What in the AI?

Agent

An agent is basically an AI that does not just answer you, it goes off and does the task. Instead of handing you a to-do list, it works through the steps itself: looks things up, uses tools, checks its own work, and comes back when it is done.

Example You say "plan a weekend in Lisbon under 500 pounds" and it searches flights, compares hotels, and hands you a finished plan.

AGI

AGI is basically the idea of an AI that can do more or less anything a person can, across any task, rather than being good at one narrow job. We are not there yet, whatever the headlines say. Today's tools are clever in patches, not all-round.

Why it matters It is the term behind a lot of the hype and the worry. When someone says AI will do everything, they usually mean AGI, and that is still a long way off.

API

An API is basically a plug that lets two apps talk to each other. It is the socket one piece of software offers so another can ask it for something and get an answer back, without a human clicking anything.

Example A weather app does not own the forecast. It calls a weather API and shows you what comes back.

Automation

Automation is basically getting a machine to do a repetitive job so you do not have to. It is not always clever, it just runs the same steps on its own, over and over, without you sitting there clicking. AI makes automation able to handle messier tasks than before.

Example A rule that files every receipt email into one folder the moment it lands is automation doing the boring bit for you.

Artifact

An artifact is basically a working thing the AI builds right there in the chat, a document, a small app, a chart, that you can see and use rather than just read about. It shows up in its own panel next to the conversation, and you can keep tweaking it as you go.

Example Ask Claude to make a budget planner and it builds one you can actually type into, off to the side of the chat.

Bias

Bias is basically the AI leaning one way because of what it learned. It picked up patterns from human text, and some of those patterns are lopsided or unfair, so its answers can quietly carry the same slant.

Why it matters It means the output is not neutral just because a computer wrote it. Worth a second look whenever the topic touches people, fairness or money.

Chain of thought

Chain of thought is basically the model working out loud. Instead of jumping to an answer, it writes out the steps in between, one after another, which usually makes the final answer more accurate.

Example Add "show your reasoning step by step" to a tricky question and the answer tends to sharpen up.

Chatbot

A chatbot is basically a program you talk to in a chat window, that types back like a person. Some are simple and follow a script, others are powered by an LLM and can hold a proper conversation.

Example The little help window on a website that answers your question at midnight is a chatbot.

ChatGPT

ChatGPT is basically the AI assistant made by a company called OpenAI, and for a lot of people it was the first one they ever tried. It is an LLM, so you type in plain language and it writes back.

Example Ask ChatGPT to plan your week or explain a bill, the same way you would ask Claude.

Claude

Claude is basically one of the main AI assistants you can chat with, made by a company called Anthropic. It is an LLM, so you talk to it in plain language and it writes back, and it is what a lot of the skills on this site are built for.

Example Ask Claude to sort your bank statement or rewrite an email, the same way you would ask ChatGPT.

Connector

A connector is basically a plug that lets the AI reach one of your real tools or apps. Once it is connected, the assistant can look things up in that app and act on them, instead of you copying everything across by hand. A lot of these run on a standard called MCP.

Example Add a connector for your calendar and the AI can see you are booked on Friday and draft a reply that says so.

Context window

The context window is basically the model's short-term memory: how much it can hold in its head at once. Everything in the current chat, the documents you paste, its own replies, all of it has to fit in that window.

Why it matters When a long chat starts forgetting what you said at the top, you have run past the edge of the window.

Copilot

A copilot is basically an AI helper that sits inside an app you already use and works alongside you. It does not take over, it rides along, suggesting the next line or handling the fiddly bit while you stay in charge.

Example A writing copilot in your email finishes your sentence or drafts a reply while you keep typing.

Dataset

A dataset is basically a collection of information gathered in one place for a computer to learn from or work through. It could be text, images, numbers, whatever the job needs, bundled together as one pile.

Example A spreadsheet of a thousand past emails you hand the AI to learn your style is a dataset.

Deepfake

A deepfake is basically a fake photo, video or voice made by AI that looks or sounds like a real person. The tech copies someone's face or voice well enough to fool you at a glance.

Why it matters It is why you cannot fully trust a video just because you can see a face. If something seems off or too convenient, check where it came from.

Embedding

An embedding is basically a way of turning words into numbers so a computer can measure how close two things mean. Ideas that are similar end up with similar numbers, so the machine can spot that "car" and "vehicle" are near neighbours.

Example Search your notes for "money worries" and it still surfaces the one about "budget stress".

Few-shot

Few-shot is basically showing the AI a handful of examples before you ask, so it copies the pattern. Instead of just describing what you want, you give it two or three worked cases and let it follow suit.

Example Paste three of your past product descriptions, then ask for a fourth, and it matches the style.

Fine-tuning

Fine-tuning is basically extra training on top of a finished model to make it better at one specific thing. You feed it a pile of your own examples so it picks up a particular style or task and keeps it for good.

