One email each morning — yesterday's AI, sortedGet it in your inbox

METAL LAB

AI STUDY · BEGINNER COURSE

Learn AI without wondering where to start

Start with what today's AI is, how to ask better questions, and what not to trust. This path combines clear explanations, short lessons, and exercises you can try right away.

Start with lesson one

LEARNING PATH

A practical order for learning AI

Each step makes the next one easier. Read a module in about ten minutes, then use the linked glossary entries whenever you want more depth.

01

1. See the big picture

Understand why AI, generative AI, and LLM can sound interchangeable, and how an engine differs from the product built around it.

02

2. Try ChatGPT and Claude

A chatbot is a product and an LLM is its engine. Ask both services the same question and compare which answer fits your work.

03

3. Ask well and catch wrong answers

A useful prompt makes the role, context, output, and constraints explicit. Treat confident answers as claims to verify, not facts to accept.

04

4. Move from answers to actions

AI agents can search, create files, and carry a task through several steps. Learn how MCP and skills connect tools and repeatable methods.

FREE MINI LESSON

Three AI exercises to try today

Doing one comparison teaches more than reading a long feature list. These exercises work in the free chat screen of ChatGPT or Claude and require no coding.

01

Compare two AIs with one question

Use the exact same request in ChatGPT and Claude. Compare accuracy, tone, and format against the work you actually need done.

I have eggs, tofu, and scallions. Give me three dinners I can make in 15 minutes, with quantities and steps in a table.
02

Turn a vague request into a good prompt

State the role, context, output, and constraints. Then ask for the same answer again with a different length or audience.

You are an editor for beginners. Explain why AI needs data centers to a middle-school student in five sentences, with one analogy and short definitions for technical terms.
03

Verify the answer

Do not trust numbers or recent claims on confidence alone. Ask for the primary source, date, and exact supporting line, then open the source yourself.

List every fact and number in this answer that needs verification. Add a primary-source URL and publication date, and write ‘not verified’ instead of guessing.

START HERE

Before you begin

Do I need to know how to code?

No. This beginner path starts with asking questions in a chat window and checking the answers. The first four modules require no technical background.

Should I start with ChatGPT or Claude?

Try both free accounts with the same task. Compare them on the work you do often, such as summarizing documents, writing, or research, rather than relying on a general ranking.

AI changes quickly. What should I trust?

Check product behavior and pricing in official documentation and new announcements in the company's newsroom. Verify numbers and dates against the primary source, while using the glossary for concepts that change more slowly.

Learn AI from the beginning — METAL LAB