19 min ·
AI Agent Examples: What Makes Each One Actually an Agent
AI agent examples sorted by what makes each one an agent: its loop, tools, stop rule, outside check and blast radius. Classify your own. Read Chapter 1.
19 min ·
AI agent examples sorted by what makes each one an agent: its loop, tools, stop rule, outside check and blast radius. Classify your own. Read Chapter 1.
20 min ·
AI agent vs chatbot comes down to who decides the next step. Place your product in one of four columns and describe it honestly. Read Chapter 1 free.
23 min ·
What is the ReAct agent pattern? See one task run three ways, where the original paper won and lost, and what a written thought costs. Pick your shape.
22 min ·
Why do AI agents hallucinate? The model writes the likeliest text where a fact is missing. Sort your case into seven causes and pick the fix that fits.
29 min ·
This AI agent architecture guide reduces any design to four parts and a loop, orders fifteen decisions by stage, and gives you a doc skeleton to copy.
11 min ·
AI agents book Kindle, paperback or PDF? How each edition handles diagrams and code, what each gives up, and a free test to run first. See the formats.
23 min ·
AI agents for backend developers, sorted: the habits that carry over, the three that break and one new trust boundary. Find your ticket and start there.
21 min ·
The best books on AI agents, compared by the author of one on shelf life, framework ties, prerequisites and evaluation. Filter the table by your situation.
28 min ·
Build an AI agent from scratch in one Python file: no framework, no API key, a scripted stand-in model, six ways to break it. Copy the file and run it.
13 min ·
Read AI agents book online free: the Preface, Chapters 1 and 2 and the glossary of AI Agents, Engineered, mapped section by section. Check the fit first.
20 min ·
Stop conditions for agent loops, made operational: three exits, four statuses, and caps set from measured successful runs. Copy the stop-policy card.
19 min ·
What is agentic AI? A model choosing its own next steps, and a label applied far more widely. Take three questions to your next vendor demo.
17 min ·
What is an AI agent? A model that picks its own next step, in a loop, within limits set by code. Run five checks on any system and get a verdict.
14 min ·
Agents vs workflows comes down to one question: who decides the next step, your code or the model? Learn the rule, then read Chapter 1 free.
14 min ·
What is an agent harness? The loop, tools, history, stop rules and recovery around the model, what breaks without each, and when to adopt a framework.
12 min ·
What is an agent loop? The four beats every agent repeats, how ReAct fits, and the three exits ranked by trust, with a worked trace. Read the walkthrough.
12 min ·
Compounding errors in AI agents are worse than pⁿ: self-conditioning, snowballing and correlated errors. See the evidence and the redesign that fixes it.
23 min ·
Design agent tools and schemas for software built for humans: choose a surface per operation, wrap it in a five-part contract. Copy the checklist.
16 min ·
MCP vs function calling: one is how a model asks for a tool, the other how programs find and share tools. See when a protocol pays and what to review.
22 min ·
What is MCP in AI? A contract between programs for sharing tools. See the count behind it, seven questions to ask of any protocol, and what to review.
19 min ·
When your AI agent calls the wrong tool, the record shows which of six causes it is. Classify the call, run five checks, apply one fix and measure it.
24 min ·
How to design tools for LLM agents: shape each tool to a task, not an endpoint. One API wrapped two ways, a symptom table and a template to copy.
25 min ·
To sandbox AI agent tool execution, isolate files, network, credentials and limits. Take the requirements table, 16 wall tests and a policy block to copy.
23 min ·
Structured output for tool calls comes in two grades. See what each guarantees, what neither covers, and copy a boundary check that holds under both.
26 min ·
What is tool calling in LLMs? An exchange: the model writes a request, your code runs the function, the result goes back. See who does each step.
21 min ·
Tool calling LLM agents pay for every definition on every pass. Sort each gap into tool, skill, protocol, rule or code, price the move, copy the skeleton.
14 min ·
Idempotent tools for AI agents turn a retry or crash-resume into a no-op, not a second charge. Learn keys, receipts and retry rules with a worked example.
23 min ·
Agent memory architecture in one page: four kinds of memory ranked by write risk, three designs compared, and a write policy to copy. Pick your design.
22 min ·
Context compaction for AI agents, built and tested: sort every item into verbatim, pointer, summary or drop, write the step, and prove the run survives it.
19 min ·
How to give an AI agent memory in four steps: an inventory, run notes, a reviewed file, a store on a trigger, and a test that it helped. Start here.
21 min ·
RAG vs agents is a choice on two levels of one product. Sort your questions, pick one of three designs, and see what each costs before you scope it.
18 min ·
RAG vs long context is settled by permissions and fit before accuracy. Use the 11-row decision table and the cached-corpus arithmetic to choose yours.
16 min ·
Write select compress isolate context engineering, one operation at a time: how each works, how it fails, what it costs and how to measure it. Start here.
25 min ·
A context engineering checklist of 40 pass-or-fail lines for long agent runs, with a budget that adds up and a policy block. Audit one recorded run.