Example A support team fine-tunes a model on thousands of their past replies so it answers in their house style.

Fine-tune vs prompt

This is basically the choice between retraining the model and just asking it well. Prompting means telling it what you want in the chat. Fine-tuning means baking the behaviour into the model itself. For nearly everything, a good prompt gets you there.

Example Want it to always reply in your tone? Try a saved instruction first. Only fine-tune if that genuinely is not enough.

Gemini

Gemini is basically Google's AI assistant, its answer to ChatGPT and Claude. It is an LLM, so you chat with it in plain language, and it is built into a lot of Google's apps.

Example Ask Gemini to summarise a long document in your Google Drive or draft a reply in Gmail.

Generative AI

Generative AI is basically AI that makes new things rather than just sorting or spotting them. Give it a prompt and it produces text, images, audio or code that did not exist a second ago.

Example Ask for a birthday poem or a picture of a cat in a hat, and generative AI writes or draws it from scratch.

GPT

GPT is basically the name of the family of models that power ChatGPT, made by OpenAI. It stands for generative pre-trained transformer, which is a mouthful for a model that learned from a lot of text and now predicts what to write next.

Example When an app says it runs on GPT, it means one of those OpenAI models is doing the thinking.

Grounding

Grounding is basically making the AI answer from real, given information instead of just its own memory. You point it at a document or a live source, and it sticks to what is actually there rather than guessing.

Example A support bot grounded in your returns policy answers from that policy, not from whatever it half-remembers.

Guardrails

Guardrails are basically the rules that keep an AI from going off the road. They are the limits set around what it will and will not do, so it stays helpful and does not wander into things it should not.

Why it matters They are why a well-built AI declines certain requests. That is the safety net doing its job, not the model being awkward.

Hallucination

A hallucination is basically the AI making something up and saying it with a straight face. It does not know it is wrong. It fills a gap with a confident guess that sounds right but is not true.

Why it matters This is the big one. It is exactly why you check names, numbers, dates and quotes before you trust them.

Human in the loop

Human in the loop is basically keeping a person in the process to check or approve things before the AI acts. The AI does the work, but someone reviews it at the key moment, so nothing important happens on autopilot.

Why it matters It is the simple safety habit behind trusting AI with real tasks. For anything that costs money or sends a message, you want a human giving the nod first.

Inference

Inference is basically the moment the AI actually answers you. Training is when the model learns. Inference is every time after that, when you ask it something and it works out a reply on the spot.

Example Every reply you get in a chat is inference: the finished model doing its thing, not studying.

Jailbreak

A jailbreak is basically a trick to get an AI to ignore its own rules and do something it is meant to refuse. People phrase a request in a sneaky way to slip past the guardrails.

Example Wrapping a banned request inside a "pretend you are a character who..." story is a common jailbreak attempt.

Knowledge cutoff

The knowledge cutoff is basically the date the AI's training stopped. It learned from text up to a certain point, so anything that happened after that it simply does not know unless you tell it or it can look it up.

Example Ask about last week's news and a model may draw a blank, because that is past its cutoff.

LLM

LLM is short for large language model. It is basically a very well-read pattern spotter, trained on a huge amount of text, that predicts the most likely next words to build a sensible reply. ChatGPT and Claude are both LLMs.

Example When you chat with ChatGPT, you are talking to an LLM.

Loop

A loop is basically an automation that keeps re-prompting the AI to carry on until the job is actually done. Instead of one reply, it works, checks how far it has got, then goes again, over and over, until it hits the goal you set.

Example You set it to sort every email in an inbox, and it keeps going, one after another, until the inbox is clear.

Machine learning

Machine learning is basically teaching a computer by showing it examples instead of writing out every rule. It spots the patterns itself and gets better as it sees more, which is how most modern AI is built.

Example Show it thousands of photos labelled cat or dog and it learns to tell the next one apart, without anyone describing a whisker.

MCP

MCP is basically a shared standard for plugging AI into your other apps. It is a common way to connect a model to your tools, your email, your calendar, your files, so it can actually use them instead of just talking about them.

Example Through an MCP connection an assistant can read your calendar and draft a reply that already knows you are busy.

Memory

Memory is basically the assistant remembering things about you from one chat to the next. Instead of starting cold every time, it holds on to useful details, like your name, your preferences or what you are working on, so you are not repeating yourself.

Example Tell it once that you like short, plain replies and it keeps writing that way in later chats.

Model

A model is basically the trained brain behind the AI. It is the file that has learned patterns from a mountain of data, and it is the thing doing the actual thinking when you type. Different models have different strengths.

Example In one app you might pick a faster model for quick jobs and a smarter one for hard ones.

