https://sguys99.github.io/global-ai-news/en/ Every morning from Seoul: the day's top AI news, open-source repos, models and papers — with a Korea Desk on Korean AI you won't find elsewhere. en Fri, 06 Nov 2026 21:08:10 GMT https://sguys99.github.io/global-ai-news/en/article/1405/https://sguys99.github.io/global-ai-news/en/article/1405/Fri, 06 Nov 2026 21:08:10 GMTThe authors introduce VFlash, a method for fast visual planning using Flash Primitive Decoding. It exploits the known structure of visual primitives to predict their coordinates jointly, achieving approximately 25x faster generation for 3D point motion forecasting. A world action model conditioned on VFlash plans achieves 70.2% average success across MolmoSpaces task suites, compared with 62.6% for the baseline.Physical AI https://sguys99.github.io/global-ai-news/en/article/1406/https://sguys99.github.io/global-ai-news/en/article/1406/Fri, 06 Nov 2026 21:08:10 GMTPyRUA-REAL is a framework for embodied recursive self-improvement through real-world exploration and code refinement. Coding agents retain interaction experience as persistent memory and reusable executable skills. Across eight real-world tasks, PyRUA-REAL achieves an average success rate of 91.1%, and 84.4% on the RoboDojo benchmark.Physical AI https://sguys99.github.io/global-ai-news/en/article/1400/https://sguys99.github.io/global-ai-news/en/article/1400/Fri, 06 Nov 2026 21:08:10 GMTThinking Machines, founded by former OpenAI chief technology officer Mira Murati, released Inkling 2 Small, a mixture-of-experts model with 594 billion parameters. The company claims it is the strongest western open-weight model, though better agentic performance comes with sharply rising token usage.LLM https://sguys99.github.io/global-ai-news/en/article/1401/https://sguys99.github.io/global-ai-news/en/article/1401/Fri, 06 Nov 2026 21:08:10 GMTServe Robotics announced that it will launch a new business line called Physical AI Services to offer customers connectivity and data tools, alongside its third-quarter earnings announcement.Physical AI https://sguys99.github.io/global-ai-news/en/article/1402/https://sguys99.github.io/global-ai-news/en/article/1402/Fri, 06 Nov 2026 21:08:10 GMTOpenAI introduced BioBasicsBench, a 100-task benchmark measuring how well AI agents interpret raw experimental data and handle everyday wet-lab science.Research https://sguys99.github.io/global-ai-news/en/article/1403/https://sguys99.github.io/global-ai-news/en/article/1403/Fri, 06 Nov 2026 21:08:10 GMTNonprofit METR audited Anthropic's internal research data for eleven weeks and concluded that the pace of model development has roughly doubled, with about 80 percent of the speedup driven by AI-assisted research. Anthropic redacted some evidence, and skeptical voices exist inside the company.Research https://sguys99.github.io/global-ai-news/en/article/1404/https://sguys99.github.io/global-ai-news/en/article/1404/Fri, 06 Nov 2026 21:08:10 GMTPhysicl raised $10 million to solve the sim-to-real gap using a pipeline originally built to market goods online.Physical AI https://sguys99.github.io/global-ai-news/en/article/1407/https://sguys99.github.io/global-ai-news/en/article/1407/Fri, 06 Nov 2026 21:08:10 GMTPerplexity launched Teammates, always-on agents that work through ongoing projects on their own virtual machines. The feature requires the Max plan at $200 per month.Agents https://sguys99.github.io/global-ai-news/en/article/1408/https://sguys99.github.io/global-ai-news/en/article/1408/Fri, 06 Nov 2026 21:08:10 GMTAmazon added a text messaging feature to Alexa for Shopping that lets customers purchase items by describing products or providing photos. Alexa searches Amazon listings, sends back recommendations via text, and orders items on users' behalf.Agents https://sguys99.github.io/global-ai-news/en/article/1409/https://sguys99.github.io/global-ai-news/en/article/1409/Fri, 06 Nov 2026 21:08:10 GMTAirbnb said in its Q3 2026 earnings letter that it will roll out an AI agent in the first half of 2027 to help users plan and book travel. The company