Now.

Active threads in the lab

• Strategic thread (Essays): security as a runtime architecture concern in agentic systems, moving from static perimeter models to continuous runtime decision-making with explicit gates, evidence requirements, and tier-specific controls.
• Tactical thread (Notes): reusable patterns for runtime security gates, evidence collection infrastructure, autonomy-tier-aware policy enforcement, and audit-trail design for security decisions.
• Compounding model: many-to-many links between essays and notes so insights build over time across governance, security, autonomy, and operating model dimensions.

What I am actively exploring

• AI-native enterprise architecture: how architecture evolves when systems can reason, not only execute.
• Agentic systems and enterprise coordination: where agents sit above systems as decision and orchestration layers.
• AI-first operating models: moving from process-centric workflows to signal-driven decision loops.
• Build Factory and productized thinking: engineering systems that can produce and iterate systems quickly.
• Architectural control planes: policy, governance, and orchestration boundaries for intelligent systems.
• The autonomous enterprise: autonomy as an architectural evolution, not a single technology event.

What I am intentionally pausing

• Generic AI commentary with low structural insight.
• Tool comparisons and model-versus-model debates.
• Hype-cycle narratives without architectural grounding.
• Code-only deep dives that skip operating-model implications.
• Generic digital transformation language without AI-native depth.

Current sharpening

The lab is now sharpening the security architecture thesis: how do agentic systems move security from a static perimeter to a runtime architecture concern? The practical answers are explicit gates per action tier, evidence-first evaluation, machine-verifiable security decisions, and audit trails that make every security decision inspectable. This connects to earlier work on fail-closed controls, trust boundaries, and machine-verifiable governance.

How I am publishing right now

I am using a practical three-part stream to keep the lab useful and honest: Pattern notes, Observations and learnings, and Thesis essays.

The expected flow is: Observation -> Pattern -> Essay -> Reuse. This keeps exploration grounded, testable, and structurally coherent over time.

Core thread across the work

Enterprises are slowly evolving from deterministic systems toward intelligent adaptive systems, and architecture is the discipline that will determine whether that transition succeeds.