From Get to Know Claude: Your AI Thinking Partner · Parmod K. Gandhi
By the end of this chapter, you will know what Claude is, where it came from, why it was built the way it was, and how to take your very first steps with it. No technical background required. All you need is a device with a web browser and a little curiosity.
For most of the history of computing, there was a frustrating gap at the heart of the relationship between humans and machines. If you wanted a computer to do something for you, you had to learn to speak its language. The computer was not going to meet you halfway.
That gap shaped the entire industry. It determined who could participate — mostly engineers and mathematicians — and who could not. And for decades it seemed immovable. Computers were powerful but alien, and the price of admission was fluency in a language no human had grown up speaking.
That is changing now. Faster than most people realize.
The author of this book started his career in the 1970s, writing code in Assembly and Fortran on DEC minicomputers, and Omron’s microcomputer. Those languages were about as close to raw machine instructions as you could get. Every line of code demanded that the human think in the computer’s terms, not their own.
The decades that followed brought better and better bridges — C, then object-oriented languages like C++ and Delphi, then the web languages, then Python and Ruby with their more readable syntax. Each generation moved programming a little closer to human language. But they were all still programming languages. You still had to learn them. The computer was still demanding that you meet it on its own ground.
What Claude and tools like it represent is something different in kind, not just degree. The programming language of the AI era is English. Or French. Or whatever language you grew up speaking. If you can read and write, you can already talk to Claude. Not approximately. Not with training wheels. Actually.
For fifty years we asked humans to learn computer languages. Now computers are learning human languages. The direction of the burden has reversed — and that changes everything about who gets to use these tools and what they can accomplish with them.
There are many AI tools available today. ChatGPT from OpenAI. Gemini from Google. Copilot from Microsoft. Grok from xAI. All of them are capable. So why does this book focus on Claude?
The short answer is trust. Not just capability — trust. And the story of why Claude earns that trust is one of the most remarkable in the brief history of artificial intelligence.
When a journalist visited Anthropic’s San Francisco headquarters in early 2026, he noticed something sitting on the common-area coffee tables: copies of The Making of the Atomic Bomb by Richard Rhodes — the Pulitzer Prize-winning account of the physicists who built the first nuclear weapon.
The book was not there by accident. Anthropic’s CEO Dario Amodei recommends it constantly. It is the lens through which he thinks about the work his company is doing. The parallel is intentional and uncomfortable. The scientists who built the atomic bomb were the best physicists in the world. They solved the hardest problem in physics. And then they watched the solution leave the laboratory. The destination was someone else’s to decide.
Amodei and the founders of Anthropic left OpenAI in 2021 precisely because they believed the AI industry was heading toward its own Oppenheimer moment — and that someone needed to be in the room who was thinking seriously about what happens when the technology leaves the laboratory.
In 2025, Anthropic signed a contract with the United States Department of Defense worth up to $200 million. The contract included two restrictions: Claude would not be used for mass domestic surveillance of American citizens, and it would not be used to power fully autonomous weapons systems — AI making lethal decisions without a human in the loop.
In early 2026, the Pentagon demanded Anthropic remove those restrictions. Weeks of pressure followed. The threatened penalties included cancellation of the contract, a national security designation normally reserved for foreign adversaries, and potential government action to compel compliance.
Anthropic’s response came from Dario Amodei directly: the company could not in good conscience accede to the request. The deadline passed without agreement. The contract was terminated.
Most companies would have found a way to say yes. Anthropic did not. That is not a small thing.
Boris Cherny is the lead engineer who built Claude Code — Anthropic’s powerful command-line AI tool. In mid-2025, he was recruited away by a well-funded competitor. He accepted the offer. Two weeks later, he was back at Anthropic.
When asked why, he pointed to one thing: Anthropic’s mission. Safety. The kind of work that, once you have left it, makes it hard to stay away. He took a seniority cut to return. His co-lead came back with him.
The best engineers in AI have choices. When they choose to come back, the reason matters.
Most people’s first instinct is to use Claude like Google. Type a question. Get an answer. That works — but it undersells what is actually happening.
A search engine finds pages that might contain your answer and leaves you to figure out which one applies. Claude reads what you wrote, understands your specific situation, and composes a response directly for you. The answer is not fetched from somewhere. It is written, for you, right now.
