Plain-English meanings for the keywords you hear in the club. Search to filter.
Want more? MIT Sloan keeps a good glossary of AI terms.
No matches. Add it with a pull request!
A set of rules that lets one program ask another program to do something. Like a waiter: you order from the menu, the kitchen does the work. Businesses use APIs to plug tools together (for example, an app calling FLUX to make a video).
Example: send a prompt to an AI model's API and get a video back.
A secret password that identifies you to an API and tracks your usage and bill. Never paste it into public code or a GitHub repo.
Store it in an environment variable or a .env file that is not committed.
A way to control a program by typing text commands instead of clicking. Faster to repeat and easy to automate.
Example: git status, or running Claude Code in your terminal.
The click-and-drag version of a tool, like CapCut's editor.
A standard that lets an AI assistant safely connect to outside tools and data (Gmail, Notion, calendars, databases). Think of it as a universal plug for AI.
Example: a Higgsfield or Notion connector inside Claude.
A ready-made link between an AI assistant and an app, often built on MCP.
A toolbox of pre-written code that makes an API easier to use in a language like Python.
An AI trained on huge amounts of text that can write, summarize, and reason. Claude, GPT and Gemini are LLMs.
The trained AI itself. Different models are good at different things (text, images, video, audio).
The instructions you give an AI. Clear prompts (goal, context, format, constraints) get better results.
A small chunk of text an LLM reads or writes. Pricing and limits are often counted in tokens.
How much text a model can keep in mind at once.
An AI that can plan steps and use tools to finish a task, not just answer one question.
A repeatable chain of steps (sometimes called a stack). Example: dictate, generate, edit, publish.
A model whose trained files are public so you can download and run it. License rules still apply, so check before using commercially.
Many are on Hugging Face.
Extra training of an existing model on your own examples so it fits your task.
A cheap way to fine-tune a big model by training a small add-on instead of the whole thing.
Using a trained model to produce an answer. This is what you pay for when you generate.
The chip that does the heavy math for AI. Renting GPUs is a major AI cost.
Running AI on your phone or laptop instead of a remote server. Better privacy and speed. Example: Google AI Edge Eloquent.
Computers you rent over the internet (AWS, Google Cloud, Azure).
Software that tracks changes to files so you can undo mistakes and collaborate.
A website that hosts Git projects (repositories) so people can share and work together.
A project folder tracked by Git. This site is one.
A saved snapshot of your changes with a message.
A proposal to merge your changes into a project, reviewed by others.
Free hosting that turns a repo into a website. This site uses it.
The front-page file of a repo explaining what it is and how to use it.
The language for asking questions of a database.
Organized storage for business data (customers, sales, inventory).
A simple text format for passing data between programs.
A trick where hidden text in a webpage or file tries to give an AI new instructions. Treat outside content as data, not commands.
When an AI states something false with confidence. Always verify important facts.
Making your business easy for AI assistants to find and recommend, like SEO for AI answers.
Improving your site so it ranks higher in search results.
The page a website shows when the address you asked for does not exist. A good one explains what happened and links back to useful pages. Code 404 means 'not found'.
Example: this site's 404 page appears if you mistype a link.
A file (sitemap.xml) that lists every page on a website so search engines can find them all. Like a table of contents for Google.
Example: this site's sitemap lists the home page, every tool guide and more.
A small file that tells search engines which parts of a site they may visit and where the sitemap is.
Hidden tags in a page that control the title, description and picture shown when you share its link in chats and social apps.
Example: the card you see when you paste a link in GroupMe or LinkedIn.
A one-sentence summary in a page's code. Search engines often show it under the page title in results.
A site made of ready-made files (HTML, CSS, JavaScript) with no server-side program. Fast, cheap or free to host, and easy to secure. This club site is one.
Your own copy of someone else's GitHub repo. You can change it freely without touching the original.
Example: fork the club repo, then edit your copy.
A side version of a project where you can try changes without affecting the main version.
Accepting a change and combining it into the main version of the project.
A website that can be installed on your phone or computer and opens like an app, without an app store.
Example: this site can be added to your home screen.
Details that can identify a person, such as a name with contact information, an ID or account number, or health details. Keep it out of AI tools unless the tool is approved for it.
A board with columns such as Ideas, In progress and Done. Cards move left to right so everyone can see the status of work.
Example: the club roadmap board.
The point where the savings from a choice equal its cost. After that, you come out ahead.
Example: how many videos per month before an AI subscription costs less than the old way.
Everything a person sees and touches on a screen: buttons, menus, colors, text and layout. Good UI looks clear and is easy to use.
