All AI news

Seven-layer outline of Carbon-Silicon Doctrine: Cross-Domain Migration Governance Code published in 188 installments

The seven-layer outline of the Carbon-Silicon Doctrine: Cross-Domain Migration Governance Code has been published, with 188 main installments and four additional negative-transfer prototypes, NT1–NT4, serving as a zero-numbered baseline and excluded from the main series. The layers are: Migration Incident Analysis (001–021), Migration Technical Mechanisms (022–051), Migration Rights and Responsibilities (053–081), Cross-Industry Migration Governance (082–109), Cross-Domain Top-Level Oversight (110–136), Civilization-Level Migration Risk Control (137–163), and Final Code Locking (164–188).

Migration Incident Analysis | Boundaries for attributing incident entropy growth

The first public baseline in the seven-layer Carbon-Silicon Doctrine: Cross-Domain Migration Governance Code has been finalized. It defines four forms of negative transfer, NT1–NT4, and expands governance from performance monitoring to cross-auditing causal integrity, permission boundaries, and temporal stability. It defines 'incident entropy growth' as a situation in which surface metrics remain normal after AI migrates across domains while underlying causal chains, semantic alignment, or permission boundaries silently break. It classifies seven forms of hidden negative transfer, including cascading drift, covert privilege smuggling, and accumulated timing deviations.

Migration Technical Mechanisms | Root causes of negative transfer

The second installment of the seven-layer Carbon-Silicon Doctrine: Cross-Domain Migration Governance Code proposes a standardized quick-assessment baseline for four types of negative transfer: NT1 mechanism mismatch, NT2 semantic drift, NT3 unknown completion, and NT4 power smuggling. It argues that negative transfer persists because the industry judges migration by fit and generalization metrics without checking mechanism equivalence or semantic parity. The installment also breaks down six native failure paths and proposes two foundational mechanisms: knowledge lineage and the copper-vein interface.

Carbon-Silicon Doctrine, layer three: establishing migration rights and responsibilities

The third installment of the seven-layer Carbon-Silicon Doctrine: Cross-Domain Migration Governance Code proposes a layer for establishing migration rights and responsibilities. It introduces the Knowledge Package (KP) framework for domain-wide rights determination, precisely dividing responsibilities among developers, migration operators, deployers, and auditors, and establishing lists for permitted, restricted, and prohibited migrations. It sets special red lines for the highest-risk NT4 power-smuggling transfer, prohibiting implicit inheritance of weights and cross-domain reuse of review permissions, while establishing the destination domain's final veto authority.

Carbon-Silicon Doctrine, layer four: cross-industry migration governance and calibration of industrial risk baselines

Layer four of the seven-layer Carbon-Silicon Doctrine, the Cross-Industry Migration Governance layer, has been published. It proposes a three-dimensional risk classification model that takes the highest rating across reversibility, scope of impact, and autonomous permissions, dividing migration into freely permitted, restricted, and absolutely prohibited tiers. It covers 11 verticals including healthcare, finance, education, justice, government, and media, establishing a dedicated hard risk-control baseline for each. It is positioned as a standardized industry record for corporate compliance, regulatory audits, and judicial use.

AI advances in robotics will not change your life anytime soon

An article produced by MIT Technology Review with Aventine says that while Elon Musk claims Tesla Optimus could eventually automate nearly all human labor for as little as $20,000 per robot and predicts public sales before the end of 2027, many robotics researchers are skeptical. Yann LeCun said at Davos that no humanoid robotics company knows how to make robots smart enough to be truly useful.

AI hacking tool helps one attacker breach multiple South Korean banks

CrowdStrike reported that a suspected Chinese-speaking user attacked multiple South Korean financial institutions from late September to early October 2026 and stole large amounts of data. More than 25,000 records containing names, contact details, income, and credit limits were stolen from Shinhan Bank alone.

Teen stranded on cliff after using Claude to plan hike says he will not use AI for route planning again

A 16-year-old used Anthropic's Claude to plan a hike on Vancouver's Crown Mountain and became stranded on a five-foot-wide ledge on the steep rock face known as Widowmaker Arete, a route requiring climbing gear. North Shore Rescue eventually airlifted him to safety.

