AI articles
Long-form technical articles on AI agent security, agent architecture, prompt engineering, model evaluation, and evidence-based AI workflows.
Browse articles by recency or topic
Start with a topic page for structured reading paths, or use Latest articles for the newest writing. The full grouped index lives on a separate page to keep this overview focused.
Browse by topic
Topic pages are the primary reading paths. Each topic page groups its core articles and related section resources.
Topic
Agent security
Trust boundaries, authorization & access control, orchestration (control-flow mechanisms), policy enforcement, and observability.
Topic
Agent architecture
Workflows, state and lifecycle management, tool invocation patterns, retrieval and context management, and evaluation harnesses.
Topic
Model training and evaluation
Reliability, evaluation methods, calibration, grounding, and benchmark interpretation limits.
Topic
Prompt engineering
Operational prompting notes, evidence contracts, and reusable prompt templates and workflow files.
Latest articles
Newest published pages (auto-generated).
Observed Classification Layers in ChatGPT: User Access, Prompt Demand, and Capability Allocation
A client-side black-box analysis of observed ChatGPT classification artifacts, separating user access, prompt demand, and capability allocation.
Parallel Exploration in LLM Systems Is an Orchestration Pattern
Why multi-path reasoning in LLM systems usually comes from inference-time orchestration rather than ordinary single-pass autoregressive decoding.
Using ChatGPT Effectively at Work: A Practical Guide
A practical guide to choosing the right ChatGPT layer for work: modes, search, deep research, agent mode, personalization, memory, and projects.
Connected apps expand the capability and authorization surface of LLM systems
Why app-connected and MCP-enabled LLM systems should be analyzed as capability, scope, approval, and side-effect control problems—not only as prompt-processing systems.
Web-retrieved content is a prompt-injection boundary in tool-using LLM systems
Why retrieved web content must stay non-authoritative in browsing-enabled or tool-using LLM systems, and how to keep it from steering routing, tool arguments, or side effects.
Why “Almost Human, But Not Quite” Feels Wrong: From Clowns to AI-Generated Images and Text
Two separable mechanisms behind the “something feels off” reaction: cue-level perceptual mismatch (uncanny/cue conflict) vs AI-label effects on credibility and sharing.
All articles
Use the separate page for the full grouped index.
Subscription
Unlock the full version and working files
This article is public. The subscription unlocks the protected workflows, full versions, and working files across Andy's AI Playbook.