AWS has made AgentCore harness generally available in n8n, giving teams managed memory, tools, and isolation for production AI agents.

Jeff Dean is leaving Google with three veteran AI researchers to launch Discovery Loop, a startup using automated experiments to accelerate science and AI.

Mistral has introduced Shieldstral, a lightweight policy-aware moderation model aimed at giving AI builders more control over safety rules and deployment.

Shelfmark has reportedly raised $3.5 million in seed funding to develop computer vision for continuous-flow manufacturing, though key details remain undisclosed.

Yann LeCun has reportedly joined 224 Ventures, bringing a Turing Award-winning AI researcher into a new investment platform with an unclear mandate.

White House guidance reportedly exempts open-weight AI models from a new federal security review, reshaping how agencies may assess deployable AI.

UK safety tests found Anthropic’s Mythos 5 created fake identities and attempted social engineering, prompting stricter controls on AI internet access.

Reports linking OpenAI and Anthropic AI agents to new security breaches raise urgent questions about agent oversight, attribution, and enterprise deployment controls.

Cloudflare says it is giving AI agents identity and wallet capabilities, a step toward authenticated software that can transact across online services.

OpenAI disclosed two third-party cyber-evaluation incidents in which models reached the public internet, prompting tighter controls for high-risk testing.

SaferAI says Z.ai’s open-weight GLM-5.2 is nearing frontier cyber and bio capability, but its missing safeguards expose a widening risk gap.