1. Most people think AI is either magic or science fiction. It is neither.

It is a tool that does specific things very well, and other things badly. Knowing which is which is most of the skill.

AI gets oversold and undersold at the same time. The marketing makes it sound omniscient. The skeptics make it sound useless. Both are wrong. The reality is more useful and more interesting.

• Magic camp: thinks AI knows everything, will replace all jobs, is about to wake up.
• Skeptic camp: dismisses AI as autocomplete that hallucinates and is overhyped.
• Reality: very capable assistant for specific tasks. Predictable strengths and weaknesses.
• Most adults: never get past the first ten minutes of trying. Quit too early.
• Today's job: replace both myths with a working mental model.

2. AI is a very capable assistant that has read most of the internet.

It predicts what useful answer comes next based on patterns in language. That is the entire trick.

Modern AI assistants like ChatGPT, Claude, and Gemini are trained on huge amounts of text from books, websites, and conversations. They learn patterns: what kind of answer follows what kind of question. When you ask it something, it predicts the most useful response, one word at a time.

• Trained on books, websites, code, conversations — most of the public internet.
• Predicts useful answers one word at a time, based on what came before.
• Not a database — it does not look things up; it generates from learned patterns.
• Not a person — no opinions of its own, no memory between conversations by default.
• Not magic — predictable behavior once you understand the basics.

3. From research labs to your phone in five years.

The brief history. AI as we know it today is genuinely new. The breakthrough paper, called Attention is All You Need, came out of Google in twenty seventeen. It introduced the Transformer architecture — the foundation of every AI assistant you will use. Twenty twenty: OpenAI released GPT-three, which was the first AI to demonstrate that scaling — making the model bigger — produced qualitatively new abilities. Researchers were genuinely shocked at what it could do. Twenty twenty-two: ChatGPT launched and became the fastest-growing consumer product in history. AI entered mainstream consciousness. Twenty twenty-three and twenty twenty-four: competitors emerged — Claude from Anthropic, Gemini from Google, Copilot from Microsoft, Perplexity for search. The ecosystem matured. Twenty twenty-five and twenty twenty-six: AI got integrated into the tools you already use — Microsoft Office, Google Workspace, Apple Intelligence, your iPhone keyboard, your inbox, your calendar. The point of this history is that you are learning AI at exactly the right moment. Five years ago this technology did not exist in a useful form. Five years from now it will be assumed background knowledge for every adult. Right now, basic competence is rare and valuable. Learning it now puts you years ahead of the average adult.

• 2017 — Transformer architecture paper. The breakthrough that made all this possible.
• 2020 — GPT-3 demonstrates real capability. Researchers shocked.
• 2022 — ChatGPT launches. AI enters mainstream consciousness.
• 2023-2024 — Claude, Gemini, Copilot, Perplexity. Multi-tool ecosystem.
• 2025-2026 — AI integrated into Microsoft Office, Google Workspace, daily apps.

4. The basic loop: you type, it responds, you iterate.

Here is how a session actually works. You type a question or a request — that is called a prompt. The AI reads it and generates a response, predicting useful text one word at a time. The response shows up on your screen, usually within a few seconds. You read it. If it is exactly what you wanted, you copy it and use it. If not — and most of the time it is not perfect on the first try — you tell the AI what to change. Make it shorter. Make it warmer. Make it more professional. Add a specific detail. Try a different angle. The AI takes your feedback and gives you a revised version. You iterate until you are happy. The single biggest mistake beginners make is treating the first response as the final answer. The first response is a draft. The skill is in the second, third, and fourth turn — the iteration. Treating AI like a slot machine where you put in a prompt and hope for gold is what produces frustration. Treating it like a writing partner who responds quickly to feedback is what produces real value. Throughout this course, we will keep coming back to this point. Iteration is the core skill of using AI well.

• You type a question or request (called a prompt).
• The AI generates a response, predicting useful text one word at a time.
• You read it. If it is right, great. If not, you ask for changes.
• Iteration is the core skill: shorter, warmer, simpler, more specific.
• Most beginners stop after the first response. The skill is in the second and third turn.

5. What a real AI session looks like.

A complete back-and-forth, annotated. This is what good usage looks like.

