🧠 Cloud & Infrastructure Watch
Anthropic launched “Dispatch”, enabling Claude to remotely control a user’s computer via phone-triggered tasks, signalling a major step toward persistent agentic workflows across devices.
Read MoreMeta is accelerating internal AI agent adoption, with Zuckerberg building a personal “CEO agent” and rolling out tools like “Second Brain” to query internal data across the company.
Read MoreCloudflare released Dynamic Workers, allowing AI agents to execute generated code in sandboxed environments at significantly faster speeds for production use.
Read MoreAgile Robots partnered with Google DeepMind to integrate Gemini models into industrial robotics, creating a feedback loop between real-world deployments and model training.
Read MoreMicrosoft and NVIDIA unveiled an AI platform to accelerate nuclear energy projects using digital twins and AI-driven optimisation on Azure infrastructure.
Read MoreOpenAI is reportedly reallocating compute from Sora to its upcoming “Spud” model, highlighting ongoing infrastructure prioritisation toward next-gen foundation models.
Reddit will require suspicious or “fishy” accounts to verify they are human, while still allowing communities to set their own rules for AI-generated content.
Read MoreGoogle’s TurboQuant memory-compression method cuts LLM memory use by roughly 6x without retraining and can speed processing on H100s by up to 8x.
Read MoreGoogle rolled out Lyria 3 Pro more broadly, extending longer-form AI music generation across more Google products.
Read MoreSierra launched Ghostwriter, an AI agent capable of building other AI agents across voice and chat, accelerating enterprise agent deployment.
Read MoreOpenAI is shutting down Sora, reallocating compute and talent toward its upcoming “Spud” model, signalling a strategic pivot away from video generation to prioritise core foundation model advancement and broader economic impact.
Read More
💻 Hardware & Semiconductor Moves
Alibaba unveiled a new AI chip optimised for agent-based workloads, focusing on lower latency and improved efficiency for multi-step autonomous systems.
Read MoreArm announced its first in-house “AGI CPU”, positioning itself as foundational silicon for the emerging agentic AI cloud era.
Read MoreHark, a new AI hardware startup by Figure AI’s Brett Adcock, committed to deploying thousands of NVIDIA B200 GPUs ahead of launching personalised AI devices.
Read More
