Technical Enough

Living document

Glossary.

Every technical term used in this course, translated to plain English. Some entries point to the lesson that introduces them. New terms get added as the course grows.

A

Agentic / Agent modeAI
An AI mode where the model takes multi-step actions autonomously, often across files or external services, with limited interruption. 'Cowork' and agent modes in modern tools are this. Use when you want the AI to take a bigger step on its own.See: Day 18
AI (Artificial Intelligence)AI
Any system that does something a human would have done, by learning from data instead of being explicitly programmed step by step. In 2026, when people say 'AI,' they usually mean LLMs (see below).
AI APIAI
The front desk for a model provider's services (OpenAI, Anthropic, etc.). You hand it a prompt, get back a completion. Most AI features in modern apps are calls to one of these.See: Day 10, Day 17
AI slopAI
Verbose, hedging, generic AI output that sounds correct but commits to nothing. Lots of 'you might consider' and 'depending on your needs.' The fix is to ask for a specific opinion: 'Pick one. If it's wrong, I'll tell you.'See: Day 18
AlertFoundation
A signal that some metric crossed a threshold. 'Error rate above 5% for ten minutes' is an alert. Alerts page a human; they should not page a room with nobody in it.See: Day 14
AnthropicAI
The company that makes Claude. One of the major model providers, alongside OpenAI and Google. See also LLM, GPT, Claude.
APIFoundation
Application Programming Interface. The front desk of a service. You hand it a request in the format it expects, and it either does what you asked or hands back a polite 'no' with a reason.See: Day 10
Authentication (AuthN)Foundation
Who you are. The login. The mechanism that proves a user is who they say they are (email + password, social sign-in, magic link).See: Day 13
Authorization (AuthZ)Foundation
What you are allowed to do. The permission rules. The mechanism that decides whether a logged-in user can perform a specific action.See: Day 13

B

BackendFoundation
The work that happens out of sight, on a server somewhere, in response to requests from the frontend. Holds the database, the secrets, and the heavy logic.See: Day 9
BackupFoundation
A copy of your data taken at a point in time. You should have one. You should test that you can restore from it. You probably don't.See: Day 11
Billing alertPractice
An automatic notification when your hosting or service costs cross a threshold you set. 60 seconds to configure. The difference between a $0 hobby and a $4,000 surprise.See: Day 12, Day 19
BranchFoundation
A parallel timeline in git. You can experiment on a branch and either merge it back into main (you liked the result) or throw it away (you didn't).See: Day 6
Breach notificationPractice
A legal obligation to tell affected users (and sometimes regulators) within a specific window, often 72 hours, when their data has been compromised.See: Day 20
Browser extensionPractice
A small program that runs inside your browser and modifies how web pages look or behave. A legitimate 'shape' for a build that just needs to tweak the web for you.See: Day 16
Build PlanPractice
Six fields, one page, no code. The graduation artifact of this course. Names what you're building, who it's for, the journey, the stack, where AI fits, and the smallest version.See: Day 18, Day 21