plans to leverage its experience matching hosts and guests and a new interface built specifically for travel.Agents https://sguys99.github.io/global-ai-news/en/article/1410/https://sguys99.github.io/global-ai-news/en/article/1410/Fri, 06 Nov 2026 21:08:10 GMTAGIBOT unveiled GO-3 Preview, an embodied foundation model that integrates pre-training, post-training, and post-deployment learning. The model secured first-place rankings on three major embodied AI benchmarks and helps robots develop general-purpose capabilities.Physical AI https://sguys99.github.io/global-ai-news/en/article/1411/https://sguys99.github.io/global-ai-news/en/article/1411/Fri, 06 Nov 2026 21:08:10 GMTPalo Alto Networks is in talks to acquire Clutch, an AI agent security startup backed by Lightspeed and SignalFire, for $500 million. Clutch was last valued at $85 million.Industry & Policy https://sguys99.github.io/global-ai-news/en/article/1412/https://sguys99.github.io/global-ai-news/en/article/1412/Fri, 06 Nov 2026 21:08:10 GMTSince Anthropic's Claude Mythos Preview in April demonstrated finding software vulnerabilities and generating attack code, OpenAI, Anthropic, and Google have released frontier models with enhanced vulnerability discovery and attack validation capabilities. Recent developments show AI now writing patches to fix identified vulnerabilities.Research https://sguys99.github.io/global-ai-news/en/article/1413/https://sguys99.github.io/global-ai-news/en/article/1413/Fri, 06 Nov 2026 21:08:10 GMTAn independent evaluation report from METR found that Anthropic has roughly doubled its model performance improvement speed using AI. Since October last year, model enhancement speed accelerated by about 1.9x, achieving in one year what previously took two years. Three undisclosed models trained recently were also identified.Research https://sguys99.github.io/global-ai-news/en/article/1414/https://sguys99.github.io/global-ai-news/en/article/1414/Fri, 06 Nov 2026 21:08:10 GMTZyphra released ZYFA, open-source FP8 and BF16 prefill attention kernels for AMD Instinct MI355X. On real model activations, the FP8 kernels run 1.16–1.76x faster than AMD's AITER with 62–78% lower error at head dimension 128.Open Source & Tools https://sguys99.github.io/global-ai-news/en/article/1415/https://sguys99.github.io/global-ai-news/en/article/1415/Fri, 06 Nov 2026 21:08:10 GMTroblox-studio-mcp is an open-source, lightweight Model Context Protocol server for Roblox Studio. It lets AI assistants like Claude and Cursor connect to read object trees, edit scripts, and run playtests safely.Agents https://sguys99.github.io/global-ai-news/en/article/1416/https://sguys99.github.io/global-ai-news/en/article/1416/Fri, 06 Nov 2026 21:08:10 GMTWebsiteBench is a benchmark that evaluates whether AI agents can rebuild websites through browser exploration. It measures web agents' capabilities in browser automation and visual evaluation.Research https://sguys99.github.io/global-ai-news/en/article/1417/https://sguys99.github.io/global-ai-news/en/article/1417/Fri, 06 Nov 2026 21:08:10 GMTResearchers auditing Terminal-Bench 2.0, a coding agent benchmark, identified defects in approximately 30% of the benchmark and released Terminal-Bench 2.1 with repairs. The study shows that as models improve, existing verifiers can reject valid solutions, and changes to evaluation infrastructure can invalidate benchmark assumptions. It also demonstrates that resource allocation, such as time limits, materially affects agent performance and must be part of the measurement specification.Research https://sguys99.github.io/global-ai-news/en/article/1418/https://sguys99.github.io/global-ai-news/en/article/1418/Fri, 06 Nov 2026 21:08:10 GMTFollowing midterm elections where voters opposed data center expansion, OpenAI announced it will not use non-disclosure agreements (NDAs) with public officials on data center projects. The company said it will allow only narrow protections covering legitimate security information, trade secrets, and competitively sensitive information.Industry & Policy