Page 82
12 articlesThe Second Opinion Economy: When Dual-Model Verification Actually Pays Off
Using a second LLM to verify the first sounds obvious. In practice, almost nobody does it well. Here's the cost-benefit framework that tells you when to bother.
The Three Clocks Problem: Why Your AI System Is Living in Three Different Timelines
Production AI systems run on three unsynchronized clocks — wall time, model knowledge cutoff, and RAG index freshness — creating silent failures that standard monitoring never catches.
The Warm Standby Problem: Why Your AI Override Button Isn't a Safety Net
As AI agents absorb tasks humans used to own, the humans nominally in charge lose the competence to take over when things go wrong. Here's how to design escalation paths that actually work.
Treating Your LLM Provider as an Unreliable Upstream: The Distributed Systems Playbook for AI
LLM APIs fail differently from every other upstream dependency — they return 200 OK while producing hallucinated garbage. Here's how to adapt circuit breakers, timeouts, fallbacks, and bulkheads for the unique failure modes of production AI.
Agent Behavioral Versioning: Why Git Commits Don't Capture What Changed
Git commits and semver fail to capture what actually changed in AI agent behavior. Learn how behavioral snapshots, flip-centered gating, and trajectory test suites define what a 'version' really means for non-deterministic systems.
The AI Delegation Paradox: You Can't Evaluate Work You Can't Do Yourself
Engineers who delegate coding to AI lose the very skills needed to verify its output. Research shows developers are 19% slower with AI tools while believing they're 20% faster — a 39-point perception gap that drives a dangerous feedback loop of declining code quality.
AI Feature Decay: The Slow Rot That Metrics Don't Catch
AI features degrade not from model changes but from the world shifting underneath — user behavior evolves, knowledge goes stale, and eval suites ossify while dashboards stay green. Here's how to detect and prevent the silent quality collapse that hits most AI features within 90 days.
The AI Skills Inversion: When Junior Engineers Outperform Seniors on the Wrong Metrics
AI coding assistants make junior engineers look 6x more productive on dashboards while masking architectural decay, measurement distortion, and a mentorship collapse that threatens the entire engineering pipeline.
The AI Team Topology Problem: Why Your Org Chart Determines Whether AI Ships
Where AI engineers sit in the org chart is the biggest predictor of whether ML projects ship or stall — a breakdown of centralized, embedded, platform, and federated team models with their failure modes and maturity progression.
CLAUDE.md as Codebase API: The Most Leveraged Documentation You'll Ever Write
Your CLAUDE.md is an API contract between your codebase and every AI agent that touches it. Learn the instruction budget constraints, anti-patterns that degrade agent performance, and the progressive disclosure architecture that scales.
Compound AI Systems: Why Your Best Architecture Uses Three Models, Not One
Production AI systems that compose a classifier, generator, and verifier consistently outperform single frontier models — delivering higher accuracy at lower cost, as long as coordination overhead stays below the 40% latency threshold.
Database-Native AI: When Your Postgres Learns to Embed
PostgreSQL extensions like pgvector and pgai now handle embedding generation, vector search, and LLM calls inside the database — eliminating the sync pipeline most RAG architectures carry and keeping vectors transactionally consistent with source data.
About Tian Pan
I'm Tian Pan, an engineer-founder focused on agentic engineering — building autonomous AI systems and scaling engineering teams. I write practical guides on system design, technical leadership, and shipping with AI agents. Previously an early engineer at Uber, Brex, and IoTeX.
