70-20-10 Learning Model
A learning approach where about 70 percent comes from hands-on practice, 20 percent from others, and 10 percent from formal teaching. The backbone of how the Bootcamp runs.
ISLANDBeginner’s compass
Find the plain-language meaning of every important Bootcamp term, then read it beside English in any of 18 translated languages.
Start with any word
New to AI? Type the word you heard, choose the language you know best, or explore one module at a time.
A learning approach where about 70 percent comes from hands-on practice, 20 percent from others, and 10 percent from formal teaching. The backbone of how the Bootcamp runs.
A ChatGPT feature that writes and runs code to analyse data, files and spreadsheets for you.
Optimising your content so AI assistants recommend and quote your business in their answers. The AI-era evolution of SEO.
A process where one or more AI agents complete a chain of tasks with limited human input.
An AI that can plan and carry out multi-step tasks on your behalf, using tools and making decisions along the way.
The practical ability to understand, use and get real value from AI tools in everyday work and life.
A way for software tools to talk to each other and share data automatically.
A hypothetical future AI that could match or exceed human ability across virtually any task.
Computer systems that perform tasks normally needing human intelligence, such as understanding language, recognising images or making decisions.
The width-to-height shape of an image, such as 16:9 or 1:1.
A NotebookLM feature that turns your documents into a podcast-style audio conversation.
Setting up tasks to run automatically with little or no manual effort.
A realistic AI-generated presenter that can speak your script on camera, for example from HeyGen.
Supplementary footage cut in to support the main video, increasingly AI-generated.
Unfair or skewed patterns in AI output that come from imbalances in its training data.
A popular design tool with built-in AI for images, text and layouts.
A ChatGPT workspace for writing and editing longer documents or code side by side with the AI.
The final hands-on project where you apply everything you have learned to a real use case.
Prompting an AI to reason step by step, which improves accuracy on harder problems.
A program you interact with in conversation, increasingly powered by generative AI.
OpenAI's popular AI chat assistant, one of the main tools taught in the Bootcamp.
Anthropic's AI assistant, known for strong writing, reasoning and handling long documents.
AI that interprets and understands images and video, recognising objects, faces, text and scenes.
How much text an AI can hold in mind at once. Its short-term memory for a conversation or document.
Microsoft's AI assistant, built into Windows and Microsoft 365.
A personalised version of ChatGPT that you set up with your own instructions, knowledge and purpose. No coding needed.
Saved preferences that tell ChatGPT how to respond to you every time.
OpenAI's image generator, built into ChatGPT.
Protecting personal and sensitive information when using AI tools.
A powerful form of machine learning using many-layered neural networks, behind most of today's AI breakthroughs.
A synthetic video or image that convincingly replaces someone's face or voice. Powerful but easily misused.
An open, cost-efficient AI model family developed in China.
Publishing your app or website so that others can access it online.
The main technology behind AI image generation, which turns random noise into a clear image step by step.
The web's address system. A domain is your website's name, such as learn-ai.co.za.
A leading AI voice tool for realistic text-to-speech and voice cloning.
A way of storing information by meaning so AI can find the most relevant pieces. The memory behind many AI apps.
Giving an AI a few examples inside your prompt so it copies the pattern you want.
Further training a model on specialised data so it performs better for a specific purpose.
A large, general-purpose AI model trained on broad data that can be adapted to many different tasks.
When an AI decides to use an external tool or action, like searching the web or sending an email, to complete a task.
Google's AI assistant, built into Google's products.
AI that creates new content such as text, images, audio, video or code, rather than only analysing or sorting existing data.
A platform for storing, sharing and collaborating on code and projects.
The family of models behind ChatGPT. The name stands for generative, pre-trained transformer.
OpenAI's marketplace where Custom GPTs are shared and discovered.
Rules and limits that keep an AI system's behaviour safe, on-topic and within policy.
When an AI confidently produces information that is wrong or made up. Always worth checking important facts.
