AI Daily brief · Saturday, 31 October 2026
STARHOT
Factory releases Droid 1.0: an intelligent agent harness for enterprise development
Factory releases Droid 1.0, a major update to its intelligent agent harness, with programmatic tool calling enabled by default (Script mode). The model can execute hundreds of tool calls within a single request in a sandbox, executing JavaScript.
Factory NewsOriginal Model launches & updates
ITHome · AI+1 more outlets
Zhiyuan released the embodied model GO-3 Preview, topped three major lists, and plans to open source by the end of 2026.
Zhiyuan launched the embodied model GO-3 Preview on October 29, adopting an embodied-native WRAM architecture that connects task reasoning, future prediction, and action generation, and has topped the three major evaluation lists: RoboDojo, RoboTwin, and RoboColiseum. It plans to fully open source by the end of 2026.
ITHome · AI
Xiaomi open-sources 0.8B document understanding model Xiaomi-OCR-0, supporting Markdown and JSON output
Xiaomi open-sources lightweight document understanding model Xiaomi-OCR-0, with 0.8B parameter size, weights and inference code have been uploaded to Hugging Face and GitHub, a single model covers three tasks of document parsing, key information extraction, and OCR visual question answering.
QbitAI
The domestically developed open-source OXR tops three document parsing benchmarks with 1.3B parameters.
The SoMark team has open-sourced the 1.3B document parsing model OXR, ranking first on three benchmarks: OmniDocBench (97.18), olmOCR-Bench (87.84), and DocParsingBench (87.76).
OpenAI NewsOfficial+2 more outletsOpenAI updates model guidelines, guided by five core values to create a more powerful AI
OpenAI updates model guidelines, guided by five core values to create a more powerful AI; the new guidelines aim to ensure AI has sound judgment, respects human autonomy, and is loyal to human intentions.
Product launches & updates
LangChain BlogOfficial
LangChain releases Managed Deep Agents v0.10, adding browser usage and asynchronous sub-agents
LangChain releases Managed Deep Agents v0.10, adding browser usage capabilities to let agents operate websites without APIs, and supporting asynchronous sub-agents to run long tasks in the background.
Industry
QbitAI
FaceWall AI's ForgeKernel live stream generated an operator from scratch, 9.8% faster than DeepSeek FlashMLA and announced open-source.
FaceWall AI's operator generation system ForgeKernel generated the DeepSeek-V4.1 FlashMLA Prefill operator from scratch on an NVIDIA B200, with a final time of 472 microseconds, 9.8% faster than FlashMLA's 518 microseconds. The entire process took 22.5 hours, 42 iterations, and 32 versions, using 1 B200 and 22 GPU hours.
Research papers
Meta EngineeringOfficialMeta rolls out JiTTesting to production: Use LLM to generate tests and catch bugs before code merge
Meta announced that Catching JiTTests has been put into internal production, using LLM to automatically generate executable tests for diffs to be merged, exposing behavioral differences before the code is landed
More worth reading
The DecoderEpoch AI: Open-source weight model GLM 5.3 constructs a complete Firefox vulnerability exploitation chain within 48 hours with zero manual inputEpoch AI researchers discovered that the Chinese open-source weight model GLM 5.3 can exploit the window between the release of security patches and their deployment to attack Firefox, covering 6 versions released from September 17 to October 26. The API cost for each successful attack is approximately 1000 US dollars.- ITHome · AIAlibaba said that users of the China market version of Apple Intelligence can choose whether to enable the Qwen integration or not.Alibaba announced that the China market version of Apple Intelligence will integrate with Qwen, allowing users to call Qwen's text and image understanding and content generation capabilities in Siri, writing tools, and visual intelligence, and they can choose whether to enable it or not. Related features will be open for testing with the latest iOS developer Beta version on November 4, and will be officially launched for Chinese users with the iOS 27.3 version in December. Alibaba stated that user requests and generated replies will not be used for model training.
- The DecoderClaude model independently selects gene therapy and completes mouse brain surgeryThe research team allowed Claude Sonnet 5 to autonomously select gene therapy, brain regions, and injection coordinates in an epilepsy mouse model, and operated a custom surgical device to complete the injection, only providing the instruction 'reduce epilepsy seizures'.
Ars Technica · AIThe U.S. Department of Justice sued those using AI to impersonate students and fraudulently obtain student aid loans, targeting 'ghost students'.The U.S. Department of Justice announced charges against five individuals suspected of impersonating 'ghost students' to fraudulently obtain federal student aid loans, calling it the first large-scale federal crackdown on such fraud.- The DecoderReddit CEO Huffman warns Google AI Mode does not drive traffic for anyone.Reddit CEO Steve Huffman stated that Google's AI Mode does not bring traffic to any website, and Reddit lost over one million Google search referrals per week in October, with the stock price falling more than 25%. Huffman believes this is a fundamental shift in Google's work methods. Reddit is testing login walls outside the United States to reduce its reliance on the open network and continues to negotiate data authorization with AI providers.
MarkTechPostMicrosoft and UC Berkeley release TokenEconomizer plugin, coding agent costs reduced by up to 30.9%Microsoft Research and UC Berkeley's research team released the open-source cost controller TokenEconomizer, which can reduce the cost of encoding agents without changing model weights and agent code.