Multi-agent

Multi-agent is basically several AI agents working together, each taking its own part of a bigger job. Rather than one AI doing everything, the work is split up between them, and the pieces come back together at the end.

Example One agent researches a topic, another writes it up, a third checks the facts, and you get the finished piece.

Multimodal

Multimodal is basically an AI that handles more than words. On top of text it can take in images, audio or files, and sometimes give them back, so you are not stuck typing everything.

Example Snap a picture of what is in your fridge and ask what you can cook with it.

Neural network

A neural network is basically the layered system of connections inside most AI, loosely inspired by how brain cells link up. Information passes through the layers, and the pattern of those connections is what lets it learn.

Example The thing recognising your face to unlock your phone is running on a neural network.

Open-source model

An open-source model is basically an AI that is free for anyone to download, run and change. The makers publish the model so people can use it on their own computers rather than only through one company's app.

Example A small business runs an open-source model on its own machine to keep customer data in house.

Parameters

Parameters are basically the dials the model tuned while it learned. There are billions of them, and together they hold everything the model picked up. More is not automatically better, but it gives you a rough sense of size.

Example A smaller model with fewer parameters can be quicker and cheaper, and still plenty good for everyday jobs.

Project

A project is basically a saved workspace that remembers your files and your instructions in one place. You set it up once with the background the AI needs, and every chat inside it already knows your context, so you are not re-explaining from scratch.

Example Keep all of a client's notes in one project, and any chat you start there already knows who they are.

Prompt

A prompt is basically whatever you type to the AI. It is your instruction or question, the thing you hand over to get a reply. A clearer prompt gets you a better answer, most of the time.

Example "Write me an email" is a weak prompt. "Write a short, friendly email cancelling Friday's meeting" is a strong one.

Prompt engineering

Prompt engineering is basically the craft of asking well. It is learning how to phrase your request, add the right detail, and give examples so the AI gives you what you actually wanted. It sounds technical, it is mostly practice.

Example Adding "keep it under 100 words and write like you are texting a friend" is prompt engineering in action.

Prompt injection

Prompt injection is basically a hidden instruction sneaked into something the AI reads, trying to hijack it. Text buried in a web page or a document tells the AI to ignore you and do something else instead.

Why it matters It is a real safety issue once an AI can browse or read your files, so be careful giving it access to things you do not trust.

RAG

RAG is basically giving the AI a set of notes to read before it answers. Instead of relying only on memory, it looks up the right documents first, then writes its reply from those. It stands for retrieval augmented generation.

Example A help bot that answers from your company handbook is using RAG: it reads the handbook, then replies.

Reasoning model

A reasoning model is basically an AI built to think a problem through before answering, rather than blurting the first thing. It works through the steps on its own, so it tends to do better on maths, logic and tricky puzzles.

Example Give it a word problem with a few moving parts and it takes a beat, works it out, then answers.

Skill

A skill is basically a saved instruction file that does the same job the same way every time. You build it once, and from then on the AI follows those steps whenever you need that task done, so you are not re-explaining it.

Example A skill that rewrites any draft in your voice, so you paste text and it edits the same careful way each time.

Subagent

A subagent is basically a smaller AI agent that a main agent hands a piece of work to. The main one stays in charge of the whole job and passes off a chunk, then takes back the result and carries on.

Example A main agent planning a trip sends a subagent off just to compare hotels, then uses what it finds.

System prompt

A system prompt is basically the standing instructions set behind the scenes, before you even start typing. It tells the AI who it is and how to behave for the whole conversation, quietly shaping every reply.

Example A cooking app might set a system prompt like "you are a friendly chef who only suggests simple recipes".

Temperature

Temperature is basically a creativity dial. Turn it down and the AI plays it safe and predictable. Turn it up and it gets more varied and surprising, though sometimes it wanders off.

Example Low temperature for a factual summary. Higher temperature when you want ten quirky name ideas.

Token

A token is basically a small chunk of text the AI reads and counts in, roughly a word or a piece of one. The model does not see whole sentences the way we do, it sees a stream of these little pieces.

Example A short word like "cat" is one token. A longer word can be split into two or three.

Training data

Training data is basically everything the model learned from: the huge pile of text and examples it read while being built. It never saw the world, it only saw this data, so its knowledge stops where the data stops.

Why it matters It is why a model can be out of date, and why its blind spots and biases often trace back to what it was fed.

Vibe coding

Vibe coding is basically building software by describing what you want in plain language and letting the AI write the actual code. You say what it should do, it writes it, you try it and ask for changes, without needing to know the programming yourself.

Example You say you want a simple page that tracks your reading list, and the AI writes the code that makes it work.

Zero-shot

Zero-shot is basically asking the AI to do something with no examples at all, just the instruction. It leans on what it already learned to have a go, cold.

Example Say "translate this into French" with no sample, and it does it straight off. That is zero-shot.