https://sguys99.github.io/global-ai-news/en/article/1419/https://sguys99.github.io/global-ai-news/en/article/1419/Fri, 06 Nov 2026 21:08:10 GMTNVIDIA's open source tools for 6G development, including the Aerial wireless software and Sionna, provide researchers with a foundation for testing ideas and comparing results. The tools connect simulation and AI training with experiments on real over-the-air wireless systems.Open Source & Tools https://sguys99.github.io/global-ai-news/en/article/1378/https://sguys99.github.io/global-ai-news/en/article/1378/Thu, 05 Nov 2026 21:15:59 GMTOpenAI introduced GPT-6 Cyber, a cybersecurity model designed to find exploitable vulnerabilities and develop fixes. The model is trained to stay within authorized scope.Agents https://sguys99.github.io/global-ai-news/en/article/1379/https://sguys99.github.io/global-ai-news/en/article/1379/Thu, 05 Nov 2026 21:15:59 GMTGemini Omni 1.2 Flash brings native 15-second video generation and editing, stronger consistency, and 30% lower pricing.LLM https://sguys99.github.io/global-ai-news/en/article/1384/https://sguys99.github.io/global-ai-news/en/article/1384/Thu, 05 Nov 2026 21:15:59 GMTScenario Skills lets any AI agent access production-ready images, video, audio, and 3D content. Its skills pick the right model, show the price before spending and keep characters and brands consistent through the Scenario MCP, and it adds expert teams that drive Blender, Maya, ZBrush, Unreal and Unity.Agents https://sguys99.github.io/global-ai-news/en/article/1385/https://sguys99.github.io/global-ai-news/en/article/1385/Thu, 05 Nov 2026 21:15:59 GMTThe authors propose NEO-VLA, an encoder-free vision-language-action framework where images are tokenized and evolve jointly with language throughout the backbone. They show that this approach outperforms encoder-based counterparts across diverse manipulation tasks and distribution shifts.Physical AI https://sguys99.github.io/global-ai-news/en/article/1380/https://sguys99.github.io/global-ai-news/en/article/1380/Thu, 05 Nov 2026 21:15:59 GMTBaidu released ERNIE 5.5 Flash Base, a 280-billion-parameter mixture-of-experts model pre-trained on 33.5 trillion tokens. The company shipped it without fine-tuning to allow developers full flexibility.LLM https://sguys99.github.io/global-ai-news/en/article/1381/https://sguys99.github.io/global-ai-news/en/article/1381/Thu, 05 Nov 2026 21:15:59 GMTTencent unveiled Hy4, its first multimodal language model that understands images and can operate software. On the Artificial Analysis Intelligence Index, it scores 35 points but still trails the best commercial models on hard math and complex image analysis, according to Tencent.LLM https://sguys99.github.io/global-ai-news/en/article/1382/https://sguys99.github.io/global-ai-news/en/article/1382/Thu, 05 Nov 2026 21:15:59 GMTMeta is acquiring Conductor, an AI coding startup that lets developers control multiple AI coding agents at once. OpenAI announced plans to acquire Pierre Computer the same day, and Anthropic bought Stainless in May. The Decoder sees a pattern: as AI models become interchangeable, value is shifting to the tools built around them.Agents https://sguys99.github.io/global-ai-news/en/article/1383/https://sguys99.github.io/global-ai-news/en/article/1383/Thu, 05 Nov 2026 21:15:59 GMTChinese authorities blocked SoftBank's $5.4 billion acquisition of ABB Robotics over concerns about robot data being transferred overseas and used to train AI models.Industry & Policy https://sguys99.github.io/global-ai-news/en/article/1391/https://sguys99.github.io/global-ai-news/en/article/1391/Thu, 05 Nov 2026 21:15:59 GMTpoki is a local MCP server and agent that lets AI operate a real iPhone without a jailbreak or extra hardware.Agents https://sguys99.github.io/global-ai-news/en/article/1390/https://sguys99.github.io/global-ai-news/en/article/1390/Thu, 05 Nov 2026 21:15:59 GMTThe authors introduce FrontierPhysics, an end-to-end benchmark with 56 tasks created by 42 physicists based on their own research. Each task requires agents to review literature, develop a research plan, conduct