Amazon Web Services (AWS) is launching a dedicated GitHub organization for its most experimental agentic AI work, reports Frederic Lardinois in today’s lead story.

On Monday, the company launched Strands Labs, where teams across Amazon will publish frontier projects that aren’t quite ready for inclusion into the production-ready version of the company’s Strands Agents SDK.

The initial release includes two projects: AI Functions, which generates code at runtime from natural-language specifications, and Strands Robots, which connects large language models to physical hardware via vision-language-action (VLA) models.

In an interview with The New Stack, Clare Liguori, a Senior Principal Engineer at AWS, said the release of this new feature to foster experimentation creates a needed boundary between the stable SDK releases and more experimental work:  “We want to make sure that the SDK continues to focus on that production-ready solution,” Liguori says.

Read the full story: AWS creates a sandbox for its agent experiments

More stories for your mind this Monday:

➮ This simple infrastructure gap is holding back AI productivity

➮ Cloudflare’s Markdown for Agents automatically make websites agent-ready

➮ Wasm vs. JavaScript: Who wins at a million rows?

➮ Why Kubernetes 1.35 is a game-changer for stateful workload scaling

➮ Claude Code comes to Roadmap, OpenClaw loses its head, and AI workslop

Join us on March 12 as Broadcom’s Jad El-Zein shows how you can automate full-stack deployments of Kubernetes infrastructure – including networking and storage policies – with a simple git commit.

The VMware “exodus” is happening one workload at a time. Learn how 300+ IT leaders are managing the transition and avoiding common technical limitations. Build your exit strategy with proven data.