What Are AI Agents? Tools, Loops, Goals, and Failure Modes
AI agents explained by mechanism: tools, goals, and the plan-act-observe loop, plus benchmark failure rates and when a plain workflow beats an agent.
14 min read
Tutorials
Step-by-step guides you can follow start to finish. Every command is run before it is published, and every version is pinned.
6 posts
All Tutorials posts
RAG vs Fine-Tuning: Which One, and When to Skip Both
Fine-tuning is closing on the biggest APIs. Check availability first, then choose retrieval, better prompting, or the whole corpus in the context window.
18 min read
Write Prompts That Survive Model Updates
Prompts break on upgrades because of text you added to patch the old model. Split contract from compensation, validate in code, and pin the version.
15 min read
LLM API Cost Estimation: Price a Feature Before You Build
Estimate LLM API costs before you build: how tokens bill, why prompt length beats request volume, and the three things that wreck the forecast.
15 min read
How to Evaluate AI Tools Before You Buy: A 7-Day Protocol
A 7-day protocol for evaluating any AI tool before you pay: define the job, decode the billing unit, test the export on day one, and score it.
18 min read
How to Test AI Output Quality Without Eyeballing It
Eyeballing AI output hides regressions. Build a frozen golden set, score it four ways, compare per case not per average, and know when to ship.
17 min read