The difference becomes clear the moment something goes wrong. If a search result does not answer your question, you search again with different words, hoping to get lucky. If Claude’s first answer misses the mark, you tell it so. The conversation continues. It adjusts.
I searched online for why my tomato plants keep wilting in the afternoon even though I water them every day. I got a lot of general articles but none of them matched my situation. My plants are in pots on a south-facing concrete patio in Ontario.
That combination of clues points pretty clearly to heat stress rather than drought. South-facing concrete acts like a radiant heater in the afternoon — the surface temperature can reach 50°C or more on a hot day, and pot soil heats up much faster than ground soil. A few things that help: move the pots back from the concrete edge if you can, water in the early morning rather than midday, and consider putting a light-coloured saucer under each pot to reflect heat. If the wilting recovers by evening, that confirms it is heat stress rather than a watering problem.
Notice what happened there. Claude did not return a list of articles about tomato care. It read the specific details — pots, south-facing, concrete, Ontario — and reasoned about the actual situation. Then it asked a follow-up question, because that is how useful conversations work.
Claude is available in several ways depending on how you want to use it. For most readers of this book, you will start with one of the first two.
Open any web browser and go to claude.ai. That is it. No installation. No account required for basic use. A free account gives you more daily conversations and saves your history. No payment information is needed for either.
This is the fastest way to start and the right choice for most people reading this book. Every example in the chapters that follow was created through this interface.
Claude Desktop is a native application that runs on Windows and Mac. You download it from claude.ai/download, install it in about two minutes, and it lives in your taskbar or dock — always one click away, even when you have fifty browser tabs open.
Claude Desktop does everything the web interface does, and adds the ability to work with files on your computer directly, create Word documents and spreadsheets, and run Cowork — the autonomous task feature covered later in this book. It requires a paid subscription.
Claude Code is for developers and technically adventurous readers. It runs in a terminal window and can read your files, write and run code, manage version control, and work directly inside your projects. If the words “terminal” and “command line” are unfamiliar to you, skip this for now — the web interface is the right place to begin.
Installation is a single command once you have Node.js installed:
npm install -g @anthropic-ai/claude-code
Every chapter in this book works with the free claude.ai browser interface. You do not need to spend a dollar to get full value from every example in this book. That said, as you grow more comfortable with Claude, a paid plan unlocks more. Here is a simple guide based on what this book covers.
| Chapter | What You Are Doing | Free | Paid |
|---|---|---|---|
| Ch 1 — Hello, Claude | First conversations, asking questions | ✓ | ✓ |
| Ch 2 — How Claude Works | Learning, context window, handoff files | ✓ | ✓ |
| Ch 3 — How to Ask Claude Well | Better prompting, CLAUDE.md | ✓ | ✓ |
| Ch 4 — Cowork | Autonomous tasks, file work | × | ✓ |
| Ch 5 — Writing | Drafting, editing, publishing | ✓ | ✓ |
| Ch 6 — 3D Design | Fusion scripts, design conversations | ✓ | ✓ |
| Ch 7 — Website and Beyond | Building websites, automation scripts | ✓ | ✓ |
| Ch 8 — Health, Finance, Friend | Health, finance, emotional support | ✓ | ✓ |
| Ch 9 — Medical Care | Understanding options, finding care | ✓ | ✓ |
| Ch 10 — Learn Anything | Deep learning, language, curiosity | ✓ | ✓ |
| Ch 11 — Your Business | Communications, business plans | ✓ | ✓ |
| Ch 12 — Your Future | Planning ahead, habits | ✓ | ✓ |
A paid plan ($20 per month) includes both the claude.ai browser interface and Claude Desktop. Claude Desktop is where Cowork lives — the autonomous task feature covered in Chapter 4. Every other chapter in this book works equally well on the free plan.
The honest answer for a novice: start free. claude.ai costs nothing to begin. Create a free account, work through the first few chapters, and see what Claude does for you. Upgrade when Chapter 4 calls your name.
Claude is not a vending machine with a fixed number of coins. The limits are based on how much work you ask Claude to do — measured in something called tokens, which is roughly a measure of words processed. A short question uses a handful of tokens. Asking Claude to write a full chapter, analyze a long document, or debug a complex problem uses thousands.