C

CCPAPractice
California Consumer Privacy Act. The California version of GDPR. Applies if your users are in California, which they probably are.See: Day 20
CDN (Content Delivery Network)Foundation
A layer in front of your hosting that caches static content (images, CSS) close to users around the world, so the trip to your stuff is shorter. Most hosting providers include one.See: Day 12
ChatGPTAI
OpenAI's consumer-facing chat product. The most famous LLM application. The whole thing is mostly a wrapper around their underlying GPT models.
ClaudeAI
Anthropic's AI assistant. The model this course cites by name when illustrating what AI tools ask you. See also Anthropic, LLM.
CLAUDE.mdAI
A project-level context file at the root of a Claude Code project. Tells the model what we're building, the conventions, the constraints, what to avoid. The single most underrated AI productivity move. See also Context engineering.See: Day 18
ClientFoundation
The local thing in a request: usually your phone or browser. The thing that starts the request and renders the response.See: Day 3, Day 4
CodebaseFoundation
A folder of files containing the text of a program, along with whatever supporting things it needs to run. Also called a repo or repository.See: Day 6
CohortPractice
A group of students taking a course together at the same time. The opposite of self-paced. Technical Enough is self-paced; cohort offerings are a possible future option.
Cold startFoundation
The delay when a serverless function hasn't run in a while and the provider has to start it up. Usually a few hundred milliseconds; occasionally noticeable.See: Day 12
CommitFoundation
A snapshot of the codebase at a moment in time, with a short note explaining what changed. Git remembers every commit forever, which is the whole point.See: Day 6
ComponentFoundation
A reusable piece of UI. A button is a component. A navigation bar is a component. Modern frontends are built out of components stacked together.See: Day 8
ContainerFoundation
Your code packaged with everything it needs (operating system, dependencies, configuration). You hand the package to a provider who runs it on computers they manage. More control than serverless, more responsibility.See: Day 12
Context engineeringAI
The discipline of structuring the information you give an AI so it can produce useful output. More important than prompt engineering. Three layers: project-level context (CLAUDE.md and equivalents), task-level context (relevant files and examples), living context (updates as you learn what works).See: Day 18
Context fileAI
Any project-level context document an AI tool reads to understand your project. CLAUDE.md (Claude), .cursorrules (Cursor), CONVENTIONS.md (Aider). The fix for most 'the AI keeps doing it wrong' problems.See: Day 18
Context windowAI
The amount of text an LLM can hold in its working memory for a single response. Measured in tokens. When you give the model more than fits, the oldest stuff falls out.
CopilotAI
GitHub's AI coding assistant. Suggests code as you type. One of the early Role 1 (writes the code) AI products. See also Cursor.
Cron jobFoundation
A piece of backend code that runs on a schedule rather than in response to a request. 'Every night at 2 a.m., send the daily digest email' is a cron job.See: Day 9
CSS (Cascading Style Sheets)Foundation
The appearance of a web page. The clothes on the HTML skeleton. Colors, layout, type, animation.See: Day 8
CursorAI
An AI-native code editor. Like VS Code but with the AI plugged into the editing flow. One of the most common Role 1 (writes the code) tools.

D

Data residencyPractice
The legal requirement that certain data stay in certain countries. EU user data, for example, often has to live on servers in the EU.See: Day 20
Data warehouseFoundation
A database optimized for analytics rather than for serving an app. Snowflake, BigQuery, Redshift. You probably do not need one until you have a real analytics question.See: Day 11
DatabaseFoundation
A purpose-built service for storing and querying structured data. Postgres, MySQL, MongoDB are all databases. They differ in shape; mostly they all do the same job.See: Day 11
DeployFoundation
Push a copy of your code from your laptop (or CI system) to the hosting provider, so it starts serving real users. The act of going from 'works locally' to 'live'.See: Day 7, Day 12
Development environmentFoundation
Where you write code. Usually your laptop. A sandbox where things break on purpose. See also Environment, Staging, Production.See: Day 7
DisclaimerPractice
A document explaining what your product is not responsible for. Important for anything that touches health, finance, or legal, where users might mistake your output for professional advice.See: Day 20
Document databaseFoundation
A database where each piece of data is its own JSON-shaped object with no enforced structure. MongoDB, DynamoDB, Firestore. Compare with relational.See: Day 11

E

EmbeddingAI
A numeric fingerprint of a piece of text (or image) that captures its meaning. Used to compare semantic similarity. The foundation of RAG and most search-with-AI features.
Encryption (at rest, in transit)Practice
Scrambling data so only authorized parties can read it. 'At rest' means stored encrypted on disk. 'In transit' means encrypted while moving over the network (HTTPS). Most providers do both by default; don't turn them off.See: Day 19
EndpointFoundation
A specific door at the server that frontend requests are allowed to knock on. 'Like a photo' is one endpoint; 'comment on a photo' is a different one.See: Day 10
EnvironmentFoundation
The specific setup where a piece of code happens to be running. Operating system, tool versions, configuration values, available memory. Different environments are why 'it works on my machine' is a sentence.See: Day 7
Environment variableFoundation
A configuration value the code reads at runtime instead of having hard-coded. Secrets (API keys, passwords) live here, not in the codebase.See: Day 7, Day 19
Error trackingFoundation
A tool (Sentry, Rollbar, Bugsnag) that captures errors as they happen, groups them, and surfaces them ranked by frequency and severity. Free Sentry account, an hour of setup, peace of mind.See: Day 14

F

Feature flagPractice
A switch in your code that lets you turn a feature on or off without redeploying. Useful for kill switches, gradual rollouts, and A/B tests.See: Day 19
Fine-tuningAI
Further training an existing LLM on a specific dataset so it gets better at a narrow task. Almost always overkill for non-engineer builds; prompt engineering and RAG cover most cases.
FrameworkFoundation
A pre-baked set of decisions someone else made for you. Picks the structure, the conventions, the libraries, writes the boilerplate. You fill in the parts specific to your idea.See: Day 5
FrontendFoundation
Everything users see and touch. The buttons, the colors, the layout, the animations. Built from HTML (structure), CSS (appearance), JavaScript (behavior).See: Day 8