22 min ·
Context engineering vs prompt engineering: what is a rebrand, what changed with agents, and three tests that separate a prompt failure from a context one.
18 min ·
What is context engineering? Deciding what a model reads on every call. Learn what is sent each time, three limits on it, and why a model seems to forget.
25 min ·
When an AI agent ignores instructions, one of four causes is at work. See what 14 studies measured, then run four checks on the failing request.
11 min ·
What is context rot? Why an AI agent degrades as its context window fills, the five failure modes behind it, and four fixes. Learn to diagnose a run.
17 min ·
An AI agent design patterns cheat sheet: 22 patterns, each with its shape, cost in calls, typical failure, the check and the simpler row. Filter the table.
17 min ·
Add approval gates to AI agents by what a wrong action costs, not by model confidence. Tier each tool, design the card, test that people still read.
23 min ·
The autonomy slider AI agents sit on has four positions, set per task. See what evidence moves a task one notch, then copy the per-task record.
23 min ·
The evaluator optimizer pattern loops generate, check and revise. See which evaluator adds information, when to stop, and test it against plain retries.
19 min ·
How to build a multi agent system as a state machine: filled worker briefs, a merge that flags conflicts and escalates, and caps. Copy the pseudocode.
16 min ·
Human in the loop AI agents fail when one person must read every line. Sign the intent, the plan, the irreversible actions and the evidence. See where.
17 min ·
When not to use AI agents: nine checks to run on a proposal, and what to build instead when one fails. Tick the checklist against your own design.
28 min ·
Agentic workflow patterns as a backend engineer sees them: four shapes, the check at each seam, the cost in calls, a symptom table. Then read Chapter 10.
27 min ·
Model an AI agent as a state machine: write the outer loop as a state table with four exits and a done check the worker cannot vote on. Copy the table.
25 min ·
Multi agent system overhead in tokens, turns and failure surface, every study cited with its denominator. Budget your split, then read Chapter 11.
19 min ·
Which multi agent systems book still applies to LLM agents? Ten compared, the classic ideas that transfer, and a rule to pick one. Filter by your job.
27 min ·
Orchestrator worker pattern: a six-field worker brief, one task briefed badly and well, and a symptom table. Copy the template, then read Chapter 11.
16 min ·
Single agent vs multi agent: start with one, split only when branches run blind and merge cheaply. Use the decision table, then read Chapter 11.
23 min ·
An AI agent evaluation checklist in six stages, from the first eval to production, each check with the evidence that proves it. Tick it against your agent.
19 min ·
Agent observability records what an agent did. See which questions a trace answers, which need a judgment, and how traces become evals. Copy the note.
19 min ·
Build a golden dataset for LLM evaluation from production traces: what to sample, how to label, how to split and when to retire it. Copy the case record.
16 min ·
How do you know if an AI agent is working? Usage, uptime and its own 'done' won't tell you. Five independent checks will, with a stop number. Read on.
17 min ·
How to debug an AI agent in six steps: freeze the run, measure a failure rate, find the first wrong step, classify it, fix it, keep the test. Start here.
23 min ·
How to monitor AI agents in production: five signals beyond uptime, an alert rule for each, the sample size behind quality, and a dashboard spec to copy.
20 min ·
How to test AI agents: exact tests for tools, parsers and stop rules, pass rates on a fixed world for the model side, and a flaky-case procedure. Copy it.
23 min ·
Write an LLM as a judge rubric with four anchored levels, a worksheet before the verdict and two bias tests. Copy the template and run the checks.
23 min ·
LLM as a judge vs human evaluation, decided per criterion: a sorting rule, an allocation table and a two-stage check that proves the judge. Copy the plan.
23 min ·
Regression testing for LLM applications needs a gate that survives noise: a paired rule on flipped cases, three verdicts and a triage list. Copy the spec.
19 min ·
Agent trajectory evaluation, done honestly: check the end state first, then the path. A hand-graded example, a metric table and a template to copy.
26 min ·
AI agent evaluation metrics in one table: what each number measures, its formula, how it misleads, and the four to report first. Copy the template.
25 min ·
LLM tracing with OpenTelemetry starts from a span contract: what each agent span must carry and never record, mapped to two moving conventions. Copy it.
16 min ·
pass@k vs pass^k: one asks if any of k attempts works, the other if all k do. See which one your product needs and how to compute both. Copy the template.
15 min ·
How many eval examples do I need? Twenty to fifty at first, hundreds to compare versions. The arithmetic, paired tests, and a worked example to copy.
15 min ·
Is LLM as a judge reliable? Only as far as you have measured it. A calibration method with Cohen's kappa, a worked example and the four biases to test.
23 min ·
An AI agent governance framework on one page: owners, tool tiers, the stop, logs and promotion evidence, each mapped to a public clause. Copy the page.
18 min ·
Your AI agent keeps looping: re-running a command, retrying an error, never finishing? Read the transcript, find which of five loops it is, and apply its exit.