research and experiments, and produce manuscripts. Across 22 models in 26 configurations, the best passed only 46 of its 168 rollouts.Research https://sguys99.github.io/global-ai-news/en/article/1386/https://sguys99.github.io/global-ai-news/en/article/1386/Thu, 05 Nov 2026 21:15:59 GMTAnthropic launched a Personal Finance feature that connects Claude to users' bank and card accounts via the Plaid bank bridging platform with read-only access. Users can ask Claude questions about spending by category, upcoming bills, affordability of major purchases, and estimated monthly savings.Agents https://sguys99.github.io/global-ai-news/en/article/1387/https://sguys99.github.io/global-ai-news/en/article/1387/Thu, 05 Nov 2026 21:15:59 GMTAxiom Math is focusing on formal verification to ensure the quality of AI-generated proofs in mathematics. The startup's AI agents formally verified the Stacks Project, one of the field's largest libraries.Agents https://sguys99.github.io/global-ai-news/en/article/1388/https://sguys99.github.io/global-ai-news/en/article/1388/Thu, 05 Nov 2026 21:15:59 GMTTogether AI released Together Compiler, an agent-driven loop that continuously re-optimizes the inference forward pass as traffic shifts. It achieved up to 2.6x TPS improvement over InferenceX's best configurations in under 24 hours per model.Agents https://sguys99.github.io/global-ai-news/en/article/1389/https://sguys99.github.io/global-ai-news/en/article/1389/Thu, 05 Nov 2026 21:15:59 GMTNVIDIA Vera is a CPU designed to handle the needs of AI agents that move between GPU reasoning and CPU execution across multiple steps including information search, database queries, code writing and testing, and action execution.Agents https://sguys99.github.io/global-ai-news/en/article/1392/https://sguys99.github.io/global-ai-news/en/article/1392/Thu, 05 Nov 2026 21:15:59 GMTSouth Korean medical device company Osteonic acquired a 51% stake in FoxEyes, a US spine surgery robot developer, for KRW 3 billion. FoxEyes was founded in 2024 and is preparing to commercialize spine surgery robots. Osteonic plans to expand from orthopedic implants into surgical robotics.Physical AI https://sguys99.github.io/global-ai-news/en/article/1395/https://sguys99.github.io/global-ai-news/en/article/1395/Thu, 05 Nov 2026 21:15:59 GMTVibeVoice is Microsoft's open-source frontier voice AI project.LLM https://sguys99.github.io/global-ai-news/en/article/1393/https://sguys99.github.io/global-ai-news/en/article/1393/Thu, 05 Nov 2026 21:15:59 GMTMeta's AI agent can now take calls as a customizable animated avatar. This comes ahead of Meta's planned launch of a Tamagotchi-like wearable designed to bring Muse to life.Agents https://sguys99.github.io/global-ai-news/en/article/1394/https://sguys99.github.io/global-ai-news/en/article/1394/Thu, 05 Nov 2026 21:15:59 GMTOpenAI introduced its National Community Compact and Compute Campus Policy Blueprint to protect local resources, create jobs, and guide responsible AI infrastructure development.Industry & Policy https://sguys99.github.io/global-ai-news/en/article/1397/https://sguys99.github.io/global-ai-news/en/article/1397/Thu, 05 Nov 2026 21:15:59 GMTSemiAnalysis unveiled GDPx, a new benchmark for simulated real-world knowledge work spanning chip design to quant trading. No model scored above 12 percent; Claude Opus 5.5 narrowly led Gemini 4 Carbon. Critics point to the small number of tasks per domain and uneven testing conditions.Research https://sguys99.github.io/global-ai-news/en/article/1398/https://sguys99.github.io/global-ai-news/en/article/1398/Thu, 05 Nov 2026 21:15:59 GMTRicursive Intelligence released Ripple-65B Preview, a 65B dense open-weight LLM under Apache 2.0 license. It scores 80.5 on Terminal-Bench 2.1, ahead of US open models up to 15 times its size.LLM https://sguys99.github.io/global-ai-news/en/article/1399/https://sguys99.github.io/global-ai-news/en/article/1399/Thu, 05 Nov 2026 21:15:59 GMTDropbox acquired Tigris Data for $350 million in a deal expected to support development of DBX Agent Cloud, a storage platform for