Horus Hiero, an open-source hieroglyph translation and multimodal model based on Qwen 3.5, has been released.
The 'GitLost' vulnerability, which leaks private repository data through prompt injection in GitHub Agentic Workflows, has been discovered.
A hardware KVM based on MCP has been released, helping AI agents directly control physical devices.
Chinese companies are expanding their ecosystems by focusing on developing world models for robotics and physical AI.
A report has surfaced of an AI coding agent misusing a Prisma command, resulting in the deletion of a production database's schema.
It's a stealth Chromium browser designed to help AI agents evade website bot detection.
The combination of cheap sensors, robotics, and multimodal models is creating new AI business opportunities that leverage data from the physical world.
Shilong Liu proposed a framework that classifies agents by their mode of evolution into artifacts, infrastructure, and model weights.
Anthropic has unveiled GRAM, a technology that can selectively remove or activate dangerous knowledge in AI models without degrading performance.
SpaceXAI's Grok 4.5 model has been added to AI Gateway, supporting text and image inputs.
With the Vercel CLI, you can now directly modify framework presets and build settings without accessing the dashboard.
Vercel has introduced a deployment check feature that verifies missing microfrontends.json files to prevent routing errors during deployment.
GPT-Live, a next-generation voice model designed to support natural human-AI interaction, has been unveiled and applied to ChatGPT Voice.
vLLM's Transformers modeling backend has been optimized to deliver throughput on par with or faster than existing custom implementations.
A new AI model jointly developed by SpaceX and Cursor is expected to be unveiled as early as this Wednesday.
Current alignment benchmarks risk overestimating safety, since models may recognize the evaluation environment or conceal risky behavior.
Alongside the release of GPT-5-Codex-Mini, OpenAI announced Codex's Skills feature and Responses API updates.
This piece covers strategies for building agents that evolve on their own by combining agent execution traces with user activity data.
Meta has unveiled Muse Image, which combines with Muse Spark to generate high-quality visuals.
Gemma 4 is a next-generation open-weight multimodal model with diverse architectures and innovative features.
MiniMax M3 uses Sparse Attention technology to dramatically reduce cost and computation even with long contexts, enabling the implementation of long-horizon agents.
Chinese AI startup DeepSeek has begun developing inference chips for its own data centers in response to US export restrictions.
OpenAI is releasing three GPT-5.6 series models, including its flagship model Sol.
The Flags SDK has reduced microtask queue overhead, speeding up bulk evaluation of multiple feature flags by 10x.
AI Briefing