AI Foundations Pro ~8 min read New · July 2026

Speak AI: the jargon, decoded.

AI conversations are full of terms that make smart people feel dumb — which is a tax on your decisions, because vendors exploit exactly that feeling. Fifteen definitions, plain English, each with the reason it matters to you.

01 The core mechanics

TermPlain EnglishWhy you care
TokenThe word-chunks AI reads and writes (~¾ of a word)It's how API usage is billed and limits are measured
Context windowHow much text the model can 'hold in mind' at onceWhy long documents work in some tools and get 'forgotten' in others
InferenceThe act of the model generating an answer'Inference costs' = what running AI actually costs vendors — and eventually you
ParametersThe model's internal dials, set during trainingMostly a bragging number; capability ≠ parameter count anymore
Training vs. fine-tuningBuilding the brain vs. giving it a specialty course'We fine-tuned a model' is a real (modest) claim; 'we built our own AI' usually isn't

02 The capability layer

RAGFetching relevant documents and handing them to the model before it answersHow 'chat with your data' products actually work — retrieval quality is the whole product
AgentAI that takes multi-step actions toward a goal, using toolsThe word every vendor now uses; ask "what actions, with what supervision?"
MultimodalHandles images/audio/video, not just textWhy you can photograph a form and ask about it
Reasoning / thinking modesGenerating deliberate steps before answeringSlower, better on hard problems — you're buying rigor with seconds
MCP / connectorsStandard plugs connecting AI to apps and dataThe power AND the security surface — see our connector-security lessons

03 The evaluation layer

HallucinationConfident, fluent, wrongNot a bug being fixed next quarter — a property to manage forever
BenchmarkA standardized model testDirectionally useful, gameable, and rarely about YOUR task
Open vs. closed modelDownloadable weights vs. vendor-hosted APIControl-versus-convenience; relevant when data can't leave the building
Prompt injectionMalicious instructions hidden in content the AI readsTHE attack to understand before connecting AI to email and files
AGIHypothetical human-level-at-everything AIA debate, not a product. Anyone selling it is selling something else
The vendor-meeting superpower

Fluency's real payoff is the follow-up question: "Agentic how — what actions can it take, and what's the human checkpoint?" "RAG over which sources, updated when?" Watch pitches get honest fast when the jargon stops working as fog.

Try it now

Take the last AI product pitch you received and reread it with the tables above open. Highlight every term doing fog-work. Draft the two follow-up questions that would clear it.

Open the Playground →

This week's challenge

Use three of these terms correctly in real conversations this week — and once, politely ask a vendor or colleague to define one they used. Fluency plus the confidence to ask: that's the whole upgrade.

Up next in AI Foundations

Your AI stack

Stop collecting tools; assemble a stack — which AIs, for which lanes, at what spend. Read the lesson →