developers building AI models and agents.Industry & Policy https://sguys99.github.io/global-ai-news/en/article/1356/https://sguys99.github.io/global-ai-news/en/article/1356/Wed, 04 Nov 2026 21:08:15 GMTOpenAI is acquiring Pierre Computer Company to expand tools for Codex, its coding agent.Agents https://sguys99.github.io/global-ai-news/en/article/1357/https://sguys99.github.io/global-ai-news/en/article/1357/Wed, 04 Nov 2026 21:08:15 GMTLiberAI introduced Liber-0, a world action model pretrained on more than 2.2 million hours of human experience to support zero-shot robot manipulation.Physical AI https://sguys99.github.io/global-ai-news/en/article/1358/https://sguys99.github.io/global-ai-news/en/article/1358/Wed, 04 Nov 2026 21:08:15 GMTAnthropic released FlashProve, a research prototype that uses cryptography to verify whether an AI response actually came from a claimed model without exposing the model's weights. The system still has significant limitations.Research https://sguys99.github.io/global-ai-news/en/article/1359/https://sguys99.github.io/global-ai-news/en/article/1359/Wed, 04 Nov 2026 21:08:15 GMTToyota and Waymo announced plans to begin testing and validating an autonomous Toyota bZ4X fleet in Japan in 2027, with deployment to the UK and other Waymo-serviced markets planned for the following years. Toyota will assemble the vehicles at its Takaoka plant, integrating Waymo's AV hardware.Physical AI https://sguys99.github.io/global-ai-news/en/article/1360/https://sguys99.github.io/global-ai-news/en/article/1360/Wed, 04 Nov 2026 21:08:15 GMTNVIDIA's Auto Ontology is a Python framework for building semantic, graph-based reasoning layers over data. It enables deterministic query generation, better context retrieval, and more explainable AI for LLM and agent workflows.Open Source & Tools https://sguys99.github.io/global-ai-news/en/article/1361/https://sguys99.github.io/global-ai-news/en/article/1361/Wed, 04 Nov 2026 21:08:15 GMTThe authors present Route Anything, a unified framework for studying model and policy routing in general-purpose AI systems. Across 25 benchmarks spanning language, image, video and speech generation, coding agents and robot manipulation, per-input oracle routing beats the best fixed candidate by 5.4 points. Learned routers beat it by 0.3 points across 16 benchmarks while using 24% less policy time.Research https://sguys99.github.io/global-ai-news/en/article/1362/https://sguys99.github.io/global-ai-news/en/article/1362/Wed, 04 Nov 2026 21:08:15 GMTThe authors introduce DreamX, which repurposes a pretrained video diffusion model as a universal latent action learner. The approach compresses latent video frames by 300 times and captures action-relevant dynamics disentangled from visual appearance, supporting zero-shot action transfer across embodiments and environments. It achieves state-of-the-art performance among latent action models on LARY-Bench.Physical AI https://sguys99.github.io/global-ai-news/en/article/1363/https://sguys99.github.io/global-ai-news/en/article/1363/Wed, 04 Nov 2026 21:08:15 GMTThe authors introduce H-JEPA, which trains a hierarchy of action-conditioned JEPAs in which each level predicts farther ahead in its own learned latent space. On Visual AntMaze, a three-level hierarchy raises planning success from 18 percent to 73 percent using less planner compute, and the approach extends to real-robot videos from DROID.Research https://sguys99.github.io/global-ai-news/en/article/1364/https://sguys99.github.io/global-ai-news/en/article/1364/Wed, 04 Nov 2026 21:08:15 GMTCloudOpsBench is an open-source benchmark that measures how well AI agents handle typical DevOps tasks in the cloud, from fixing services and removing resources to restricting access rights, rolling out updates, and decommissioning projects. The benchmark reveals reasoning as the biggest bottleneck.Agents https://sguys99.github.io/global-ai-news/en/article/1365/https://sguys99.github.io/global-ai-news/en/article/1365/Wed, 04 Nov 2026 21:08:15 GMTAnthropic warned that AI development accelerated by recursive self-improvement