Older posts
A growing collection of previous posts on artificial intelligence and emerging tech trends.
The Core Metrics Every AI Agent Needs in Production
Production-ready AI agents in 2026 need a balanced scorecard covering task success, tool use, latency, cost, safety, and escalation — not just answer quality.
Development
From Business Goals to Agent Behaviour: Mapping OKRs to AI Actions
Learn how OKRs, desired and prohibited behaviours, and new governance frameworks are reshaping how enterprises deploy and control AI agents in 2026.
UX
How to Write Effective OKRs for AI Agents
Learn how AI agent OKRs shift focus from technology outputs to business outcomes like resolved cases, reduced review time, and safety in 2026.
Artificial Intelligence
Agentic AI Has Changed the Threat Model for Digital Products
Agentic AI has shifted from answering questions to taking actions, and the security frameworks defining how to deploy it safely have finally converged in 2026.
Development
Data Residency Is Becoming a Design-Tool Requirement
Figma has launched local data hosting in Japan, and the move reflects a broader 2026 shift toward residency-first enterprise design and AI tools.
Tech
What Happens When AI Becomes a Paid Product Feature?
The 2026 AI pricing reset changed how we pay for intelligence. Here is what convergence, bundling and Meta One mean for your subscription choices.
UX
When AI Behaves Unexpectedly: A Better Model for Incident Reporting
OpenAI's September 2026 misalignment reporting framework borrows from aviation safety culture, but its real test lies in consistent application.
Ethics
How to Make Your Design System Usable by AI Agents
Design systems in 2026 have evolved into agent-ready infrastructure, combining machine-readable schemas, MCP servers, agent skills, and CLIs.
Knowledge Architecture
The Case for Optimism: Why the AI Safety Debate Could Make AI Better
Responsible AI has matured into a formal discipline in 2026, but agentic systems, thin assurance markets and capacity gaps are testing its limits.
Ethics
Safety Gates for AI: What Product Teams Can Learn From Frontier Model Development
Red teaming has become a mandatory release gate for frontier AI in 2026, reshaping how product teams handle evaluation, governance and audit trails.
Development
When AI Agents Act Alone: Designing Human Oversight Into Autonomous Systems
Bounded autonomy has become the 2026 standard for enterprise AI agents. Learn how permissions, oversight, and UX design keep humans in control.
UX
The AI Race Has a Coordination Problem: Why No Lab Can Slow Down Alone
Frontier AI faces a collective action problem where rational competition produces dangerous outcomes. Here is how coordination could break the cycle.
Ethics