This manual opens with a test, and it isn't about AWS. Whatever you have been doing to study — the highlighting, the re-reading, the videos at double speed — hasn't worked, and the reason was never discipline. Studying fails when the method doesn't match the way your brain actually stores information, and no amount of extra hours fixes a mismatch; it just spends more hours proving the point. Five honest minutes here will save you two weeks of running someone else's system badly.
So here is mine, stated plainly so you know this manual isn't neutral: I am a memory-palace learner. I put facts in rooms and walk past them, and you will see that habit all through the pages ahead. But mine is mine. The loop in the next section works with any of the five styles below once the AI knows which one is yours — and finding that out is the only homework tonight.
The sixty-second test
For each of the ten questions, pick the single answer that sounds most like you — first instinct, no overthinking, no wishing you were someone else. Jot the letter down and keep moving. When you finish, count your letters: the tallest pile is your primary style. Ties are normal and real; if you tie, read both definitions and keep the one that makes studying feel like downhill instead of uphill.
The five styles, defined — find yourself in one
You think in pictures. Information arrives as shape, color, and position, and one good diagram teaches you faster than three pages of prose. Walls of text slide right off you — which is exactly why re-reading slides never worked and never will.
Run the CCP like this: one diagram per domain, color-coded, redrawn from memory until it reproduces itself on blank paper. Tell your AI: "Explain everything as a diagram or table first — describe the picture in words — then quiz me on it."
You learn by hearing and by saying. If you can explain it out loud, you know it; if you only read it silently, it evaporates by morning. Silent study has always felt dead to you for exactly this reason, not because you were bad at it.
Run the CCP like this: teach each topic aloud to the AI in your own words and let it correct you; use voice review on breaks — the Gemini sidebar in Section 05 is built for you. Tell your AI: "Explain it like a podcast, then make me say it back before we move on."
Words are your interface: lists, summaries, definitions, and above all rewriting. Highlighting failed you because highlighting is recognition, not recall — your style wants the words to pass through your own hand before they count as learned.
Run the CCP like this: ask for tight bullet summaries, then rewrite them by hand without looking, and turn every miss into a flashcard. Tell your AI: "Give me tight bullets I can rewrite in my own words, then check my rewrite."
Your hands learn before your eyes do. Watching a demo without touching anything is why nothing stuck — for you, knowledge only arrives after your fingers have been somewhere and your brain has felt the clicks.
Run the CCP like this: small drills in the AWS free tier, break-fix tasks, real consoles clicked for real — one tiny mission per session. Tell your AI: "Give me one hands-on task in the free tier, then quiz me on what I clicked and why."
Locations and journeys are the oldest, strongest memory system you own, and yours is set to receive. Facts want an address; flat lists are why studying always collapsed on you, no matter how hard you pushed.
Run the CCP like this: the palace build in Section 03, one room per domain, anchors moved in as exaggerated scenes. Tell your AI: "Give me a memory-palace route for each domain, with stops and scenes."
Now make the loop yours
The six-step loop in Section 02 is written in palace — style E — because it's mine. If you tested E, run it exactly as written and move on. If you didn't, keep the skeleton and swap the middle: test first, teach from your misses, do a bedtime ritual, sleep eight hours, recall in the morning — that part never changes. Visual learners redraw their diagram of the day from memory; auditory learners record a one-minute recap of the three facts out loud; read/write learners rewrite their three anchors by hand; kinesthetic learners run one small free-tier drill last thing before bed. The AI translates the whole loop once it knows your letter — you'll add one line to the kickoff prompt in Section 05, and from the second session on, the system will be running your style, not mine.
You don't need more hours. You need a system that spends the hours you already have on the right things — and then protects your sleep like it's part of the studying, because it is. Everything below is one loop: let the AI test you first, teach you from your misses, and let sleep lock it in. Forty-five minutes a day is enough.
If we're walking through this together, this page is the map we'll follow. If you're reading it on your own, it's still enough — you can run the whole thing without me.
Evelyn
Eight hours of sleep is not the reward for studying. It is studying. While you sleep, your brain replays what you learned and moves it from short-term into long-term storage — and that transfer is the entire point of studying. Forty-five focused minutes followed by a real night of sleep beats four exhausted hours followed by four hours of tossing. You are never "behind" when you sleep. You are finishing the work.
The loop — six steps, same every day, shown palace-style
The AI quizzes you before it teaches you. No warm-up reading.
