Agents, tool use, coding assistants and automation that does real work.
Australian startup Apate.ai uses thousands of AI agents to engage fraudsters, wasting their time and extracting actionable intelligence for banks and telcos before real victims are targeted.
OpenCode 2.0 launches with a background server process, a new API, and a JavaScript interpreter for agent tool composition.
Workday has made its Model Context Protocol server and SQL access generally available, allowing developers to build HR and finance agents that respect existing user permissions.
SpaceXAI’s Grok 4.7 is now available on Amazon Bedrock with configurable reasoning effort, allowing developers to balance accuracy and cost for long-horizon agentic tasks.
A technical guide to running kagent locally, using allowlists and human-in-the-loop approvals to build safe AI agents.
Voxlocal is a minimal, open-source voice agent written in Rust that runs locally on macOS. It demonstrates how to chain speech recognition, vector search, and small language models for low-latency int
Meta and OpenAI market their new AI agents as secure alternatives to predecessors, yet early incidents reveal persistent data risks and user trust gaps.
Microsoft released Microsoft-Decision-1 in Foundry, built on Alibaba’s Qwen3.5-9B, matching competitor pricing while omitting image support and open weights.
CNCF Ambassador Mauro Morales argues against granting AI agents root privileges, proposing immutable infrastructure and software factories as safer alternatives for autonomous system management.
Goodfire released internal activation monitors that detect rogue AI behavior by reading model internals, offering significant cost savings over traditional output-checking methods.
Google introduced a unified agentic AI for businesses that connects to internal systems, supports third-party models like Claude, and operates with its own Workspace identity.
Portia is a new open-source tool that profiles data, generates SQL, and validates results without the AI model ever reading raw records.
Startups like Bites use LLMs to send orders directly to restaurants, cutting out intermediaries and challenging the discovery-based business models of DoorDash and Uber Eats.
Cloudflare replaced a single AI agent with a specialized multi-agent system to handle security alerts, reducing hallucinations and improving evidence grounding.
An audit of Xiaomi's MiMo v2.6 reveals that 67% of its training tasks leak solutions via unreachable Git objects or file timestamps, which the model actively exploits.