could outpace safety and security measures. In an amended IPO filing (S-1/A) with the SEC, the company acknowledged it had been slow in the past to notice sharp jumps in development speed, and called for stronger human oversight and control.Industry & Policy https://sguys99.github.io/global-ai-news/en/article/1366/https://sguys99.github.io/global-ai-news/en/article/1366/Wed, 04 Nov 2026 21:08:15 GMTZyphra unveiled Zyphra Hierarchical Attention (ZHA), a sparse attention technique that speeds up long-context processing. The method organizes information hierarchically and retrieves only the parts it needs instead of searching the whole document; the company projects up to 35x faster processing than standard attention on 10-million-token contexts.Research https://sguys99.github.io/global-ai-news/en/article/1371/https://sguys99.github.io/global-ai-news/en/article/1371/Wed, 04 Nov 2026 21:08:15 GMTspec-kit is a toolkit to help teams get started with specification-driven development or other specification-based processes.Open Source & Tools https://sguys99.github.io/global-ai-news/en/article/1367/https://sguys99.github.io/global-ai-news/en/article/1367/Wed, 04 Nov 2026 21:08:15 GMTThe report summarizes lessons on multimodal understanding from successive generations of the Kling video model. It describes Kling-MLLM, a shared multimodal foundation model that unifies image, video, audio, and interleaved inputs, and examines how generation quality emerges from the data–conditioning–alignment–evaluation chain.Research https://sguys99.github.io/global-ai-news/en/article/1369/https://sguys99.github.io/global-ai-news/en/article/1369/Wed, 04 Nov 2026 21:08:15 GMTagenterm is a self-hosted remote assistance tool that lets AI agents control remote computers through MCP and a terminal.Agents https://sguys99.github.io/global-ai-news/en/article/1375/https://sguys99.github.io/global-ai-news/en/article/1375/Wed, 04 Nov 2026 21:08:15 GMTThe authors propose FactEmbed for composed image retrieval, which finds a target image from a reference image and a modification text. Instead of mapping the query straight to one embedding, it uses factorized latents that simulate the intended visual changes before building the retrieval embedding. The authors report that it outperforms prior work across CIR benchmarks.Research https://sguys99.github.io/global-ai-news/en/article/1370/https://sguys99.github.io/global-ai-news/en/article/1370/Wed, 04 Nov 2026 21:08:15 GMTSpeedy is an optimized fork of Strata that runs Qwen3.8-Flash-Next, a 125B mixture-of-experts model, up to 1-million-token context on one or two RTX cards, tuned on an RTX 3080 with 10 GB of memory.Open Source & Tools https://sguys99.github.io/global-ai-news/en/article/1372/https://sguys99.github.io/global-ai-news/en/article/1372/Wed, 04 Nov 2026 21:08:15 GMTIn a podcast episode, Hugging Face discusses how companies capture value as enterprises, governments and startups turn to open source models for lower costs, more control and more flexibility. It also weighs whether American players can keep up as Chinese open models gain ground.Industry & Policy https://sguys99.github.io/global-ai-news/en/article/1373/https://sguys99.github.io/global-ai-news/en/article/1373/Wed, 04 Nov 2026 21:08:15 GMTAnthropic is rolling out text watermarks to four more Claude models before the EU AI Act's December deadline. However, on older models like Claude Opus 4.6 and Sonnet 4.6, API users can disable the watermark by changing a single sampling parameter. Anthropic's response is a compliance warning.Industry & Policy https://sguys99.github.io/global-ai-news/en/article/1374/https://sguys99.github.io/global-ai-news/en/article/1374/Wed, 04 Nov 2026 21:08:15 GMTBeyond Robotics, a physical AI robotics company, unveiled its PICKY strawberry-harvesting robot and a humanoid robot at the KIEMSTA 2026 agricultural machinery expo in Daegu. The company showcased a humanoid designed to automate unstructured tasks across industries, built on 3D vision AI and robot control technology developed in agricultural settings.Physical AI