CCID report: Qi An Xin leads China's cybersecurity market for seventh year, also retaining top spots in three segments

CCID Consulting's 2025–2026 China Cybersecurity Market Research Annual Report says the market reached RMB 93.08 billion in 2025 and Qi An Xin ranked first by market share for the seventh consecutive year, with RMB 4.39 billion in revenue. It led in endpoint security, security management platforms, and security services. Its 17.5% share in endpoint security led for an eighth year, its 18.3% share in security management platforms led for a fifth year, and it led in security services for an eighth year.

Li Yuan of Benewake: making eyes for machines, and eleven years of choosing the hard path

Benewake founder Li Yuan has spent eleven years building lidar for drones, trains, cars, and robots. In the first half of this year, Benewake delivered 250,000 lidar units for robots; robotics accounted for more than half of company revenue, and 2025 robotics revenue doubled from 2024. Li believes measurements such as distance and position from lidar should become learning material that helps machines understand the physical world.

Yunqi Conference outlook: Alibaba's roadmap for next-generation Qwen and multimodal models

At the Yunqi Conference, Liu Daheng, who leads the Qwen large language model, outlined the roadmap: the Qwen4 series is coming soon, followed by Qwen4.5 and Qwen5, which are planned to move toward 5T–10T parameters. Qwen3.8 has already reached 2.4T parameters.

Former Tencent Hunyuan vision-model lead Hu Han starts a multimodal AI venture, targeting a valuation of hundreds of millions of dollars

Hu Han, formerly head of the algorithm center for Tencent Hunyuan's vision foundation models, has made progress on a new venture focused on multimodal AI. It plans to raise tens of millions of dollars at a target valuation of several hundred million dollars, with Vision Plus Capital as the financing advisor. Multiple investment firms have approached Hu over the past month. He believes the model could eventually be sold to large technology companies such as DeepSeek.

vLLM v0.31.0 ships with DeepSeek-V4.1-Flash support and fast restart

vLLM v0.31.0 is live, spanning 717 commits from 307 contributors, including 96 first-timers. Highlights include DeepSeek-V4.1-Flash support with FlashMLA mega attention and NVFP4 KV caching, a vllm preload fast-restart path that keeps weights in GPU memory, Model Runner V2 speculative decoding, and large-scale serving features such as MoonEP balanced all-to-all and DeepEPv2. Full release notes are on GitHub.

How AVEVA is building a safer path to deploying industrial autonomous AI

Arti Garg, AVEVA's chief technology expert, says industrial AI is shifting from predictive analytics to foundation models, physical AI, and agentic AI, and requires connected data, governance frameworks, and human oversight for safety. AVEVA's responsible AI framework emphasizes safety, efficiency, and human oversight, arguing that AI should augment rather than replace people in critical decisions. Garg also participates in an IEEE working group developing a standard method to measure AI's environmental impact across power, energy, resources, water, and carbon.

Naive AI raises $400 million seven months after founding, backed by an eight-person team led by Dai Jifeng

Tsinghua-affiliated Naive AI, registered as Beijing Zhiyan Huisheng Technology Co., Ltd., completed three funding rounds totaling $400 million in seven months. Its post-money valuation is $1.42 billion, with investors including Tencent, Boyu, Sequoia China, IDG, and Matrix Partners.

What I know about Zhang Lei: an MSRA algorithm race and half the Chinese AI scene

In 2002, Microsoft Research Asia held an internal competition for face-detection algorithms. Zhang Lei, less than two years out of his PhD, led interns Xiao Rong, Zhu Long, Chen Longbin, and others in reducing the false-positive rate from 'one false alarm for every positive' to 1% in three months, beating the team of face-recognition expert Li Ziqing. Their approach was eventually integrated into Windows. During 20 years at Microsoft, Zhang also built a celebrity-recognition system that could identify 200,000 people with 99% accuracy and released the MS Celeb dataset. OpenAI later drew on its image-text matching approach when training the 400-million-example CLIP model.

Zhijian Dynamics' Feng Zongbao: Achieving robot-built robots with 600 trajectories under a 0.5-millimeter gap

At the IROS 2026 venue, Feng Zongbao, head of reinforcement learning at Zhijian Dynamics, presented a solution that enables a robot to autonomously tend two CNC machine tools, completing grasping, loading, insertion, removal, and placement under the condition of only a 0.5-millimeter gap between the workpiece and the chuck, with training using only 600 real robot trajectories.

