You
Ask a question, upload a photo or use a service.
Start here / no jargon
AI explained.
Finally.
You do not need to know how to code to understand AI. Here is the short, honest version.
DOT / EXPLAINERCertainty is a style choice. Evidence is the receipt.
The simple version
A very clever pattern finder.
AI is software that learns patterns from examples. Some tools recognise images, some recommend a route, and some generate words or pictures. They can be useful without understanding the world as a person does.
Keep this in mind
“Useful first draft. Human final say.”
For everyday tasks, treat AI like a keen assistant whose work you still review.
How it works
Four steps, one idea.
A simplified way to understand many AI tools. Different systems work differently, but this is a good starting map.
Data
The system has learned from many examples.
Pattern
It looks for relationships that fit your request.
Prediction
It produces a likely response, recommendation or action.
A quick look back
This didn't happen overnight.
1950
Alan Turing proposes a way to discuss machine intelligence.
1956
Researchers formally name the field “artificial intelligence”.
1966
ELIZA shows how computers can imitate conversation.
1997
Deep Blue defeats chess champion Garry Kasparov.
2011
Siri brings voice assistance to millions.
2016
AlphaGo beats Go champion Lee Sedol.
2022
ChatGPT makes generative AI widely accessible.
Now
AI tools are moving into everyday work and life.
You've already met AI.
It has been quietly helping with everyday tasks for years.
Spotify recommendations
Gmail spam filters
Google Maps routes
Amazon suggestions
Face Unlock
Netflix recommendations
The words people use
A little less alphabet soup.
Tap a card for the plain-English meaning, or see the full glossary — every term, why it matters, and where to read more.
Two-minute check-in
What have you sussed?
Five questions. Helpful explanations, whether you get them right or not.
Your next small step
Ready to try AI for yourself?
Advanced layer
Go one orbit deeper.
Clear enough for a beginner. Substantial enough to keep the curious ones busy.
Context is not memory
A context window is the material a model can actively consider in one response. Persistent memory is a separate product feature. Long context helps, but relevance and structure still matter. And every word of it is billed ↗
Training is not searching
A base model learns statistical patterns during training. Search and retrieval systems add current or private sources at answer time. That distinction explains why some answers cite the web and others do not.
Agents add a loop
Give a model tools, a goal and a repeatable observe–act–check loop and you have the beginnings of an agent. Useful, powerful and still in need of boundaries. What an agent actually is ↗