Ideas into perspective.
Field notes on investment research, applied AI, venture building, security, and the next generation of talent.
Explore The Observatory · Private research by invitationMore from the blog
Build with conviction. Verify with evidence.
A starting point for the TaylorVentureLab journal: connecting capital, applied AI, security, and talent through better questions—and evidence that can change our minds.
EU AI Act Q1 2025: Phased Deadlines & Annex III Education Classification
The EU AI Act is rolling out in phases through 2027. If your AI touches education, credentialing, or student profiling—here's what you need to prepare now.
UK AISI: Building Evaluation Artifacts for Third-Party Testing Expectations
The UK AI Safety Institute is shaping third-party evaluation expectations. Here's what evaluation artifacts should look like for frontier model assessments.
Philippines NPC: AI + Children Transparency & DPIA Requirements
The Philippines NPC has issued targeted guidance on AI systems processing children's data. Here's what child-oriented transparency and CPIA requirements mean for your products.
Singapore PDPC/IMDA 2025: AI Assurance Sandbox & Practical Testing Tooling
Singapore continues its practical approach to AI governance with new sandboxes, PET guides, and elevated data protection standards. Here's what's available for 2025.
NIL Legal Landscape 2025: Antitrust Pressure & State-Federal Dynamics
The NIL legal landscape in college sports remains unsettled. Here's how antitrust litigation and state actions are reshaping what schools, athletes, and dealmakers must comply with.
Philippines 2025: Cyber & AI Talent Demand Converges with GJobs Hiring Innovation
High-visibility breaches, government upskilling strategy, and platforms like PasaJob/GJobs are driving urgent cyber and AI talent demand in the Philippines.
Port Gating as a Control Plane: Designing Zero‑Trust That Doesn't Break Operations
Closed-by-default is easy to say and hard to ship. Here's how to implement port gating as an auditable control plane that keeps production moving.
Ontology Before Tooling: Why Modeling the System Comes First
Tools create data. Ontologies create understanding. Here's why modeling entities and relationships first is the only way to build autonomous security and governance.
The Cost of Identity Sprawl—and the Controls That Reduce It
Identity sprawl is the modern perimeter failure. Here's how to measure it, reduce it, and prevent it from returning—without slowing the business.
DNS Resilience as a Security Primitive
DNS is a control plane disguised as plumbing. If it fails—or is subverted—everything fails. Here's how to engineer DNS resilience as a security capability.
Why Explainability Is an Operational Requirement, Not a Marketing Feature
In regulated environments, 'the model said so' is not an acceptable control. Explainability is what makes automation auditable—and survivable.
Risk Scoring That Doesn't Lie: How to Weight Reality
Most risk scores are cosmetic. Here's a practical method to weight exposure, control strength, and freshness—so your score reflects reality, not politics.
Board Governance as a Security Control
Governance isn't paperwork. It's the operating system for trust. Here's how to structure board-level oversight that actually reduces risk.
How I Invest: A Governance‑First, Operator‑Led Style
Investment without governance is speculation. Here's the framework I use: operator credibility, control structures, and compounding potential.
Deterministic AI for High‑Stakes Decisions: Beyond Pure Statistics
When decisions have consequences, 'probably correct' isn't good enough. Here's why bounded, explainable computation is emerging as a requirement for regulated environments.
Want to discuss these topics?
I'm always interested in conversations about security architecture, AI governance, and systems optimization.
Get in touch