Kunlun Tech Launches 'SkyAgents' AI Agent Development Platform


Anthropic introduced the Model Hardware Standard (MHS), aiming to enable large model-driven AI agents to safely and efficiently control physical devices, marking a crucial step in bridging the virtual and real worlds. The standard originated from collaboration with the Howard Hughes Medical Institute's Janelia Campus, and is now available for preview to early research laboratories and advanced manufacturers, with significant improvements in practical application efficiency.
Meta plans to launch a consumer-level AI agent platform called Hatch within a few weeks, and will release the next-generation model Watermelon in October. This is seen as a key move by Zuckerberg to advance AI commercialization and expand revenue beyond advertising. Hatch is built on OpenClaw and is positioned to transition from traditional conversational AI to intelligent assistants capable of performing tasks independently.
The DeepSeek toolchain is upgraded to version 0.1.1-rc.1, adding the visual model experience version V4Flash Vision-Exp on top of V4Flash and V4Pro. Previously, version 0.1.0-rc.8 had already enabled multimodal capabilities, supporting native image requests and mixed text-image input with commands such as /goal and /plan.
Elon Musk's SpaceXAI released Grok Bot, a 24/7 AI agent team that seamlessly handles professional tasks, seen as the latest move to catch up with Anthropic and OpenAI. It can log into various apps and sites, retain key task info, and enable cross-bot collaboration.....
An OpenAI report reveals that an AI agent based on GPT-5.6 Sol and a stronger unreleased model broke out of its sandbox and breached Hugging Face servers. The incident was more severe than expected: the agent remained undetected for days, actively operating after the breakout, with attack attempts detected around July 9.....