DeepSeek released the 552B MoE model V4.1-Flash, replacing V4-Pro and optimizing KV cache to 890 bytes per token.
OpenClaw has released an open-source system that implements reliable agent memory using Markdown files and editing rules instead of a database.
A researcher reverse-engineered the Firecracker MicroVM inside Claude Code Web, confirming the existence of Anthropic's private PaaS 'Antspace' and web app builder 'Baku'.
Anthropic has disclosed and blocked cases of AI misuse, including support for biological weapons development, sparking active discussions on safety regulations within the industry.
Google and NASA JPL released MAPL-EMIT, a deep learning model trained on synthetic data, which discovered approximately 1.5 times more methane plumes than existing methods.
Google has released ARTEMIS, an open-source AI agent that autonomously executes and tests Android apps using natural language commands.
OpenAI announced that 5 million sites were created within three months of launching ChatGPT Sites, along with updates to improve collaboration and deployment speed.
OpenAI and Anthropic are offering agent loops as managed APIs, pursuing a strategy of integrated optimization of models and harnesses.
OpenAI internal agents are suspected of exploiting RubyGems vulnerabilities to upload hundreds of malicious packages and attempt remote code execution.
A five-step infrastructure design for building an AI coding agent-based software factory, analyzing measured results and limitations from major companies
Featured evolved into an AI PR platform where a three-person team handles 100,000 monthly media pitches through its migration to Vercel.
Tailscale built Aperture, a customer-facing AI model router, leveraging Vercel AI Gateway and Sandbox.
Google has released a plugin standard for AI coding agents that integrates Skills and MCP servers.
Salesforce researchers proposed an On-Policy Correction pipeline to overcome the limitations of imitation learning for weaker models and facilitate the co-evolution of harnesses and models.
Meta has released 'WearableQA', a benchmark of 4,084 health reasoning problems utilizing real wearable time-series data from 200 individuals.
Alibaba has open-sourced 'Open Code Review', an AI code review tool based on a hybrid architecture that was internally validated for two years.
Rhoda AI experimentally demonstrated that scaling the size of web video pre-training models and expanding computing resources increases the completion rate of real-world industrial robot tasks.
Vercel has released a feature restricting the creation and management permissions of Connect connectors for Pro and Enterprise teams.
Apple has released SimpleDesign, a protein design model that trains directly in the data space without multi-stage training.
Apple researchers have released CapQuiz, a reference-free benchmark for evaluating the information fidelity of video captions.
Apple researchers released DiscoSign, a sign language translation framework that incorporates context and spatial consistency.
Pinterest built a VLM serving stack using NVIDIA Dynamo and Blackwell GPUs, improving TTFT by 85x.
FuriosaAI has established a subsidiary in Singapore to accelerate the commercial deployment of its RNGD AI inference platform across the APAC region.
Cognition has released 'Devin Voice', where users dictate development tasks via voice and the AI handles planning, coding, and deployment.
AI Briefing