Chief A.I., Oh! / AI Basics Blueprint v1.0
A start-anywhere, no-prerequisites guide to the basics of AI. By the time you finish, you can navigate the three flagship LLMs, pick the right one for the job, and use the novice-to-intermediate features each one ships with. No coding required.
00 · What you get
Concrete capabilities, not vibes. Each one is taught in the tier that owns it.
Use words like model, token, context, prompt, hallucination, agent, and embedding without bluffing.
Turn a vague need into a prompt that gets you something useful on the first try, then iterate without starting over.
Know when to reach for ChatGPT vs Claude vs Gemini. Stop using whichever one is open in your browser by default.
Projects, Custom GPTs, Gems, memory, file uploads, voice mode, image gen. Confidently, in the right one.
Recognize where image gen, voice, transcription, agents, and coding tools fit, so a new product launch doesn't disorient you.
01 · Wayfinder
Pick the card that sounds like you. Each one drops you into the right tier. Nothing in this guide assumes you know anything yet.
02 · The scale
Each tier assumes you understand the one below it. If a tier feels easy, skim it; if a tier feels hard, the tier below it is the gap. Don't skip down without going back.
AI vs ML vs LLM vs generative AI. Training vs inference. Tokens, context, parameters, hallucinations, model families. The mental model under everything else.
System vs user prompts. Role, goal, constraints, format, examples. The Funnel Method. Iteration patterns. The three failure modes that ruin most prompts.
Who makes each. What each is best at. Where each lives. What each costs at the consumer tier. A decision matrix so you stop defaulting to whichever tab is already open.
The intermediate features inside each LLM that let you stop re-explaining yourself. Where Projects beats memory. When a Custom GPT is overkill. How Gems plug into Google Workspace.
Image generation, voice, transcription, agents, coding assistants. Not so you build all of them, just so you know what they are and which ones might earn a place in your week.
Every bolded term across the tiers is defined in the glossary. If you hit a word you don't know, jump to the glossary, search it, then come back. Treat it as your reference librarian, not a chapter you read once.
03 · How to use this guide
Don't try to memorize anything. The goal of the first read is to know what's in here, not to master it. Most things will half-stick. That's fine.
Open ChatGPT, Claude, or Gemini in a second tab and actually try every example as you go through Tiers 2 and 3. Reading without doing is the most common reason people stay beginners.
Use the techniques from Tier 2 every day for a week. Then sit in Tier 4 for a week. Skill comes from concentrated repetition, not breadth.
You're going to forget some terms. That's not a failure, it's the design. The glossary exists so you can recover the word without re-reading the whole tier.
04 · One opinion before you start
The biggest mistake new AI users make is treating an LLM like Google. Google answers a question. An LLM does a job. If your prompt looks like a search query, you're going to get a search-query-shaped answer. From the moment you finish Tier 1, write prompts that read like a one-paragraph email to a smart, fast intern who has never met you. Role, goal, what good looks like, what to avoid. That single habit is worth more than memorizing any feature list.