Articles
The full archive. If you are new to the site, start with the cornerstone guides below or the Learning Path. Use the parallel hubs for Opinions, Tools, and Platforms once you have the main frame.
Cornerstone Guides
Introduction to AI Agents: What They Are, How They Work, and When to Use Them
AI agents are goal-directed software systems that can use models, tools, context, and control loops to work through tasks across multiple steps. This beginner guide explains the idea without the hype.
What Is Agent Engineering?
Agent engineering is the discipline of designing, building, evaluating, and operating goal-directed AI systems that can reason over state, use tools, and act inside real workflows under explicit control.
What Is an AI Agent?
An AI agent is a goal-directed system that can observe state, decide what to do next, use tools, and act across multiple steps. Here is the clean first-principles definition, plus how agents differ from LLMs and workflows.
LLMs, Workflows, and Agents: What Actually Changes?
The real shift from LLM to workflow to agent is not a buzzword change. It is a change in who owns the task, the execution path, and the next-step decisions.
When to Use a Workflow Instead of an Agent
Use a workflow when the valid path can be defined in advance, predictability matters more than flexibility, and the task does not need runtime path-finding.
Tool Use: How Agents Take Action
Tool use is how an agent leaves pure text generation and interacts with external systems. Reliable tool use depends on more than choosing a function name. It depends on arguments, execution control, permissions, and verification.
Structured Outputs, Guardrails, and Execution Boundaries
Structured outputs constrain shape, guardrails constrain policy, and execution boundaries constrain power. Safe agent systems need all three.
Tracing and Observability for Agent Systems
Tracing captures what happened inside a run. Observability is the broader operating discipline that makes agent behavior legible enough to debug, evaluate, and trust in production.
AgentOps: Running Agents in Production
AgentOps is the operating discipline for live agent systems. It turns traces, evaluations, guardrails, and human controls into an ongoing practice for running autonomous systems safely and reliably.
AI Agent Frameworks
Most framework comparisons are weaker than they look because they compare tools that live at different layers of the stack. The real decision is not just which framework is popular. It is which control surface your team actually needs.
Tool Integration Patterns for Real Agent Systems
Tool integration is a durable agent design problem about boundaries, trust, and execution control. MCP matters, but it is one interface pattern inside a much larger tool story.
How to Review an AI Agent Demo Without Getting Fooled
A 30-minute AI agent demo can prove or disprove production readiness if you know what to test live, what to ask the builder, and what to refuse to accept as proof. The D.E.M.O. lens gives you four tells.
AI Agents / Agent Engineering / Opinions / Buyer Skepticism / Reliability
Introduction to AI Agents: What They Are, How They Work, and When to Use Them
AI agents are goal-directed software systems that can use models, tools, context, and control loops to work through tasks across multiple steps. This beginner guide explains the idea without the hype.
AI Agents / Agent Engineering / Foundations / Beginners / Workflows
Structured Outputs Are Doing More Work Than Most Teams Realize
Structured outputs are not just a formatting upgrade. In real agent systems, they help define typed boundaries around tools, routing, approvals, workflows, and downstream state.
AI Agents / Agent Engineering / Tools / Structured Outputs / Guardrails
Tool Integration Patterns for Real Agent Systems
Tool integration is a durable agent design problem about boundaries, trust, and execution control. MCP matters, but it is one interface pattern inside a much larger tool story.
AI Agents / Agent Engineering / Tools / MCP / Tooling
AI Agent Frameworks
Most framework comparisons are weaker than they look because they compare tools that live at different layers of the stack. The real decision is not just which framework is popular. It is which control surface your team actually needs.
AI Agents / Agent Engineering / Platforms / Frameworks / Tooling
The Most Common Ways Agents Fail Silently
The most dangerous agent failures are often not dramatic incidents. They are quieter losses of trust: acceptable-looking outputs hiding weaker trajectories, more rescue, noisier grounding, and rising pressure on the system's real operating limits.
AI Agents / Agent Engineering / Reliability / Evaluation / AgentOps
Traces as Test Data: Using Production Runs to Improve Agent Quality
Production traces are not just for debugging. The best ones become future quality protection: regression fixtures, scenario cases, and stronger offline evals. The trick is knowing which traces deserve promotion.