G

GDPRPractice
General Data Protection Regulation. The EU's privacy law. Applies to anyone with EU users (which is most apps). Includes the rights to know, delete, and be notified of breaches.See: Day 20
GitFoundation
Undo for an entire project. Remembers every change ever made to every file in the codebase. The reason engineers can be brave with experiments.See: Day 6
GitHubFoundation
A website (owned by Microsoft) where people store their git histories in the cloud and collaborate. The most common host for open and private codebases.See: Day 6
GPT (Generative Pre-trained Transformer)AI
The family of models OpenAI is best known for. 'GPT-4', 'GPT-5' refer to specific versions. Often used loosely to mean any LLM.
GraphQLFoundation
An alternative to REST for designing APIs. Lets the client ask for exactly the data it wants in one request. Less common than REST in 2026; pick REST unless you have a reason.See: Day 10
GuardrailPractice
An operational discipline that experienced builders just know, that nobody tells beginners until they have been burned. Day 19 covers the five most important ones.See: Day 19

H

HallucinationAI
When an LLM confidently produces something false. Not a bug exactly; a consequence of how the model works. The reason you do not let AI be the final word on anything high-stakes.
HostingFoundation
Renting a computer connected to the internet to run your backend on. The four shapes: serverless, managed service, container, virtual machine.See: Day 12
HTML (HyperText Markup Language)Foundation
The structure of a web page. The skeleton. 'Here is a heading, here is a paragraph, here is a button, here is an image.'See: Day 8
HTTPSFoundation
HTTP with encryption. The 'S' is for secure. The default for the entire web in 2026; if a site is not on HTTPS, treat it as broken.

I

Index (database)Foundation
A behind-the-scenes structure the database keeps to make certain queries fast. Forgetting to add an index is one of the most common reasons a fast app becomes slow at scale.See: Day 11
Iteration loopAI
The back-and-forth between you and an AI tool to converge on a working output. Productive iteration follows three habits: specific feedback (not vague), verify before trust, and the five-attempt kill rule.See: Day 18

J

JavaScriptFoundation
The language that runs in the browser. Also runs on the backend (via Node.js). The most widely deployed programming language on earth.See: Day 5, Day 8

K

Kill rule (AI iteration)AI
If you have asked an AI the same thing five times and the output is still wrong, the AI is not going to get there in attempt six. Stop. Reset the context, reframe the task, or do it yourself.See: Day 18
Kill switchPractice
A documented way to turn the whole product off in under five minutes when something has gone wrong you cannot fix in real time. A feature flag, an env var, or a button at the hosting provider.See: Day 19

L

Language (programming)Foundation
A set of words and grammar rules the computer agrees to take seriously. Python, TypeScript, Ruby, Go, etc. Differ from each other the way English differs from German: same job, different costumes.See: Day 5
LibraryFoundation
A single package of code that does one specific thing (parse a date, send an email, encrypt a password). You use many libraries inside a framework.See: Day 5
LLM (Large Language Model)AI
A statistical model trained on a huge amount of text that predicts the next token. The thing under the hood of ChatGPT, Claude, Copilot, and most 'AI' products.
LocalFoundation
Software that runs on the physical device in front of you. Yank the Wi-Fi, it keeps working. Compare with Remote.See: Day 3
LogFoundation
A timestamped line of text recording something that happened: a request, an error, an action. You read logs when something has gone wrong and you want to figure out why.See: Day 14

