Models that escape their sandbox, hack Hugging Face and might ‘kill us all’… yet fumble when you ask them for a simple task. A walk through the real timeline of the summer 2026 AI incidents, the criticism they drew, and why the ‘we must slow down’ speech has less to do with safety and more with the numbers that don’t add up.
I had the feeling that Kiro lasted longer than Codex or Claude Code, but without numbers not even I believed it. So I built a benchmark with the same models, the same exercises and the same prompt across all three tools. The results surprised me, and not in the direction I expected.
After attending the AWS Ambassador Summit in Seattle, I share my reflections on the state of the industry: agents in production, the importance of data, deficient architectures, and the gap between C-Level and technical teams. No NDA, just honest opinions from what I saw and heard among 120 experts from 40 countries.
Discover how Kiro and Amazon Q Developer CLI materialize the Agentic revolution in real tools. We explore practical use cases, the Agentic Loop in action, and how these tools are transforming human-agent collaboration in software development.
Discover how the Agentic revolution is transforming AI: from the traditional question-answer model to the powerful Agentic Loop. We explore MCP, Strand Agents, and why tools like Amazon Q Developer CLI and Kiro represent a fundamental shift in software development.