Example: the Install app button and the search box on this site.
How easy, fast and pleasant it feels to get something done with a product, from first click to finish. UI is the look, and UX is the whole journey.
Example: if finding the meeting time takes one tap, that is good UX.
A simple black-and-white sketch of a page layout that shows where things go before any design or code. Cheap to change.
A clickable mock-up of a product used to test an idea with real people before building it fully.
Example: a Figma prototype of an app.
A page layout that adjusts to any screen size, from phone to laptop, so it works well everywhere.
Example: this site's menu wraps on small screens.
Designing so everyone can use a product, including people who use screen readers, keyboards only, captions or high-contrast text. Often required by law and good for everyone.
Example: adding alt text to images.
The part of a website or app people see and use, built with HTML, CSS and JavaScript.
The behind-the-scenes part that stores data, runs logic and talks to other systems, usually on a server.
Example: a database that saves orders.
Working on both the front end and the back end of an application.
The language that defines the structure of a web page: headings, paragraphs, links and images.
The language that controls how a web page looks: colors, fonts, spacing and layout.
The programming language that makes web pages interactive, such as search boxes, calculators and menus.
The readable address of a website, such as example.com, that you register and renew each year.
The internet's phone book. It turns a domain name into the server address where the site lives.
The service that stores a website's files and delivers them to visitors. Static sites can often be hosted free, as this one is on GitHub Pages.
The secure version of the web's connection. The padlock in the address bar means data between you and the site is encrypted.
Showing two versions of something (an ad, a page, an email) to different people to see which performs better.
Example: two ad videos, compare clicks.
The share of visitors who take the action you want, such as signing up or buying. Conversions divided by visitors.
Example: 5 sign-ups from 100 visitors is 5%.
AI that creates new content, such as text, images, audio, video and code, from a prompt.
Example: FLUX 3 making a video from a sentence.
A way of building software that learns patterns from examples instead of following hand-written rules.
Example: a spam filter trained on labeled emails.
A model made of layers of connected math units, loosely inspired by the brain. Most modern AI is built on them.
A model that can handle more than one kind of input or output, such as text, images, audio and video.
Letting an AI look up your own documents before it answers, so replies are grounded in real sources instead of memory alone.
Example: a chatbot that answers from a company's policy manual.
The examples a model learns from. The quality, variety and rights of this data shape what the model can do.
When an AI treats groups unfairly or gives skewed answers because of patterns or gaps in its training data.
AI-made or AI-edited media that makes it look like a real person said or did something they did not.
Example: never post one without permission and clear labeling.
Rules and checks that keep an AI system within safe, intended behavior.
A design where a person reviews or approves an AI's work before it takes effect.
Example: an agent approves each AI-written client email before it is sent.
A way of working in short cycles: build a little, get feedback, adjust. The opposite of planning everything up front.
A popular Agile method with fixed-length work cycles (sprints), regular check-ins and clear roles.
A short, fixed period (often one to four weeks) in which a team finishes a set of planned work.
A one-sentence request from the user's point of view: 'As a member, I want to add a prompt so I can share what works.'
Anyone affected by or interested in a project, such as customers, managers, members or investors.
When a product clearly solves a real problem for a group of customers who want it enough to keep using and recommending it.
A significant change of direction for a startup after learning that the first idea is not working.
Building a business with your own money and revenue instead of outside investment.
How much cash a company spends each month beyond what it earns. Divide cash on hand by burn rate to get runway.
How many months a company can keep operating before it runs out of cash.
The share of customers who stop buying or cancel in a period. Lower is better.
What it costs, on average, to win one new customer, including marketing and sales.
The total profit you expect from a customer over the whole relationship. Healthy businesses earn much more per customer (LTV) than they spend to win one (CAC).
Revenue minus the direct cost of delivering the product, shown as a percentage of revenue.
How well a business or system handles growth without costs or problems growing just as fast.
Example: a static website serves 10 or 10,000 visitors at about the same cost.
Tools and practices that turn company data into reports and dashboards people use to make decisions.
A single screen that shows key numbers and charts so you can see how things are going at a glance.
A central database built for reporting and analysis, combining data from many systems.
Pulling data out of source systems, cleaning and reshaping it, and loading it somewhere you can analyze it.
The rules for who owns data, who can use it, how accurate it must be and how it is protected.
The stages of building a system: plan, analyze, design, build, test, deploy and maintain.
Practices that bring software development and IT operations together to release updates quickly and reliably.
Automatically testing and releasing changes every time someone updates the code. A GitHub Action that checks every pull request is one example.
