Prompt Rules Are Suggestions, Not Guardrails
Three production databases deleted in nine months by AI agents that understood their rules and violated them anyway. The agents aren't the problem. The deployments are.
On AI governance, agentic engineering, and what actually matters in enterprise technology.
The Storage Pool Column You Never Configured
Azure lets you deploy SQL Server VMs with a single-disk storage pool and never mentions that the column count is permanent. By the time you're chasing I/O, the only fix is a rebuild.
Azure Backup's Expiry Dates Are Lying
When you stop protection and retain data in Azure Backup, the expiry dates on your recovery points become cosmetic. They display. They don't trigger. Your data lives forever — and so does the bill.
The AI Governance Gap
AI collapsed the cost of building software to near zero. The cost of governing it didn't change. Most companies haven't noticed the gap yet — but their breach reports will.
Where Your Cloud Budget Actually Goes
Azure Advisor says you're optimized. The audit says otherwise. Here's what I actually find when I look at where the money goes across enterprise tenants — and none of it is in the recommendations dashboard.
Agentic Engineering vs. Vibe Coding
AI didn't make everyone a developer. It made everyone think they're a developer. The difference matters more than most companies realize.
The Most Dangerous Engineer in the Room
Confidence without understanding is the highest-risk pattern in enterprise IT. The 6 Ps aren't a bumper sticker — they're the methodology that separates the engineer who prevents the outage from the one who causes it.