Confirmed that the MSK Express Kafka transaction bottleneck lies in the coordinator and presented an optimization sequence.
The company is advancing the construction of 50MW-class AI data centers in Ansan and Cheongna.
Kanzhun released the 33B dense model Nanbeige4.3-33B under the Apache 2.0 license.
The Kakao Recommendation Team applied Data Catalog and Analyst Persona to unify AI analysis standards.
Baidu has released Unlimited-OCR 2, which uses cross-page context to parse long documents.
Following the 2025 National Information Society Agency fire, public-sector DR standards have been strengthened, requiring recovery designs tailored to A1–A4 tiers.
The working draft defines 140 rules for handling tables, forms, and diagrams to resolve inconsistencies between different parsers.
A vulnerability chain involving MathLive and YouTube JSONP allowed attackers to execute arbitrary JavaScript in victims' authenticated sessions via public shares, enabling private data exfiltration.
The feature uses signals like label changes and PR merges to propose prompt refinements, which users must review and approve before application.
The new system routes and orchestrates models automatically, achieving a score of 60 on the Artificial Analysis Coding Agent Index while using about half the tokens and 40% less time per task.
Google's Gemini 4 Carbon scores 59 on the Intelligence Index v5, leading in complex reasoning and scientific tasks while ranking second only to Claude Opus 5.5 on coding agents.
The expansion includes the US National Science Foundation, EMBL, and the Francis Crick Institute, which will distribute free Claude credits to researchers.
The plugin allows users to start agents in the cloud from a laptop and control them via phone, saving files and history when the sandbox stops.
The feature allows developers to update models and configurations via a single ID, supporting 125 new models added in the last 90 days.
The hierarchical network architecture demonstrates 1.3x compute efficiency on validation loss and up to 5.37x gains on out-of-distribution code compared to standard BPE baselines.
Post-trained on DeepSeek V4.1 Flash, the model targets software engineering tasks with shorter reasoning and fewer tokens.
The limit increases from 500 MB to accommodate large toolchains and dependency sets, while the total image size cap stays at 15 GB.
The update removes the need for an interactive terminal, allowing agents to run enablement commands from a linked project directory.
The new system uses Turborepo to identify and run only the specific tasks affected by a code change, leveraging Fluid compute for faster feedback.
The feature is in private preview, requiring users to request access from Box and unmount before stopping the sandbox to ensure data persistence.
The acquisition allows Mithril to continue developing bispecific antibodies for LAMA2-CMD and other ECM-related genetic diseases with expanded resources and infrastructure.
AI Briefing