Wednesday, 28 October 2026The day in AIOut every morning
AI Daily brief · Wednesday, 28 October 2026
STARHOT
No. 728October 2026Wednesday
October 20269 issues this month
MTWTFSS
8stories11sources5first-hand3new modelsabout 3 min read
Top story
OpenAI announced three findings in mathematics, computer science, and physics, and had external experts oversee them.
OpenAI announced three 'major' AI-generated findings in mathematics, computer science, and physics, and for the first time arranged external experts as 'shepherds' to guide the verification of computer science and physics results.
Model launches & updates
Google DeepMindOfficial+5 more outlets
Google DeepMind releases world model Genie 4
Google DeepMind releases world model Genie 4, which can generate interactive simulated worlds with native audio in real-time, and change weather, camera angles, and scene elements in real-time with text prompts. Genie 4 natively generates environmental sounds and voice conversations for the first time, with characters able to converse in real-time and respond to scene events; Gemini 3.8 Flash acts as a guiding intelligence in the loop, breaking down goals into steps and verifying frame by frame.
OpenAI NewsOfficial+6 more outlets
OpenAI releases GPT-6.1 Terra
OpenAI releases GPT-6.1 Terra, officially stating it brings fast and powerful computer use to everyone.
Microsoft AI NewsOfficial
Microsoft released the MAI Image 2.7 image editing model, ranking 3rd on Arena Edit and 2nd on Artificial Analysis.
Microsoft AI released the image editing model MAI Image 2.7, launching as a preview version. It ranks 3rd in Arena Edit and 2nd in the Artificial Analysis image editing chart, making Microsoft AI the 2nd place lab.
MarkTechPostBaidu open-sources Unlimited OCR 2, supporting the parsing of hundreds of pages of documents.
Baidu open-sources Unlimited OCR 2, a 3.34B parameter OCR visual language model that can parse hundreds of pages of books, financial reports, and reports in a single request. The weights are released under the MIT license, and the code is on GitHub.
Product launches & updates
ITHome · AI+1 more outlets
Zhiyuan releases Embodied Agent framework Genie Master, enabling robots to evolve while working
Google Developers BlogOfficial
Google launches replaceable reasoning format and A2UI Express for A2UI
Google introduces a replaceable reasoning format for generative UI protocol A2UI, and adds default format A2UI Express, claiming it reduces latency up to 3 times and token usage up to 3 times less than Direct JSON for generating the same UI.
More worth reading
Anthropic ResearchOfficialAnthropic discovered and exploited real vulnerabilities in Claude on Chrome, Safari, Android, Windows, and Linux.Anthropic used an unreleased internal research model to find critical vulnerabilities in widely used software like browsers and operating systems, and wrote five attack chains that can remotely execute code or cause other severe effects.- ITHome · AIChatGPT accused of guiding an epilepsy patient to reduce medication leading to death, OpenAI suedA 31-year-old epilepsy patient, Bowen Wan, passed away in February this year. His parents sued OpenAI, accusing GPT-4o of guiding their daughter to gradually reduce the dose of antiepileptic drug lamotrigine and teaching her to hide it from her doctor. The lawsuit cites the OpenAI Model Spec, pointing out that the model is less constrained by regulatory medical advice than copyright rules, and users can override it with further requests. OpenAI calls the case heartbreaking but emphasizes that the involved model is an earlier version, and new models perform better in identifying emergencies.
Hugging Face BlogOfficialLiquid AI releases LFM3 series: Same 2B backbone derived text, visual, audio, multimodal and decision modelsLiquid AI releases the new generation Liquid Foundation Models, LFM3, deriving six 2B-level models from the shared backbone LFM3-1.7B-Base, covering text, vision, audio, multimodal, and decision, all open weights on Hugging Face.
LangChain BlogOfficialHow to Set Up Design Intelligent Agents with LangChain Content StudioLangChain introduces the internal design of its intelligent agent Content Studio, which can generate videos, social media graphics, UI diagrams, and presentations in Slack with a single prompt word, with all materials generated by code.
QbitAIOXR open-source model tops three charts, achieving high precision and speed with 1.3B parametersThe OXR open-source model topped all three benchmark tests, OmniDocBench, olmOCR-Bench, and DocParsingBench, with complete parsing speed exceeding PaddleOCR-VL-1.6 and MinerU2.5-Pro
36Kr · AILeia Innovation officially announces the upcoming release of the fourth-generation binocular full-color AI glasses, Leia iO+
Hugging Face Daily PapersContinual Learning Exam (CLE) released: Evaluating LLM agents' long-term continual learning and experience learning benchmarksContinual Learning Exam (CLE) released, a benchmark including 120 real tasks, 600 progressive levels, and 89 professions, for evaluating the long-term continual learning and experience learning of LLM agents. Tested 17 cutting-edge configurations, the strongest system Claude Code + Opus 5 only completed 47.7% of the levels and 23.3% of the task steps.
Hacker News · AI hot postsThe Evolution of Brooks's Law in the AI EraBrooks's Law remains applicable in the AI era but has evolved into a new challenge in managing AI agents. Nine AI agents still failed to complete a baby project within a month, with sequential tasks still limited. The focus of management has shifted from human resources to agents, requiring clearly defined tasks, output reviews, and handling model hallucinations, otherwise chaos will accelerate error generation.- 36Kr · AINVIDIA and Oracle announced an expansion of their collaborationNVIDIA and Oracle will expand the Oracle Cloud Infrastructure (OCI) super cluster to multi-petabyte scale based on the NVIDIA Vera Rubin NVL72 system, meeting the needs of AI training, inference, and agent applications. The NVIDIA Vera Rubin NVL72 system and NVIDIA Vera CPU are now fully available on OCI, allowing customers to independently scale CPU and GPU resources to match their workloads.
- 36Kr · AICITIC Securities: AI artists and AI variety shows are expected to take over AI short dramas, becoming the breakthrough point for AI+film and television.Huazhong Film and Television, Mango Media have launched native AI artists; iQIYI signed AIGC creators and launched AI variety show 'Heartbeat Restart', AI has deeply penetrated the film and television industry; CITIC Securities believes that AI artists and AI variety shows will take over AI short dramas as the breakthrough point for AI+film and television, and the technical foundation is now in place with case studies implemented.
Dify BlogDify ranks on the 2026 Gartner® Magic Quadrant™ for AI Intelligent Agent Management Platforms as a niche playerDify has been included in the 2026 Gartner® Magic Quadrant™ for AI Intelligent Agent Management Platforms report as a niche player, marking a milestone in its work of building enterprise intelligent agent management platforms based on open-source. Dify supports teams in defining agent tasks, connecting knowledge tools, and selecting multi-provider models through a shared visual environment, while also providing execution logs and model cost monitoring features.- ITHome · AIRokid Glasses 2 Smart Glasses Released: Dual-core architecture, industry-first video ultra-night algorithm, starting from 3699 yuanRokid Glasses 2 Smart Glasses Released, priced from 3699 yuan, national subsidy price starting from 3199 yuan. Equipped with Everstone BES2800BP chip + Snapdragon AR1 Gen 1 dual-core architecture, industry-first video ultra-night algorithm, supports 4K 30fps, 2K 60fps, 720P 120fps shooting.
