Michael Greenlaw on AI, DevSecOps, and FedRAMP Automation: Speed, Risks, and the Need for Fundamentals

Tom and Michael discuss Microsoft’s strategy and licensing shifts pulling organizations toward Windows and Azure, and the frustration of incompatible cloud service naming across AWS/Azure/GCP. Michael describes using Claude to learn and build with LLMs rapidly, calling it “fast Google,” but warns about hallucinations, compounded errors in chained tasks, human limits in reviewing massive AI-generated code, poor handling of branching strategies, and the need for humans to own architecture and fundamentals.

They debate layoffs attributed to AI versus post-COVID overhiring and shareholder optics, and note AI’s growing impact on vulnerability discovery and exploitation.

Book recommendations: “How to Win Friends  ...