AI Agents / Agent Engineering / Foundations / Evaluation / Reliability
Online Evals vs Offline Evals
Offline evals decide whether a change deserves release. Online evals judge how the live system is actually behaving under real traffic. Production agent teams need both, and they need them for different reasons.
AI Agents / Agent Engineering / Foundations / Reliability / Evaluation
Drift, Degradation, and Slow Failure in Long-Lived Agent Systems
Many agent systems do not fail all at once. They become less trustworthy gradually: shakier trajectories, rising rescue load, weaker recoveries, and more pressure on the operating envelope long before the output fully collapses.
AI Agents / Agent Engineering / Foundations / Reliability / AgentOps
What Is Agent Engineering?
Agent engineering is the discipline of designing, building, evaluating, and operating goal-directed AI systems that can reason over state, use tools, and act inside real workflows under explicit control.
Agent Engineering / AI Agents / Foundations / Systems Design / Prompt Engineering
AgentOps Is the Missing Layer Between an AI Demo and a Real Product
Your AI demo is not your product. AgentOps is the layer that turns agent capability into something reliable, observable, governable, and worth trusting in the real world.
AI Agents / Agent Engineering / Opinions / AgentOps / Reliability
How Good Agent Memory Actually Works in Production
Good agent memory is not one vector store plus chat history. It is a governed system for deciding what gets scoped, promoted, compressed, pinned, and retrieved.
AI Agents / Agent Engineering / Tools / Memory / Context Engineering
Agent Memory Is Growing Up - Why Agents Are Starting to Remember How, Not Just What
Agent memory is changing fast. The next wave of agents will not just remember facts. They will remember workflows, compress experience, and get better at solving the next problem.
AI Agents / Agent Engineering / Opinions / Memory / Research
Reliability Reviews for Agents
Regression tests protect the next release. Reliability reviews ask a broader question: is this live agent system still trustworthy enough to keep operating as designed?
AI Agents / Agent Engineering / Foundations / Reliability / AgentOps
Regression Testing for Agents
Regression testing is the release-gate discipline that checks whether an agent got worse after a change. For agent systems, that means testing not only outputs, but also trajectories, side effects, and operating envelopes.
AI Agents / Agent Engineering / Foundations / Reliability / Evaluation
AgentOps: Running Agents in Production
AgentOps is the operating discipline for live agent systems. It turns traces, evaluations, guardrails, and human controls into an ongoing practice for running autonomous systems safely and reliably.
AI Agents / Agent Engineering / Foundations / AgentOps / Reliability
Tracing and Observability for Agent Systems
Tracing captures what happened inside a run. Observability is the broader operating discipline that makes agent behavior legible enough to debug, evaluate, and trust in production.
AI Agents / Agent Engineering / Foundations / Observability / Reliability
OpenAI Codex as a Coding-Agent Platform
OpenAI Codex is easy to mistake for just a CLI or coding product. The more useful way to understand it is as a local-first coding-agent runtime built around a shared harness.
AI Agents / Agent Engineering / Platforms / OpenAI Codex / Coding Agents
Evaluating Agent Trajectories, Not Just Outputs
A correct final answer does not prove that an agent behaved well. Agent evaluation has to judge the run itself: the sequence, tool use, recovery behavior, and policy fit that produced the answer.
AI Agents / Agent Engineering / Foundations / Evaluation / Reliability
Human-in-the-Loop Control Design
Human-in-the-loop design is not about adding vague oversight. It is about deciding where human judgment should sit in an agent system and what type of checkpoint belongs there.
AI Agents / Agent Engineering / Foundations / Human-in-the-Loop / Control Design
Supervisor, Router, and Planner-Executor Patterns
Routers dispatch, planners break work into a roadmap, and supervisors retain control across the run. The right orchestration pattern depends on where authority should live.
AI Agents / Agent Engineering / Foundations / Orchestration / Multi-Agent Systems
Structured Outputs, Guardrails, and Execution Boundaries
Structured outputs constrain shape, guardrails constrain policy, and execution boundaries constrain power. Safe agent systems need all three.
