00Start here · zero judgment

You Said AI What?

The part of the program where we admit AI is confusing, everyone fakes it sometimes, and you are allowed to know nothing yet. No jargon, no quizzes, no judgment — just enough to survive the next meeting without nodding along to something you don't understand.

What "AI" actually is

A really good guesser with a very big memory. It doesn't "know" things the way you do — it predicts what answer sounds right, based on patterns it saw in data. Sometimes that looks like magic. It's not magic. It's statistics with better marketing.

What it isn't

It isn't conscious. It isn't secretly watching you with opinions. It isn't a person you should feel guilty about delegating to. And it isn't always right — confident ≠ correct. That's the whole ballgame.

Myths that need busting

“AI thinks like a human.”

It doesn't. It's a prediction engine. Thinking is yours to do — AI just drafts the options faster.

“The AI is biased, so I can't use it.”

It can be biased — that's why you review the output. Knowing that is step one of using it responsibly, not a reason to avoid it.

“If the AI says it, it must be true.”

It says things confidently. Confidence is not accuracy. Check the important stuff, always.

“AI will do my whole job.”

It does pieces of jobs. The humans who keep the judgment, the review, and the accountability are the ones who thrive.

“I have to know how the model works to use it.”

You don't need to know how the engine works to drive the car. You need to know how to steer and where to brake.

The vocab cheat sheet

The words people throw around — decoded in one line each.

LLM

A really good next-word guesser. 'Large Language Model' — the engine behind most chatbots.

Prompt

What you say to the AI. Garbage in, garbage out — this is the part you control.

Hallucination

When the AI makes something up with total confidence. It does this. Check important facts.

Token

A chunk of text the AI reads (roughly a word). More tokens = more context = better answers.

Fine-tuning

Teaching a pre-trained model a specialty. Like sending the intern to a conference.

RAG

Giving the AI a document to look at before it answers — 'Retrieval-Augmented Generation.' Reduces hallucination.

Agent

An AI that can take steps and use tools on its own. Exciting and needs supervision.

Context window

How much conversation the AI remembers. Run out and it forgets the beginning.

Model

The actual AI brain — trained on data, then used to predict answers.

Now you can say "AI" without faking it

From here, the six tracks take you deeper — durable capability, evidence-based assessment, and judgment that transfers between tools. Pick the one that matches what you need next.