M

Managed serviceFoundation
A hosting shape where the provider runs your code on always-on computers and gives you a URL. You pay a steady monthly amount and rarely think about scaling. Heroku, Render, Fly.io, Railway.See: Day 12
MetricFoundation
A number tracked over time. Request count per minute, average response time, error rate. You watch metrics on dashboards and set alerts on them.See: Day 14
MFA (Multi-Factor Authentication)Practice
The 'enter the code from your authenticator app' step at login, in addition to your password. Modern auth services give it to you free; turn it on.See: Day 13
MicroserviceFoundation
A backend split into many small backends, each responsible for one thing. Useful at large scale, overkill for almost everything else. Compare with Monolith.See: Day 9
Migration (database)Foundation
A change to the database schema applied in a controlled way. 'Add a verified column to the users table.'See: Day 11
Model (AI)AI
The trained statistical artifact at the heart of an AI system. GPT-5, Claude Opus 4.7, Llama 3 are all models. When AI providers release a 'new model,' this is what they mean.
MonitoringFoundation
How you find out your software is broken before your users tell you. Logs (detail), metrics (health), error tracking (what specifically broke).See: Day 14
MonolithFoundation
A backend that is all one big program. Less fashionable than it used to be; still the right answer for most early-stage builds.See: Day 9
MVP (Minimum Viable Product)Practice
The smallest version of your idea that would be useful to someone on day one. Usually narrower than your first instinct. The 'smallest thing' field in the Build Plan.See: Day 15, Day 21

O

OAuthFoundation
The protocol that powers 'sign in with Google' and similar. You do not need to understand it; you need to know that when AI mentions OAuth, it is talking about social logins.See: Day 13
ObservabilityFoundation
The umbrella term for 'having enough signal to figure out what is happening in production.' Logs plus metrics plus traces add up to observability.See: Day 14
On-callFoundation
The arrangement where one engineer at a time is responsible for responding to alerts. Healthy teams rotate this. Unhealthy teams have one person silently doing it forever.See: Day 14
OpenAIAI
The company that makes GPT and ChatGPT. One of the major model providers. See also Anthropic, GPT, ChatGPT.
Operating SystemFoundation
The thing your laptop or phone runs (macOS, Windows, Linux, iOS, Android). Your code runs on the OS, possibly inside a framework, possibly using a bunch of libraries.See: Day 5

P

Personal data (PII)Practice
Information about an identifiable person: name, email, phone, location, IP. The moment you store any, real obligations attach (GDPR, CCPA, etc.). Don't collect what you don't need.See: Day 11, Day 19, Day 20
Postgres (PostgreSQL)Foundation
An excellent, free, open-source relational database. The default answer for 'where should this data live' on most modern builds. Available hosted from Supabase, Neon, Render, AWS RDS.See: Day 11
Privacy PolicyPractice
A required document explaining what data you collect, why, how long you keep it, and who you share it with. Be specific and honest; boilerplate that lies is worse than nothing.See: Day 20
ProductionFoundation
The live environment that real users see. Where breakage costs money and reputation. The opposite of development and staging.See: Day 7
PromptAI
What you type to an AI model: the instruction or question. The 'recipe' you give the model. Everything above the prompt is what every AI course teaches; this course is what's below.
Prompt engineeringAI
The craft of writing prompts that reliably produce the output you want. Real skill; smaller than the AI ecosystem makes it sound. Largely solved by being specific and showing examples.
Pull (git)Foundation
Give me the latest commits from the team. The reverse of push.See: Day 6
Push (git)Foundation
Send my commits up to the team. The reverse of pull.See: Day 6

Q

Query (database)Foundation
A request to the database. 'Give me all users who signed up in the last week.' Written in SQL for relational databases, in other dialects for document databases.See: Day 11

R

RAG (Retrieval-Augmented Generation)AI
Pattern where you fetch relevant context from your data first, then send it to the LLM along with the user's question. How most production AI features stay grounded in real, current information.
Rate limitFoundation
The maximum number of requests an API will accept from you per unit time. Day 19 returns to this; every endpoint of your own backend should also have one.See: Day 10, Day 19
RBAC (Role-Based Access Control)Foundation
A common pattern for authorization where users have roles (admin, member, guest) and each role has a set of permissions.See: Day 13
RegionFoundation
Which physical data center your code runs in. AWS us-east-1 is in Virginia. Pick a region close to your users; pick one with data-residency requirements in mind.See: Day 12
Relational databaseFoundation
A database where data lives in tables (rows and columns) with a strict schema. Postgres, MySQL, SQLite. The default for most new builds. Compare with Document.See: Day 11
RemoteFoundation
Software that runs on a different physical device somewhere else, reached over the internet. Yank the Wi-Fi, it stops working. Compare with Local.See: Day 3
Repository (repo)Foundation
Same as codebase. A folder of code tracked by git, usually hosted on GitHub.See: Day 6
RequestFoundation
The note going from client to server. 'Hi, I'd like to like this photo.' Compare with Response.See: Day 4, Day 10
ResponseFoundation
The note coming back from server to client, in reply to a request. 'Done.'See: Day 4, Day 10
Responsive designFoundation
A frontend that behaves correctly across different screen sizes. Rearranges itself on phones, tablets, laptops, and desktops without anyone designing four versions by hand.See: Day 8
RESTFoundation
A specific style of API where each thing the service can do has its own door (endpoint), and the doors are named like web pages. The default API style in 2026.See: Day 10