The future cost of quick shortcuts in code or systems. Like borrowing time now that must be repaid later.
Protecting systems, networks and data from attack, theft and damage.
Scrambling data so only someone with the right key can read it.
Logging in with more than a password, such as a code on your phone. One of the simplest ways to protect an account.
A fake message or site designed to trick you into giving up passwords, money or data.
Example: an email that looks like your bank asking you to log in.
The delay between asking for something and getting a response. Lower latency feels faster.
An automated path that moves data from where it is created to where it is stored and used.
Data that anyone can access and reuse, often published by governments or research groups.
The top leader. Sets vision and strategy and answers to the board.
Runs the company's information systems and technology: software, data, security and IT operations. This is a core Information Systems career path.
Leads building the company's technology and product. In a startup the CTO often leads engineering.
CIO usually runs internal systems; CTO usually builds what customers use.
The letters change by industry. In insurance and finance it is usually Compliance (follows laws and regulations). In sales-focused firms it may be Commercial (revenue). Some companies use it for Creative or Customer.
Always check which one a company means.
Runs money: budgets, reporting, fundraising, investor relations.
Runs day-to-day operations and makes the strategy actually happen.
Leads brand, marketing and growth.
A newer role guiding how a company adopts and governs AI.
Protects company data and systems from threats.
The group of top executives whose titles start with 'Chief'.
Senior managers who sit below the C-suite and run departments.
Decides what to build and why, working with engineers and designers.
A number that shows whether you are winning, such as leads per week.
What you gain compared to what you spent.
The simplest version you can launch to test an idea.
Who your customer is: other companies or individuals.
Software you rent monthly instead of buying (Notion, CapCut Pro).
Software that tracks leads, clients and conversations.
Big systems that run finance, inventory, HR and operations together.
Following the laws, regulations and rules for your industry.
An investment firm that buys and improves companies.
Investors who fund early-stage startups in exchange for ownership.
The total revenue possible if you won every customer in your market.
Whether you make money on each customer after the costs of getting and serving them.
The backend and frontend of how AI products work, plus the security and data basics every user should know.
The part of a website or app you see and touch: layout, buttons, colors, text. It runs in your browser or phone. Business view: it is the storefront. This club site is almost all frontend (HTML, CSS, JavaScript).
The part you do not see: servers, databases and logic that store data, check logins and do the work. Business view: the warehouse, accounting and staff behind the storefront. An AI chat app's frontend is the chat box; the backend sends your message to the model and returns the answer.
A computer that is always on and answers requests from other computers, such as sending a web page or running an AI model. 'The cloud' mostly means servers owned by someone else that you rent.
The device or app that asks a server for something, like your browser or phone. Client and server work as a pair.
Proving who you are (password, code, fingerprint). Authorization is different: it decides what you are allowed to do once you are in. AI agents need both, so they can only touch what you permit.
A second proof besides your password, such as a code on your phone. It stops most account takeovers. Turn it on for GitHub, email and any AI account.
A setting stored outside your code, often used for secrets like API keys. The file is usually named .env and should be listed in .gitignore so it is never uploaded.
A file that lists what Git must not upload, such as .env files and large temporary files. It protects secrets.
Software whose code is public and can be used or changed under a license. Open-weight AI models are a close cousin. Always read the license before using something commercially.
The file that tells others what they may do with your project. MIT is a common permissive choice: use and change it freely, keep the credit, no warranty. Without a license, the default is that nobody may reuse your work.
A link that lets two systems exchange data, like an AI agent reading your calendar. Each integration needs permissions and creates risk, so connect only what you need.
An automatic message one system sends to another when something happens, such as 'new order received.' It is how many automations trigger without you checking.
A saved copy of something so it loads faster next time. If you change a website and still see the old version, your browser cache is the usual reason: hard refresh.
Putting a finished version live where users can reach it. GitHub Pages deploys this site each time a change is merged.
The share of time a system is working. Businesses promise it in service agreements, for example 99.9%.
A separate copy of data kept in case the original is lost or damaged. Version control helps for code, but important data still needs its own backup.
A record of who did what and when. It is how you review what an AI agent did and how companies prove compliance.
Give a person or AI only the access it needs to do the job, and nothing more. The most important safety rule for AI agents and connected apps.
A job that runs by itself on a schedule, such as 'every morning at 7, search for news and summarize it.' See the scheduled tasks guide.
A computer simulated inside another computer. Some AI agents get their own private VM in the cloud to browse and run tasks, kept apart from other users.
A cap on how many requests you can send in a time period. Free AI tiers and APIs use them to control cost and prevent abuse.