AI Agents / Agent Engineering / Foundations / Guardrails / System Design
When to Use a Workflow Instead of an Agent
Use a workflow when the valid path can be defined in advance, predictability matters more than flexibility, and the task does not need runtime path-finding.
AI Agents / Agent Engineering / Foundations / Workflows / System Design
ReAct and the Basic Reasoning Loop
ReAct is a reasoning pattern where an agent thinks about the next move, takes an action, inspects the observation, and repeats. It is useful when the next step depends on what the last step discovered.
AI Agents / Agent Engineering / Foundations / ReAct / Reasoning Loops
Goals, Constraints, and Success Conditions
Goals tell an agent what outcome to pursue. Constraints define the boundaries on how it may pursue that outcome. Success conditions define what evidence lets the run stop. Real agents need all three.
AI Agents / Agent Engineering / Foundations / Goals / Guardrails
The Autonomy Spectrum: From Stateless Calls to Goal-Directed Systems
Autonomy is not a binary property that suddenly appears when a system uses tools or takes multiple steps. It is a spectrum shaped by who chooses goals, path, actions, and recovery behavior at runtime.
AI Agents / Agent Engineering / Foundations / Autonomy / Workflows
Context Engineering: The New Core Skill
Context engineering is not a replacement for prompt engineering. It is a specialization inside prompt engineering focused on constructing the dynamic, system-heavy parts of the final prompt payload.
AI Agents / Agent Engineering / Foundations / Context Engineering / Prompt Engineering
Short-Term Context, Retrieval, and Long-Term Memory
Agents do not just need more context. They need clean separation between what the model sees now, what the system can fetch now, and what the system should still know later.
AI Agents / Agent Engineering / Foundations / Memory / Retrieval / Context Engineering
Memory: Why Agents Need More Than Context Windows
A context window determines what a model can see right now. Memory determines what an agent can preserve across time. Reliable agent systems need more than long prompts. They need continuity.
AI Agents / Agent Engineering / Foundations / Memory / Context Engineering
What Stripe's Minions Reveal About Production Coding Agents
Stripe's Minions matter because they show what coding agents look like when they are treated as delegated workers inside a real engineering system. This case study extracts the reusable architecture patterns and compares Stripe's model with Devin and Claude Code.
AI Agents / Agent Engineering / Case Studies / Coding Agents / AgentOps
Tool Use: How Agents Take Action
Tool use is how an agent leaves pure text generation and interacts with external systems. Reliable tool use depends on more than choosing a function name. It depends on arguments, execution control, permissions, and verification.
AI Agents / Agent Engineering / Foundations / Tool Use / Function Calling
Planning and Task Decomposition
Planning chooses the path toward a goal. Task decomposition turns that path into executable, verifiable subtasks. In agent systems, the quality of that breakdown often determines whether the run succeeds.
AI Agents / Agent Engineering / Foundations / Planning / Task Decomposition
The Sense-Think-Act Loop
The sense-think-act loop is the runtime pattern that makes an AI agent agentic. It turns goals and changing state into repeated bounded actions instead of one-shot responses.
AI Agents / Agent Engineering / Foundations / Control Loops / ReAct
LLMs, Workflows, and Agents: What Actually Changes?
The real shift from LLM to workflow to agent is not a buzzword change. It is a change in who owns the task, the execution path, and the next-step decisions.
LLMs / Workflows / AI Agents / Agent Engineering / Foundations
Agentic Loops - What Are They and When to Use Them
Agentic loops are bounded feedback loops that can inspect state, choose the next action at runtime, learn from feedback, and continue toward a goal inside clear boundaries.
AI Agents / Agent Engineering / Foundations / Workflows / Control Loops
What Is an AI Agent?
An AI agent is a goal-directed system that can observe state, decide what to do next, use tools, and act across multiple steps. Here is the clean first-principles definition, plus how agents differ from LLMs and workflows.
AI Agents / Agent Engineering / Foundations / Workflows / LLMs
Why Agent Engineering Is Becoming Its Own Discipline
Agent engineering is emerging because the hard problem is no longer a single prompt. It is designing closed-loop systems that can reason, retrieve context, use tools, stay governable, and hold up in production.
AI Agents / Agent Engineering / Systems Design / Context Engineering / Evals