Tsinghua AIR's first PhD graduate Li Jianxiong founds BASAL Intelligence; 5Y Capital leads investment in action-native embodied models

Li Jianxiong, the first PhD student at Tsinghua University's Institute for AI Industry Research (AIR), founded embodied AI company BASAL Intelligence. 5Y Capital led the round, with Ivy Capital, Linear Capital, and Huashan Capital participating. Tsinghua AIR associate professor Zhan Xianyuan is the chief scientist.

Anthropic releases Claude Haiku 5.5 at GPT-6 Luna pricing

Anthropic released Claude Haiku 5.5, the first Haiku-series update in about a year. Its pricing matches OpenAI's GPT-6 Luna. On the same day, Anthropic also cut prices for Sonnet 5.5 and its subscription plans.

Huawei has separate storage for KV cache, but specifications for dedicated SSDs are still undecided

At its Huawei Connect conference, Huawei launched OceanStor M900. Nvidia has also introduced CMX Context Memory Storage. Both move reusable KV cache from inference into tiered external storage, trading storage for compute to reduce costs.

Galaxy General, Tsinghua, and collaborators propose LATENT, teaching humanoids tennis from imperfect human data

The GaLbot team at Galaxy General, together with Tsinghua, Peking University, and others, proposed LATENT, which trains humanoid robots to play tennis using about five hours of imperfect motion data from five amateur players. The method first trains a motion tracker and compresses it into a latent action space with a Latent Action Barrier constraint. On a Unitree G1, it achieved a 90.90% success rate on real forehands and 77.78% on backhands.

Analysis of 1,933 IROS 2026 papers finds six emerging shifts in robotics

IROS 2026 accepted 1,933 papers. A multi-label analysis counted about 809 papers in Robot Learning / Embodied AI and 564 in Navigation / Planning.

IROS 2026 awards: LT-Mem wins best paper; humanoid tennis and end-of-arm tray work also recognized

The IROS 2026 Best Paper Award went to Yumin Lee and colleagues for LT-Mem. The work addresses temporal forgetting in long-running robots with a three-layer structure of Live Memory, Delta Memory, and Meta Memory, and the concept of volatility.

CARES 4.0 launches in Hong Kong: CAS's CAIR introduces a multimodal clinical agent system

The Centre for Artificial Intelligence and Robotics (CAIR) at the Hong Kong Institute of Science and Innovation, Chinese Academy of Sciences, launched CARES 4.0, a multimodal clinical agent system, at the 4th Hong Kong Embodied Intelligence and Medical Technology Forum. Built on the Harness agent framework, it uses in-house foundation models for CT, MRI, ultrasound, endoscopy, and EEG, and has been validated at several top-tier hospitals. The CARES 4.0 website is open. Its framework and general toolkit are available, medical toolkits are being released in stages, and its in-house open-source ultrasound and surgical-video models will be integrated and made freely available.

KargoBot launches large-scale AI transport network, cabless heavy trucks hit the road

KargoBot announced on September 22 in Qipanjing Town, Ordos, the launch of a large-scale AI transport network, putting manned heavy trucks, driverless trucks with retained cabs, and cabless transport robots into the same transport system. Its driverless transport fleet has reached the scale of about one hundred vehicles, with operations covering 6 provinces including Inner Mongolia, Xinjiang, Gansu, Shaanxi, and Shanxi, and it cooperates with 30 customers, 8 automakers, and about 30 ecosystem partners.

Dialogue with ONTIME Mobility COO Han Feng: The stronger the world model, the more real data becomes the true moat

ONTIME Mobility COO Han Feng believes that as world model capabilities grow stronger, the importance of real data will not diminish: synthetic data is responsible for expanding known long-tail scenarios, while real data is responsible for discovering unknown problems and providing calibration and validation. ONTIME Mobility does not train large models; instead, it collects real data through routine mobility services, Robotaxi, and offline operations and maintenance networks, and delivers training materials to autonomous driving, world model, and embodied intelligence companies through a data closed loop.

Wu Jiajun at IROS 2026: structured representations are scaffolding for robot reasoning

At the RoBoWoMo workshop at IROS 2026, Stanford assistant professor Wu Jiajun presented 'Building Physical Agents via Structured Representations'. He argued that structured representations are not answers to hard-code into a system, but scaffolding that helps robots learn efficiently and communicate with people.

vLLM-Omni technical report proposes a unified serving runtime for omni-modality generation

The vLLM project shared the vLLM-Omni technical report, describing a unified serving runtime for omni-modality generation. It argues that speech assistants, visual generation, world models and robot loops diverge from a single text decode loop into multi-stage autoregressive pipelines, iterative diffusion and stateful sessions, which LLM servers and diffusion stacks each cover only partially. vLLM-Omni acts as a shared control plane with an orchestrator, specialized engines, a connector and one session path. Paper and repo links are included.