Let me walk you through what a real first session looks like. You type: help me write a polite email asking my landlord to fix the heater. The AI generates a complete draft email — subject line, greeting, body, sign-off. You read it. It is fine, but a little long and a little soft. You type: make it shorter and more direct. I am frustrated but professional. The AI sends back a tighter version that is firmer in tone and about half the length. You read it. This time it nails it. You copy and paste into your email client, fill in the bracketed placeholders, and send. Total time: under two minutes. Compare that to writing the email from scratch yourself — probably ten to fifteen minutes if you wanted it to land right. You just saved over ten minutes on one task. Now multiply by every email you write this week. That is the rhythm of good AI use. Quick first draft, one or two iterations, copy and ship. Most of this course teaches you to make this rhythm habitual across every kind of writing and thinking task you do.

6. Five things AI is genuinely great at.

Five things AI does genuinely well. One: drafting. Turning a vague idea into a polished first version. Email, blog post, proposal, cover letter, message — give it the basic information and a goal, get back a complete draft you can edit. Two: summarizing. Long emails, long documents, long meetings — AI extracts the key points, decisions, action items, deadlines. Saves enormous reading time. Three: explaining. Any concept, at any level. Explain compound interest like I am twelve. Explain it like I have a finance degree. Use a sports analogy. Use a cooking analogy. The same concept, infinitely re-explainable until it clicks. Four: rewriting. The same content in different tones, lengths, audiences. Make this email warmer. Make this paragraph shorter. Translate this into language a customer would understand. Adapt this for LinkedIn versus Twitter versus an internal memo. Five: brainstorming. Twenty options when you cannot think of one. Twenty subject lines. Twenty headlines. Twenty names for a project. Twenty creative angles for a presentation. AI is excellent at quantity when you need optionality. These five capabilities, used routinely, change how much you can get done in a day. The rest of the course teaches you to use them habitually.

• Drafting — turning a vague idea into a polished first version.
• Summarizing — long text → key points, action items, decisions.
• Explaining — concepts at any level, with examples and analogies.
• Rewriting — same content in a different tone, length, or audience.
• Brainstorming — twenty options when you cannot think of one.

7. Sarah, a real estate agent, before and after AI.

A composite portrait of what a few weeks of practice produces. Real numbers, real time saved.

Let me show you what is genuinely possible. Sarah is a real estate agent — composite portrait drawn from actual adults learning AI. Before: she spends about three hours a day on writing-heavy tasks. Listing descriptions. Client follow-up emails. Showing summaries. Property comparison emails. Buyer-seller communications. Three hours a day, every day. Week one of learning AI: she covers the basics — prompting, iteration, asking for options. By the end of week one, she drafts five listing descriptions in twenty minutes that used to take her two hours. Week two: she discovers AI for client follow-up emails. She drafts a template, AI customizes per client, she reviews and sends. Saves about an hour a day. Week four: she has a personal prompt library — twenty saved prompts for the situations she runs into routinely. She uses AI for offer drafts, contract review (with a human attorney still doing legal review), market summaries, neighborhood research. After four weeks: writing tasks take about an hour a day instead of three. Ten hours per week reclaimed. Annually: about five hundred hours saved. She uses the time for actual client meetings and brings in about two extra closings per month. The investment was a few hours of structured learning. The return was substantial professional and financial benefit. This pattern repeats across professions.

8. AI is not free. It is just much cheaper than the alternative.

Honest tradeoffs. AI is genuinely useful but it is not free. You pay in four ways. One: attention. Every output has to be reviewed before you send it, sign it, or commit to it. AI sounds confident even when it is wrong. The cost of skipping review is occasionally embarrassing or worse. Plan to read everything AI gives you. Two: privacy. AI services see what you type. Most have reasonable policies, but you should not paste passwords, social security numbers, or genuinely confidential business secrets into AI assistants without understanding the privacy settings. We will cover this in detail in class thirty-nine. Three: subscription. The free versions are okay; the paid versions are dramatically better. Twenty dollars a month for ChatGPT Plus, Claude Pro, or Gemini Advanced gets you faster responses, longer outputs, better models, and useful features. For most adults, the subscription pays for itself in the first hour of saved time. Four: learning curve. The first five or ten hours of using AI feel awkward. You are figuring out how to talk to it, what it can and cannot do, what makes a good prompt. After about ten hours, things click and the tool starts feeling natural. The investment is small relative to the lifetime return. The takeaway: AI has real costs, but the time saved is dramatic, and the return on investment shows up within days for most adults.