22 min ·
AI agent reliability for enterprise buyers: fifteen vendor questions, the answers that should worry you, and a test for your own tasks. Copy the list.
23 min ·
Durable execution for AI agents replays your code against a journal. Trace a crash line by line, find the one step that can run twice, then decide.
20 min ·
How much does an AI agent cost to run? Build the figure from six measured terms, a task mix and your own rates, with no price list to go stale. Fill it in.
20 min ·
How to make AI agents more reliable in six steps, in build order: what each mechanism buys, what it costs and the test that proves it. Run the sequence.
17 min ·
How to reduce AI agent costs: six levers priced on one illustrative workload, ranked by return per effort, each with its quality risk and proof. Rank yours.
23 min ·
An indirect prompt injection attack hides instructions in content an agent reads. See 24 documented cases, graded by evidence, then test your own agent.
23 min ·
How to shadow deploy an AI agent, then canary it: five rungs, what each must prove, the arithmetic behind the slice, and a rollout plan you can copy.
27 min ·
AI agent guardrails are checks in code that decide if an action happens. See 19 candidates sorted by where each runs, its cost and its test. Start here.
21 min ·
The blast radius of AI agents can be estimated before you grant access. Fill in one sheet per agent, answer three questions per row, and drill the stop.
21 min ·
How to prevent prompt injection in AI agents: you can't fully, so lower the rate cheaply and cap the damage in code. See which defenses hold up.
15 min ·
AI agent failure modes sorted for diagnosis: symptom, cause, trace signature and fix for each, cross-checked against MAST. Keep it open beside your traces.
16 min ·
The lethal trifecta AI agents carry: private data, untrusted content, an outbound channel. Run the audit, cut one leg, cap the blast radius. Start here.
19 min ·
How to choose an AI agent framework without a ranking: eight jobs it can own, a scoring rubric, a three-day spike and the red flags. Copy the rubric.
18 min ·
AI agent ROI is hours saved minus the cost of checking and of the errors that land. Follow one worked case from first sum to corrected sum, then run yours.
18 min ·
Search "AI agent works in demo fails in production" and you get taxonomies. One case shows what a demo measures and nine questions to ask before a date.
21 min ·
AI agents for product managers: the six decisions that are product calls, the artifact that records each, and a requirements page to copy before kickoff.
20 min ·
Build an LLM classifier without a training set: a label spec, hand labels split into a dev set and a sealed test, and the numbers to report. Copy the spec.
17 min ·
Build vs buy AI agents, decided layer by layer: own the evals, tool contracts and traces, rent the model and framework, and test every exit. See the table.
23 min ·
A coding agent in a legacy codebase fails where a new hire struggles. Audit one service, fix it in a safe order and count before and after. Copy the audit.
23 min ·
LLM classifier vs fine tuned classifier: a decision table with three verdicts, the studies on both sides, and the test each must pass. Copy the recipe.
29 min ·
An agentic coding workflow for teams: the artifact, signer and meaning of done at each stage, plus a size rule and an owner rule. Copy the agreement.
16 min ·
AI agent pricing models compared on the cost curve: margin per seat, task, usage and outcome, and why outcome pricing comes last. Run the worksheet.
23 min ·
AI agent use cases for startups, sorted into eight recurring shapes, each with its check, starting autonomy and usual failure. Place your own idea.
31 min ·
Spec driven development with AI agents on one backend feature: the interview, the full spec, criteria turned into tests, and drift. Copy the worked spec.
22 min ·
An agentic AI curriculum ordered by prerequisites: a dependency table, seven public course schedules compared, and a rationale to paste. Copy the template.
23 min ·
AI agents for beginners: four ideas and 14 terms to learn first, and ten topics to skip for now, each with the symptom that brings it back. Start here.
64 min ·
AI agents lecture slides for a full thirteen-week course: 80-minute lectures with presenter notes, in-class exercises and a CC BY 4.0 attribution line.
27 min ·
AI agents assignments for students, as a bank of ten designs: what the grader checks, how each can be gamed and what it costs to grade. Copy the rubric.
29 min ·
An AI agents learning roadmap in four stages: the chapters to read, one build, one proof and the check that ends each stage. Find your stage today.
26 min ·
An AI agents study guide: 87 glossary terms mapped to 27 chapters, a 35-question self-test with a separate answer key, and one question. Start free.
18 min ·
The best way to learn AI agents is to build one small agent a program can check. Get the first project, a 20-task eval plan and the arithmetic to copy.
21 min ·
Learn AI agents from scratch in three stages: the loop, then tools, then evals. Get the build, the exit check and the too-early signs for each stage.
22 min ·
Teaching AI agents to computer science students without tying the course to one vendor: what goes stale, a syllabus audit and a tool policy to copy.
13 min ·
An AI agents course syllabus for instructors: why the loop comes first, how to grade verification over demos, and how peer courses compare. Get the kit.