Every concept in 101, and the week it lands.

model vs product

Difference between a model (GPT, Claude, Gemini) and a product (ChatGPT, Cursor)

What AI is

lab landscape

Who the major labs/vendors are and their model families

What AI is

how llms work

Intuitive explanation: next-token prediction, training data, no memory between chats

What AI is

prompt response turn

Prompt, response, turn, conversation

What AI is

context window

Fixed amount of text the model can hold; what falls out

What AI is

hallucination

Confident, fluent, wrong; tone is not evidence

What AI is

training vs inference

How the model was made vs each use

What AI is

ai vs ml vs genai

AI vs machine learning vs generative AI definitions

What AI is

limits strengths

What models are good and bad at (math, counting, recent facts, private data)

What AI is

completion drive

Model is rewarded for finishing; declares done early (fail-mode framing)

What AI is

prompt basics

Context, one ask, output format, iterate

Prompting and directing

ask it to ask

Having the model interview you before acting

Prompting and directing

context management

What to paste, @-mention files, context as a budget

Prompting and directing

definition of done

Writing acceptance criteria / spec before asking the model to build

Prompting and directing

ai ide

AI code editor (Cursor, Windsurf, VS Code + Copilot)

Tools and workbench

chat vs agent mode

Chat answers vs agent edits files/runs commands; approve/reject

Tools and workbench

files folders project

A project is a folder of text files

Tools and workbench

never paste secrets

Keys in chat or repo are compromised

Secrets and security

idea validation

Finding and validating an idea; Mom Test, JTBD

Product and discovery

scope mvp

Smallest useful version; cutting scope

Product and discovery

build buy automate

Does a tool exist / would a prompt do / would a spreadsheet do before building

Product and discovery

privacy data retention

Chat retention, training on your data, enterprise settings

Responsible and safe use

verify output

Check the rendered result, not the chat; trust but verify

Verification mindset

reasoning models

Thinking/reasoning model tiers, when to use them

What AI is

multimodal

Text, image, audio, video in and out

What AI is

few shot examples

Giving examples in the prompt

Prompting and directing

chain of thought

Asking for step-by-step reasoning

Prompting and directing

iteration refinement

Refining through follow-ups; when to start a new chat

Prompting and directing

terminal basics

The command line exists; reading output; when to paste it

Tools and workbench

localhost vs deployed

Running on your machine vs on the internet

Tools and workbench

reading a diff

Reviewing what the agent changed

Tools and workbench

git commit push

Commit as save point; push to remote

Version control and the record

github repo

Repository, README, issues

Version control and the record

branch pr merge

Branch, pull request, review, merge

Version control and the record

preview deploys

A live URL per PR / branch

Version control and the record

issues as plan

Writing intent in an issue before building

Version control and the record

transcripts as record

Keeping AI conversations as evidence / audit material

Version control and the record

html css js named

HTML/CSS/JS exist; named not written

Web app anatomy

frameworks named

React/Next.js or equivalent as a kit

Web app anatomy

pages routes components

Pages, routes, components, navigation

Web app anatomy

forms inputs validation

Collecting and validating user input

Web app anatomy

state

Remembering things while the page is open

Web app anatomy

deploy hosting

Deploying to Vercel/Netlify/Firebase etc.

Ship and operate

code review skills

Reviewing AI-written code even if you cannot write it

Verification mindset

embeddings vectors

Embeddings, vector similarity, vector databases

What AI is

frontend backend db

Three-box model: browser, server, database

Web app anatomy

tables rows columns types

Relational basics

Data

keys relationships

Primary key, foreign key, relationship

Data

crud

Create, read, update, delete

Data

sql exists

SQL named; agent writes it; student reads it

Data

baas platform

Supabase / Firebase / Convex / Neon as the database product

Data

migrations

Schema changes as files

Data

local vs prod data

Test data vs real data; env parity

Data

pii data classification

What data is about a person / regulated; handling it differently

Data

authn vs authz

Who you are vs what you may do

Users and access

sessions magic link oauth

Login mechanisms

Users and access

protected routes

Pages that require login

Users and access

row level security

Database-enforced access policy

Users and access

threat question

What can the wrong person see/do; test written first

Users and access

red team by peer

Someone tries to break it before launch

Users and access

rag

Retrieval-augmented generation in plain terms

AI inside the product

fine tuning

Fine-tuning vs RAG vs prompting tradeoffs

AI inside the product

ci cd

Automated checks on every PR

Ship and operate

testing basics

Writing/running tests with the agent; watching one fail

Ship and operate

tokens

Unit of text the model processes and you pay for

What AI is

system prompt role

System prompt / persona / role instructions

Prompting and directing

structured output

Asking for JSON or a defined shape

Prompting and directing

reading errors logs

Reading error messages and logs instead of pasting blind

Tools and workbench

env vars dotenv

Environment variables, .env, .gitignore

Secrets and security

public vs secret keys

Anon/publishable vs service-role/secret keys

Secrets and security

key rotation

Replacing a leaked key

Secrets and security

secret scanning

Scanning repo/transcripts for secrets

Secrets and security

server vs client secrets

Which side of the wire may hold a key

Secrets and security

api request response

Endpoint, request, response, docs

Integrations and APIs

api keys rate limits

Credentials and limits

Integrations and APIs

email integration

Transactional email (Resend, SendGrid)

Integrations and APIs

llm api call

Calling a model from code: messages, params, cost

Integrations and APIs

spend caps cost

Setting provider spend limits; estimating cost

Integrations and APIs

text generation app

Generation features (summaries, drafts)

AI inside the product

agent definition

Model + tools + loop toward a goal

AI inside the product

automation vs workflow vs agent

Choosing the simplest mechanism

AI inside the product

human in the loop

Approval step before effect

AI inside the product

error monitoring

Sentry or equivalent

Ship and operate

maintenance longevity

Dependency updates, handoff docs, keeping it alive

Ship and operate

bias fairness

Bias in outputs and training data

Responsible and safe use

disclosure transparency

Telling people AI was used

Responsible and safe use

copyright ip

Ownership of generated content and training-data concerns

Responsible and safe use

when not to use ai

Tasks AI should not do; human judgment

Responsible and safe use

foundation model

Large general-purpose pretrained model

What AI is

responsive mobile

Works on a phone

Web app anatomy

accessibility

Usable without a mouse / with a screen reader

Web app anatomy

ux states

Loading, empty, error states

Web app anatomy

ui design basics

Layout, hierarchy, design systems/shadcn/Tailwind named

Web app anatomy

custom domain dns

Domain and DNS

Ship and operate

analytics

Pageviews, events, funnels

Ship and operate

usability testing

Watching a real person use it

Ship and operate

demo presentation

Presenting the product, not the conversation

Product and discovery

Waiting for 201

mcp tools for agents · backups recovery · owasp basics · prompt injection · ai app security · roles · webhooks · function tool calling · multi agent · agent memory · evals testing ai · observability tracing · scheduled jobs automation · user interviews · shadow ai policy · ai governance frameworks · receipts evidence · cross model review · fail modes catalog · catch log metacognition

Deliberately not taught

prompt templates library · cli coding agent · no code builders · file storage · test vs live mode · payments integration · third party auth oauth apps · chat app build · image generation · agent frameworks · open models local · voice agents · performance basics · seo basics · monetization business · environmental cost