In practical terms: the free plan gives you roughly 10 to 20 exchanges in any 5-hour window, depending on how complex your requests are. A paid Pro plan ($20 per month) gives you roughly 5 times that. The window resets 5 hours after you send your first message of the session.
What does that mean for the work in this book? For casual use — asking questions, drafting short pieces, learning — the free plan will take you a long way. For heavier work — writing, editing full documents, longer projects — you may hit the ceiling in a serious session. When you do, Claude will tell you. You simply wait for the window to reset, or continue with a new session using a handoff file as described in Chapter 2.
The honest advice: start free. Use it. If you find yourself hitting the limit every day, that is a good sign — it means Claude is genuinely useful to you. That is the right time to upgrade.
Being honest about Claude’s limits makes it more useful, not less. Here is what you need to know before you begin.
This is the most important thing in this section. Claude can be confidently wrong. It does not always know when it is wrong. It can state incorrect facts in the same calm, clear tone it uses when it is completely correct.
This does not make Claude useless. It makes verification important. Use Claude to learn, explore, draft, and think. Verify specific facts and claims from authoritative sources before acting on them. For medical, legal, or financial decisions, Claude is a starting point — not a replacement for a qualified professional.
Each new conversation starts completely fresh. Yesterday’s productive session is gone today. You are starting from zero every time — unless you give Claude a brief summary of your context at the start. Chapter 2 covers exactly how to handle this.
Claude does not browse the web in real time unless you have enabled that feature. Its knowledge has a cutoff date. For very recent events or the latest software versions, it may not have current information — and it will tell you so.
Open claude.ai in your browser right now. You do not have to wait until you have finished reading. The best way to understand what Claude is, is to talk to it.
Here is a simple way to begin. Think of something you have been genuinely curious about — not a test, not a trick, something real. A question you have never quite found a satisfying answer to. A topic you have always wanted to understand better. A problem you are actually dealing with.
Type it in plain language, the way you would ask a knowledgeable friend. Do not worry about phrasing it correctly. Claude understands context and will ask for clarification if it needs more information.
I have always wondered why the sky is blue but sunsets are orange and red. I know it has something to do with light but I never understood the actual explanation.
The key is that sunlight contains all colours mixed together, and different colours of light scatter differently when they hit the atmosphere. Blue light has a short wavelength and scatters easily — it bounces off air molecules in all directions. When you look at any part of the sky during the day, scattered blue light is coming at you from everywhere, which is why the whole sky looks blue. At sunset, sunlight travels through much more atmosphere to reach your eyes. By the time it arrives, most of the blue has already scattered away. What remains is the longer-wavelength light — the oranges and reds — which is why the sky near the horizon glows warm at dusk.
Follow that thread wherever it leads. Ask the follow-up question. Disagree with something it said. Ask it to explain something differently. The more you put into the conversation, the more you get out.
I did not go looking for a tool. I went looking for a thinking partner.
I have been building things at the intersection of hardware and software since the 1970s — hand-built microcomputers before the IBM PC existed, minicomputers for Ontario Hydro, speech recognition research that earned me an appointment at Carnegie Mellon. I have seen waves of technology come and go. I knew how to evaluate tools.
What I was not prepared for was the first time I asked Claude something real. Not a test. An actual problem I was stuck on — one I had been turning over for two days. Claude asked two clarifying questions and gave me three possible explanations in order of likelihood. The second one was right. Eight minutes.
That was not magic. Claude did not know something I could not have eventually figured out. It helped me think faster and more clearly than I could alone. The difference was speed, availability, and patience — it never got frustrated, never needed to be somewhere else, never made me feel foolish for not seeing something obvious.
Since that night I have used Claude to write books, build websites, debug hardware, learn new tools, draft difficult communications, and explore ideas I would never have developed on my own. Not because Claude is smarter than me. Because together, we are faster and more thorough than either of us alone.
That is what this book is teaching you to build. Not a dependency. A partnership.
The full book covers 12 chapters — all in the same plain English, no jargon style.
Writing · Health ·
Finance & Stocks · Business ·
Learning · 3D Design · and more
Price of a cup of coffee · Available on Amazon Kindle & Paperback
Get notified when new articles are published or when errata updates are posted for Get to Know Claude. No spam. Unsubscribe any time.