S

ScaleFoundation
How well a system handles more traffic. Scales horizontally means adding more copies of the server. Scales vertically means making the existing server bigger.See: Day 12
SchemaFoundation
The structure of your data. 'Users have a name, an email, a created-at timestamp, and an avatar URL' is a schema.See: Day 11
ScriptPractice
A small piece of code that runs once and does one specific thing. Often a perfectly good shape for a build that doesn't need a UI.See: Day 16
ServerFoundation
The remote thing in a request: a computer in a data center that processes requests and sends back responses.See: Day 3, Day 4, Day 9
ServerlessFoundation
A way of renting computer time on the internet where you pay only when your code is actually running, not all the time. The name is a lie; there are servers, you just don't think about them.See: Day 12
Server-side rendering (SSR)Foundation
When the server prepares the actual HTML before sending it, instead of letting the user's device assemble it. Good for speed and SEO, more complex.See: Day 9
SessionFoundation
A logged-in browser tab. The mechanism that remembers you are signed in between requests.See: Day 13
SLO (Service Level Objective)Foundation
A specific target for system health, like '99.9% of requests should succeed.' Useful for grown-up products; overkill for most v1 builds.See: Day 14
SoftwareFoundation
A set of instructions, that runs somewhere, that takes an input and produces an output. Every other thing in this course is a question about one of those three parts.See: Day 2
Stack (tech stack)Practice
The combined set of tools you've picked for your build: frontend framework, backend framework, database, hosting provider, auth service, monitoring. The fourth field of the Build Plan.See: Day 18
StagingFoundation
A copy of production that only the team can see, where new features get tested under realistic conditions before they ship. The middle environment between development and production.See: Day 7
Status codeFoundation
A three-digit number that summarizes how an API request went. 200 means fine, 404 means the thing doesn't exist, 401 means not allowed in, 500 means the server had a stroke.See: Day 10
Streaming responseAI
When an AI model returns its answer one token at a time as it generates, rather than waiting until the whole thing is done. Why ChatGPT appears to type at you.

T

Terms of ServicePractice
The contract between you and the user. What they can do, what you can do, what happens if either breaks the deal. A standard document every public-facing build needs.See: Day 20
Token (auth)Foundation
A small piece of cryptographic proof that a user is who they say they are. Sessions are usually backed by tokens.See: Day 13
Token (LLM)AI
The chunk size an LLM thinks in. Roughly a few characters or a short word. You're billed per token; the context window is measured in tokens. Not the same as an auth token.
TracingFoundation
A more advanced version of logs where each request gets a unique ID and you can follow it through every system it touched. Useful at scale; overkill for v1.See: Day 14
TypeScriptFoundation
JavaScript with type annotations bolted on. The default language for new web apps in 2026 because catching type mistakes at write-time beats catching them at run-time.See: Day 5

U

UI (User Interface)Foundation
The surface of the screen. The colors, the buttons, the typography, the spacing. What an Instagram screenshot of your product would look like.See: Day 8
UX (User Experience)Foundation
What it actually feels like to use the thing over time. The journey from sign-up to first value. Where users get stuck. Related to UI but a different job.See: Day 8

V

Vector databaseAI
A database optimized for storing embeddings and searching them by similarity. The storage layer behind most RAG systems. Pinecone, Weaviate, pgvector (Postgres extension).
Version controlFoundation
The practice of tracking every change made to a codebase over time. In 2026, this means git (or something git-shaped). Nothing modern works without it.See: Day 6
VM (Virtual Machine)Foundation
A slice of a physical computer rented out with a full operating system you control. You install your stuff, you manage updates and security. Most flexibility, most responsibility.See: Day 12

W

Web appPractice
Software that lives at a URL and runs in a browser. The right default shape for most builds in 2026. Cheap to host, easy to distribute, updates instantly.See: Day 16
WebhookFoundation
The reverse direction of an API call. Instead of you knocking on the service's door, the service knocks on yours when something happens. 'Hey, that charge you set up just succeeded.'See: Day 10

Missing a term you expected to find? It probably belongs here. The glossary is a living document and grows with the course.

← Back to the map