Skip to main content

On-ramp · Plain English

Start Here

A plain-English on-ramp for people who saw the demo and want to understand it — not build it yet. It decodes the words and the mental models behind agentic engineering, then points you at the four pillars when you want to go deep. Not a tutorial. Not "AI for dummies". The shortest path from "what am I looking at?" to "I get it."

Saw the demo. Want to understand it.

If you have watched an agent write code, fix a bug, or wire up a tool and thought “I have no idea what half those words mean” — this is for you. No coding required. The goal is understanding, not building.

You will leave able to follow the conversation: what an agent is doing, why a context window runs out, and what someone means by a harness. When a word clicks, a link takes you to the deep version.

The path

Three steps, about 20 minutes. Not rails — skip ahead whenever a word already makes sense.

Decode the words

The exact terms from the demo — agent, context window, MCP, harness — in plain English, each paired with the precise meaning.

Open the Decoder

See it in action

Watch a real agent do a real task, with every move decoded as it happens. The words, but moving.

Watch a demo, decoded

Understand the ideas

Six short reads, one mental model each — why it forgets, why it makes things up, why the same model differs across tools.

See the mental models

Two ways in

Decode the words, or watch them in action. Most people bounce between both.

Understand the ideas

Six short reads, one mental model each. Analogy first, paired with the precise term — so it sticks without dumbing it down.

Ready to go deeper?

The on-ramp gets you fluent. These are the expert pillars where each idea gets the full treatment — the same concepts, no training wheels.