Keeping a person involved to guide, check and approve what the AI produces, rather than letting it run unsupervised.
Generating a new image using an existing image as a starting reference.
Turning a still image into a moving video clip.
Editing part of an image by regenerating a selected area.
Learn AI's reframe of 'AI': treat these tools as a capable assistant you delegate work to, rather than a mysterious artificial intelligence.
The date after which a model has no built-in knowledge of world events.
Documents you upload to a Custom GPT so it can answer from your own material.
An AI model trained on huge amounts of text so it can understand and generate human-like language. The engine behind ChatGPT, Claude and others.
Matching a character's mouth movements to a soundtrack or voice.
A no-code AI tool that builds working websites and apps from a plain-language description.
A type of AI where systems learn patterns from data and improve with experience, instead of being programmed with fixed rules.
A visual no-code automation platform for building multi-step workflows.
A feature where ChatGPT remembers useful details about you across conversations.
A leading text-to-image tool known for its artistic quality.
The trained 'brain' behind an AI tool: the underlying system that has learned from data and produces the output.
An open standard that lets AI assistants securely connect to apps, data and tools.
AI that can work across several types of input and output at once, such as text, images, audio and video.
The simplest working version of an idea that you can build to test whether it delivers value.
AI built to do one specific task well, like the tools we use today, as opposed to general intelligence.
Telling an image AI what to leave out of the picture.
A computing structure loosely inspired by the human brain, made of connected nodes, that underpins most modern AI.
Building apps and automations with visual tools and little or no programming.
Technology that reads text out of images or scanned documents.
A model whose underlying weights are freely available to download and run, unlike closed commercial models.
Coordinating multiple AI tools, steps or agents to work together on a bigger task.
The internal settings a model adjusts as it learns. More parameters usually means a larger, more capable model.
South Africa's data-protection law governing how personal information may be collected and used. Important when using AI with real data.
The instruction, question or input you give an AI tool to tell it what you want.
The skill of writing clear, well-structured prompts to get better, more reliable results from AI.
A small trial that shows an idea can actually work before you invest fully.
A newer type of model that thinks for longer before answering, for tougher problems.
Using AI in a way that is ethical, fair, transparent and safe.
A technique where an AI looks up relevant information from a document set before answering, improving accuracy.
A training method where human ratings teach a model to give more helpful, safer answers.
The value or payoff you get back relative to the time or money you put in.
A popular AI video creation and editing platform.
A number that controls the random starting point of an image, letting you reproduce or vary results.
OpenAI's text-to-video generation model.
Turning spoken audio into written text. Also called transcription.
A popular open image-generation model that you can run yourself.
An AI tool that creates full songs, with music and vocals, from a text description.
An artificial voice generated by AI.
A behind-the-scenes instruction that sets an AI's role, tone and rules for a whole conversation.
A setting that controls how creative or random an AI's output is. Lower is more focused, higher is more varied.
Creating a picture from a written description.
Turning written text into natural-sounding spoken audio.
Creating a video clip from a written description.
The small chunks of text, roughly part of a word, that a language model reads and generates. Usage and limits are counted in tokens.
The large collection of text, images or other examples that an AI model learns from.
The breakthrough AI architecture, the 'T' in GPT, that powers modern language models.
Increasing an image's resolution and detail using AI.
A specific, real-world task or problem that you apply an AI tool to.
Google's advanced text-to-video model.
Creating software by describing what you want in natural language and letting AI write the code.
Creating a synthetic copy of a specific person's voice from a sample of their speech.
Talking to an AI assistant out loud and hearing it reply in real time.
Hidden or visible marks that flag content as AI-generated.
An automatic message one app sends another when something happens, often used to trigger automations.
OpenAI's speech-to-text model, widely used for transcription.
Chaining tasks and tools together so a whole process runs automatically.
A no-code automation tool that connects apps so actions happen automatically.
Asking an AI to do a task with no examples, relying on its existing knowledge.