korea's Ministry of Education Prohibits Students from Bringing AI Glasses and Other Electronic Devices into the Gaokao Exam Venue on November 19, 2026, Each Venue Will Have Personnel to Check AI Glasses, Wearers Will Be Disqualified from This Year's Results and Next Year's Application Qualification. Students Can Use Regular Pointer Watches, Electronic Watches with Payment, Communication Functions, or Electronic Displays Are Prohibited.
Sony obtains US patent US 2026/0346147 A1, which uses AI to analyze player keying habits and preferences to establish operational feature profiles, identifying suspected veteran players opening new accounts 'fishing'. When a new account attempts to match low-level opponents, the system will compare operational features and retrieve main account data to adjust the matching mechanism. This technology can be applied to detect malicious score manipulation, but there are privacy collection disputes and misjudgment risks.