• You pay with attention — review every output before sending.
• You pay with privacy — sensitive data needs careful handling.
• You pay with subscription — $20/month gets you the good versions.
• You pay with learning curve — first 5 hours are awkward; after that it clicks.
• But: time saved is dramatic. ROI in days for most adults.

9. AI vs. Google vs. just doing it yourself.

Three different tools, three different jobs. Google search: best for finding specific facts, websites, recent news, places, products. You ask Google a question, you get back a list of links to sources, you click through and read. AI assistants: best for synthesizing answers, drafting documents, explaining concepts, brainstorming, and rewriting. You ask AI a question, you get back a complete formed answer. The two complement each other. For current news, prices, dates, and verified facts — Google. For first drafts, summaries, and explanations — AI. Increasingly, tools like Perplexity combine both. Your own brain: best for deep judgment, real creativity, sensitive context, and final decisions. AI gives you the first ninety percent of a draft; your judgment shapes the final ten percent that makes it actually right for your specific situation. AI does not replace either Google or your brain. It is a third tool that complements both. The skilled adult uses all three: Google for facts, AI for drafts and synthesis, brain for judgment. Most adults right now use only Google and brain, which means everything takes longer than it needs to. Adding AI to the toolkit is the upgrade.

• Google: best for finding facts and websites. AI: best for synthesizing answers.
• Yourself: best for deep judgment + creativity. AI: best for first drafts + iteration.
• AI does not replace either. It complements both.
• Use AI where speed and volume matter. Use yourself where judgment matters.

10. Five common beginner mistakes — all easy to fix.

Five mistakes that catch every beginner. One: vague prompts. Write me an email gets you generic slop. Write me a polite email to my landlord asking him to fix the heater, mentioning that the issue started Tuesday and I work from home, gets you something useful. Specifics in equals quality out. Two: single-shot use. Accepting the first response and walking away frustrated. The first response is the draft. The second prompt is where the magic happens. Three: trusting facts blindly. AI sometimes invents things — names, dates, statistics, sources. Most of the time it is right; occasionally it is wrong. Verify anything that matters, especially legal, medical, financial, or factual claims that will affect important decisions. Four: stopping too soon. Most beginners try AI for ten minutes, get a mediocre result on a vague prompt, conclude it is not for them, and quit. The skill develops in the next two hours. Push past minute ten. Five: privacy mistakes. Pasting passwords, social security numbers, or confidential business data into AI without thinking. The AI service sees everything you type. Treat it like a competent stranger who will read your message — that mostly tells you what is okay to share and what is not. We will cover all five of these in detail across the course. The good news: every one of these mistakes is easy to fix once you know it.

• Vague prompts — "write me an email" gets generic slop. Specific prompts get good output.
• Single-shot — accepting the first response. Iteration is the skill.
• Trusting facts blindly — AI sometimes invents things. Verify the important ones.
• Stopping too soon — quitting after a bad first try. Most adults give up at minute 10.
• Privacy mistakes — pasting passwords or sensitive data without thinking.

11. The five AI tools you should actually try.

Five tools to know. ChatGPT from OpenAI: the most popular AI assistant. Strong general capability across writing, analysis, brainstorming, learning. Free version works for casual use; ChatGPT Plus at twenty dollars a month is the version most adults use. Claude from Anthropic: my favorite for long-form writing, nuanced tasks, and conversations that need careful reasoning. Excellent at picking up on tone and intent. Free version available; Claude Pro at twenty dollars a month for serious use. Gemini from Google: integrated with Gmail, Google Docs, Sheets, and Drive. If your work or personal life lives in Google Workspace, Gemini meets you where you already are. Microsoft Copilot: integrated with Word, Outlook, Excel, Teams. If you live in Microsoft Office at work, Copilot is right there in your existing apps. Often comes with your work account at no extra cost. Perplexity: AI-powered search with cited sources. Best for research questions where you want both an answer and the sources that back it up. My recommendation for getting started: pick one of ChatGPT or Claude as your primary, learn it thoroughly, then add the others as you need them. Trying to learn five tools at once produces shallow knowledge of all five.