Anthropic releases Claude Haiku 5.5, matching GPT-6 Luna on benchmarks while cutting prices by 90%

Anthropic released Claude Haiku 5.5. Its official benchmark scores exceed GPT-6 Luna across the board, making it the new small-model gatekeeper, replacing DeepSeek V4.1 Flash and GLM-5.3-Flash.

OpenAI starts rolling out GPT-6 to free ChatGPT users; chat can generate interactive interfaces

OpenAI announced that GPT-6 is beginning to roll out to free and Go ChatGPT users, replacing GPT-5.6 Luna with GPT-6 Luna. Paid users on Plus, Pro, and other plans will use GPT-6 Sol.

The same story is shown in the picks as 《OpenAI 在 ChatGPT 全球上线 GPT-6 并推出 Intelligent UI》

Kaiming He team proposes NAT-ARC, using ImageNet pretraining to solve the ARC challenge with vision models

Kaiming He's team proposed NAT-ARC, a purely visual approach to solving ARC. By adding ImageNet MAE pretraining before the VARC pipeline, its best single model achieved 63.4% pass@2 on ARC-1, rising to 70.2% with an ensemble.

StepFun's first foundation-model-native agent phone, STEPX Neo, to launch on October 13

StepFun Terminal will hold a 'Ready Builder One' launch event in Shanghai on October 13 to introduce its first foundation-model-native agent phone, STEPX Neo. The model, operating system, and hardware were all designed natively for agents. The event will also announce the company's latest ecosystem partnerships.

Sometimes I read on flights, and sometimes…

Sometimes I read on flights, and sometimes…

Microsoft unveils Surface Laptop Ultra and Windows 11 updates

At its first in-person event in two years, Microsoft launched the new Surface Laptop Ultra and announced a series of Windows 11 changes, while outlining its vision for local AI and agent workflows in personal computing.

Ben Affleck discusses machine learning and visual-effects workflows

Ben Affleck said he became more interested in technology after moving from film to digital and can write 'pretty terrible Python scripts', since visual-effects workflows have included machine learning for years. He described convolutional neural networks as predecessors to transformers that process tensors and perform edge detection and feature extraction, making it easier to isolate green-screen footage and replace the background.

Aravind Srinivas Calls Perplexity Decider Demo “Cool”

@sanchitmonga22 says Perplexity’s Decider model speedruns “Who Wants to Be a Millionaire?” and calls it “the best decision model.” Aravind Srinivas calls the demo cool: https://twitter.com/sanchitmonga22/status/2107959692458922157

Artcraft uses Claude Opus 5.5 to create open-source alternatives to seven Adobe apps

Artcraft shifted from controllable AI image and video editing tools to an open-source app suite. Using Claude Opus 5.5, it recreated the interfaces and tools of Photoshop, Illustrator, Premiere, Lightroom, After Effects, InDesign, and Acrobat Pro in Rust, and offers WebAssembly versions that run in a browser.

OpenAI rolls out GPT-6 and Intelligent UI in ChatGPT for everyone

OpenAI says GPT-6 and Intelligent UI are now rolling out in ChatGPT for everyone, offering fast, interactive answers and tools for users’ tasks. Mark Chen encourages users to try Intelligent UI, saying every ChatGPT completion now has the potential to create a custom UI and calling it phenomenal for learning new things. https://twitter.com/OpenAI/status/2107894997538525580

Noam Brown says LLMs surpass top human experts on some research problems

Noam Brown argues that LLMs have crossed an important threshold by surpassing top human experts on some research problems, with the gap between slightly below and slightly above expert performance helping explain the seemingly sudden surge in math results. He emphasizes that capabilities remain jagged and models are still worse than humans in many ways, but expects continued improvements to bring breakthroughs in other domains and a new era of scientific discovery.

Great to partner with NVIDIA as we kick off this next chapter for Windows. Thank you @JensenHuang and @sriramk for joining us today!

Great to partner with NVIDIA as we kick off this next chapter for Windows. Thank you @JensenHuang and @sriramk for joining us today!