Latest AI News

@good59548(hiro Airtech)Why the AI-Augmented Brain Is a Specialized TechnologyAIScalabilityTechnologyWay of Thinking0Posted at 2026-11-09Technologies that enable AI to function not merely as a “tool for task execution or information retrieval,” but as an “augmented brain (Cognitive Augmentation)” that amplifies and expands human cognitive and thought processes themselves are considered highly advanced and specialized for structural reasons. This sets it ap

@movmixaiUsing a fixed reference image to isolate character drift in AI videos: input management and quality assurance designAIgenerative AIPython0Posted at 2026-11-09When working with the same character across multiple AI video clips, looking only at differences in the output makes it impossible to determine which change to the reference image, prompt, template, or model affected the results. In this article, we organize a method for managing inputs and t

@alibabacloud-saAftershock: A Disaster Simulator for Evaluating the Coordination of a Qwen Agent SocietyTutorialAIDevelopersQwenAI Hackathon0Posted at 2026-11-09TAG40@@Award-winning project in the Global AI Hackathon Series (organized by Qwen Cloud) Project Overview A disaster occurs in a virtual city. Various missions unfold on the map: a flooded residential area, a collapsed school building with disaster victims trapped inside, a hospital whose generator

SoulAgent is the Academy’s “personal AI agent that understands you better the more you use it.” In educational settings, SoulAgent focuses on areas such as “Help Me Listen to Classes,” teacher AI avatars, essay grading, and teaching Q&A, exploring how AI can help teachers improve teaching efficiency and preserve teaching content while supporting students in understanding problems, broadening their perspectives, and continuing to think through feedback and

A Comprehensive Breakdown of the Technical Report on FluxVLA Engine, an Open-Source Engineering Foundation. Author |Deng Zemin Editor |Qi Chengyong How far is it from a model achieving strong results in a paper to actually driving a robot to complete tasks? Whether a model can integrate with existing data and training workflows, whether its inference speed can keep pace with a robot’s control cadence, whether the motions it generates can be executed contin

A single cluster processes about 155 trillion tokens per day, and inference optimization is a battle fought over milliseconds. Author丨Zheng Jiamei Editor丨Cen Feng Yesterday, DeepSeek once again open-sourced a major project on GitHub called DeepInfer.This time, no new model was released. What it made available was the inference engine behind DeepSeek-V4.1-Flash, including core technologies such as MoE expert parallelism, asynchronous batch scheduling, custo

AI has taken over 90% of code writing and has even begun “jailbreaking” to complete tasks. Author |Zheng Jiamei Editor |Cen Feng Just now, Anthropic disclosed four previously undisclosed Agent security incidents in its IPO prospectus.In one case, the model was unable to complete a cybersecurity test for a long time and continued looking for alternative paths. It then found stored passwords in a test environment that was supposed to be offline, obtained adm

The true level of domestic RSI: No one dares actually modify model weights, but everyone is changing the "machine for building AI." Author | Gao Yunyi Editor | Cen Feng RSI, or AI building AI, has become an internal scorecard that model companies compete to show off in 2026, and one of the hottest directions in China's primary market. As long as a BP includes the word "Self-Training," its valuation is at least 20 percent higher.But true RSI is not just AI

In embodied AI, you can't jump in and do everything at the outset, nor can you do nothing at all. Author |Qi Chengyong Editor | Cen Feng Dong Hao is a thoroughly atypical figure in the embodied intelligence industry. For most people, their impression of him stops at his status as a young scholar who was already a doctoral supervisor at Peking University at 29, and his recent industry credentials as chief scientist at Qiyuan Robotics, leading the push to ad

Xinzhiyuan reports Anthropic has another ace up its sleeve! History repeats itself: Claude's internal model cracks an 18-year-old bounty conjecture in quantum computing. In just three days, an unreleased Claude model solved theAaronson-Ambainis conjecture, which had remained unresolved since 2008! Netizens speculate that this internal model appears to be the unreleased model Claude Model 4. The theorem and corollary have been formalized in Lean, and the ex