• ChatGPT (OpenAI) — most popular. Strong all-rounder.
• Claude (Anthropic) — best at long writing + nuanced tasks.
• Gemini (Google) — integrated with Gmail, Docs, Drive.
• Microsoft Copilot — integrated with Word, Outlook, Excel, Teams.
• Perplexity — AI-powered search with sources.

12. Right now: open ChatGPT or Claude. Type one prompt.

The single most valuable five minutes of this entire course is the first time you actually try it.

Concrete plan. Right now, before you watch the next class, open chatgpt dot com or claude dot ai in a browser. Free signup, no credit card needed for the basic versions. Once you are in, type this exact prompt: explain artificial intelligence to me like I am a complete beginner, using examples from email and work. Hit enter. Read the response that comes back. Notice three things. One: it is complete sentences in proper structure, not raw data. Two: it actually makes sense and uses examples that are relevant to you. Three: it adapted to your specific request — beginner level, email and work examples. Now type a follow-up: make it shorter and use more analogies. Read the second response. Notice that it changed based on your feedback. Just doing these two prompts gives you the core experience of AI. You typed something. It responded. You asked for changes. It iterated. That is the entire pattern, and you just learned it. The rest of this course is variations on this same theme applied to email, writing, productivity, learning, and a dozen other domains. Welcome to the course.

• Open chatgpt.com or claude.ai in a browser. Free signup, no credit card.
• Type: "Explain artificial intelligence to me like I am a complete beginner. Use email and work examples."
• Read the response. Notice: it is complete sentences, structured, and makes sense.
• Type: "Make it shorter and use more analogies."
• Read the second response. Notice: it adapted to your feedback.

13. AI is a capable assistant that predicts useful answers. Use it daily.

Five strengths, five weaknesses, five tools, five mistakes. The whole foundation.

Recap of class one. AI is a very capable assistant trained on most of the internet, predicting useful answers one word at a time based on patterns in language. Not magic, not autocomplete — somewhere genuinely useful in between. Five things AI does genuinely well: drafting, summarizing, explaining, rewriting, brainstorming. Five things it does poorly or unpredictably: current events past its training cutoff, exact math, sensitive judgment calls, knowing what it does not know, handling private data. Five tools to be aware of: ChatGPT, Claude, Gemini, Microsoft Copilot, Perplexity. Five common beginner mistakes, all easy to fix: vague prompts, accepting the first response without iterating, trusting facts without verification, quitting too soon, and privacy slips. This week, your only homework is to actually open ChatGPT or Claude and type one prompt. Five minutes. The single most valuable thing you can do for the rest of this course is to be hands-on rather than purely watching. Class two — the major AI tools beginners should know, when to use which one, and how to set up your free accounts. We will see you there.

• AI = very capable assistant trained on most of the internet, predicting useful answers.
• Five strengths: drafting, summarizing, explaining, rewriting, brainstorming.
• Five weaknesses: current events, exact math, sensitive judgment, knowing limits, privacy.
• Five tools: ChatGPT, Claude, Gemini, Copilot, Perplexity.
• Five mistakes: vague prompts, single-shot use, trusting blindly, quitting early, privacy slips.
• This week: ChatGPT or Claude open. One prompt. Five minutes. The course starts now.

Mastery quiz

1. Which of these is named in the 'five things AI is genuinely great at'?
   • Performing exact arithmetic calculations
   • Drafting a polished first version from a vague idea
   • Retrieving today's verified news and prices
   • Remembering you between separate conversations
2. In the Sarah example, roughly how much time did her writing tasks drop to after four weeks?
   • Three hours a day
   • About an hour a day
   • Zero — AI did it all
   • About four hours a day
3. According to the 'tradeoffs' section, which is one of the four ways you 'pay' for using AI?
   • Electricity bills from running it
   • Attention — every output must be reviewed before you rely on it
   • A one-time hardware purchase
   • Mandatory training certification
4. How does the lesson divide the jobs of Google, AI, and your own brain?
   • Google for drafts, AI for facts, brain for typing
   • Google for facts, AI for drafts and synthesis, brain for judgment
   • AI replaces both Google and your brain
   • Brain for facts, Google for judgment, AI for news