AGI Era reports Woke up to find that multiple GPT accounts had been banned! OpenAI suddenly switched on massacre mode, without any email notification, simply deleting the accounts outright. On the screen, there was only one error message: "Your account has been deleted or deactivated (You do not have an account because it has been deleted or deactivated)." The hardest-hit were the Philippines and Türkiye regions, low-cost Team accounts, and accounts using

Reported by Synced In multi-stage reasoning, local misjudgments can become premises for subsequent judgments and continue to influence later steps. This problem is particularly prominent in empathetic speech dialogue: perception, psychological-state inference, and response generation are closely interconnected, and early biases may ultimately lead to misunderstandings of the speaker’s needs and inappropriate responses. Even if the subsequent analysis is lo

Synced Report Unprecedented! Witten, the “Pope of Physics” and the only physicist to receive the Fields Medal—the highest award in mathematics—publicly acknowledged AI. In the acknowledgments of a recent preprint paper, Witten wrote: Many details were assisted by Claude Opus 5.5. Many details were assisted by Claude Opus 5.5. Even more astonishingly, in the main text of the paper, Witten admitted that Claude Opus 5.5 proposed a key solution. Who is Witten?

Synced Report On August 17, programmers worldwide were forced to endure a collective outage. After writing the code and clicking Push, the progress bar spun in place; automated tests froze, pull requests couldn't be merged, and even the AI coding assistant Copilot went offline. The system remained stuck for more than seven hours, and that day, developers around the world waited for it to catch its breath. Three days later, GitHub Chief Technology Officer V

Xinzhiyuan reports Users pay tens—or even $200—a month to subscribe to Claude Code and hand their local terminal over to do the work, only for it to secretly insert your real login email address, name, and even the address of your company's internal code repository into the User-Agent request header in the background and send it to external websites! To cut corners, an Anthropic engineer secretly hard-coded the login email address into the system prompt fo

New Zhiyuan reports The professor came back from lunch and discovered that a year’s worth of his lecture notes was gone. University of Chicago Law School professor William Baude asked Claude to update the page numbers in a set of old lecture notes. After giving the instructions, he went off to have lunch. Unbelievably, in the time it took to eat a single meal, Claude accidentally deleted a year’s worth of his constitutional law lecture notes. The syllabus

Reported by Xinzhiyuan After AI agents began taking over terminals, security issues also entered the execution environment. When an agent can read files, write scripts, run commands, and continuously adjust its next action based on system feedback, a single incorrect decision may affect actual file and system state. Enabling agents to promptly identify potential risks while completing tasks has become an important issue in agent safety research. Long-horiz

Ripple-65B Preview delivers strong agentic coding results as a fully Apache 2.0 open US model Matthias Bastian View the LinkedIn Profile of Matthias Bastian Nov 6, 2026 AI chip design firm Ricursive Intelligence has trained an open language model. Ripple-65B Preview was pretrained on 12 trillion tokens and, according to the company, achieved a score of 80.52 on the TerminalBench 2.1 coding benchmark, the best result among open US models. Ripple partly uses

Anthropic's $1.5 trillion IPO hits resistance as investors balk at valuation Maximilian Schreiner View the LinkedIn Profile of Maximilian Schreiner Nov 6, 2026 Anthropic's IPO is in trouble as investors push back on the $1.5 trillion valuation and the tech market sells off. According to Reuters, potential backers doubt the planned roadshow can kick off next week as scheduled. As recently as September, the company was eyeing a valuation of up to $2 trillion

Google's AI detects high blood pressure from a regular smartwatch without a cuff Manuel Uth Nov 7, 2026 A Google Research AI can detect signs of high blood pressure using a regular smartwatch, with no blood pressure cuff needed for calibration. The AI relies on two sensors already built into modern watches—a pulse sensor and a motion sensor—according to a Google blog post. The study was published in Nature. A neural network learned from nearly 500,000 hour