It explains your misses in plain words and attaches vivid anchors.
You rewrite three facts as image-sentences, in your own words.
Five minutes walking your memory palace, right before bed.
Eight hours. This is when today's learning actually gets saved.
Next morning: quiz on yesterday before any new material.
Tonight, in fifteen minutes
- Confirm your Gmail works. That is the only account you need. Both DeepSeek (chat.deepseek.com) and Z.ai (chat.z.ai) are free in the browser and sign in with Google — nothing to install, nothing to pay.
- Open either one, start a new chat, and paste the kickoff prompt from Section 05. On your phone? Screenshot the prompt page and type it in — two minutes. Then just answer its ten diagnostic questions. Guessing is expected; the misses are the point.
- When the AI ends the session and tells you to sleep — obey it. Tonight, finishing the diagnostic is the whole win. The plan starts tomorrow, rested.
Here is why the way most people study fails a tired brain. Rote memorization asks an exhausted mind to do the one job it is worst at: raw repetition without meaning. When your attention is thin, re-read slides past without sticking — you recognize the page and still fail the question. Worse, marathon sessions feel productive while actually producing little, which is how someone ends up studying until 2 a.m. and remembering almost none of it at 9 a.m.
Anchors flip the job. Instead of memorizing words, you glue each fact to a picture your brain already likes keeping: something vivid, physical, and slightly ridiculous. Ten hours of re-reading loses to five minutes spent building the right image — and five minutes is all a tired person can honestly afford, which is exactly why this method fits your life right now.
What an anchor looks like
An anchor is one sentence that plants an image and carries the fact inside it. Make it yours: motion, sound, and absurdity all make images stickier. Here are five ready-made ones from the exam's core services — steal them, then rewrite each in your own words tonight, because the rewriting is where your memory engages.
Checks IDs (users and roles), follows the guest list (policies), hands out wristbands (permissions). Least privilege means the bouncer only opens the doors you actually paid for — nobody wanders into the back office.
The fence is yours to guard: security groups are the guards on each gate, network ACLs patrol the fence line, and you decide which gates open to the street (the internet) and which stay shut.
Buckets are bins with unique names, objects are whatever you toss in, you pay for what each bin holds, and versioning keeps the old stuff you shoved to the back — just in case.
Press the button (a trigger), it runs your code, you pay per press, and when nobody is pressing it costs nothing at all. There is no server to feed, house, or babysit — that's "serverless."
Amazon owns the building, patches the walls (engine updates), and handles the boiler (automated backups). EC2 by contrast is the empty rental you furnish and maintain yourself — that contrast is the shared responsibility model in one image.
Build your memory palace in ten minutes
A memory palace is simply a route you know by heart, with facts placed along it as scenes. It works because spatial memory is the oldest, strongest memory system you own — and because a route has an order, so nothing gets lost. You already have the route. Here is the setup:
- Pick a route you could walk blindfolded — your apartment, or the walk to work. Number the stops: front door, hallway, kitchen, living room.
- Give each exam domain a stop, biggest domain in the biggest room. Cloud Technology and Services (34%) takes the living room; Security and Compliance (30%) takes the front door; Cloud Concepts (24%) takes the hallway; Billing (12%) takes the kitchen.
- Move your anchors in as exaggerated scenes. A bouncer carding guests in your doorway. A vending machine humming in your kitchen corner. The stranger and more alive the scene, the longer it survives.
- Walk the whole route every night, right before sleep. See each scene once, in order. Five minutes. Doing it last means the walk is the freshest thing your brain carries into sleep — and sleep consolidates what is freshest.
In the morning, before any new material, walk the route again from bed and tell your AI what you remember — it will quiz the gaps. Recall is the actual lift; re-reading only feels like lifting. The night walk loads the truck and sleep drives it; the morning walk is you checking what actually arrived.
This is the part of how I study that actually moved my scores, and at first it feels like cheating. It isn't. Most multiple-choice questions hand you the answer before they show you the choices — the question stem contains everything you need, and the choices exist mostly to make you doubt what you already know. Reading the choices first is how careful people talk themselves out of right answers: every plausible-sounding option that isn't yours lays down a little fog. So don't read them. Not yet. The question comes first, alone, and you squeeze it for everything it has.
Three moves, every question, same order
Cover the choices with your hand. Read only the question, twice. Find the keywords: MOST, LEAST, cost-effective, operational overhead, highly available, managed, temporarily, auditor.
Answer it in your head before looking. The keywords point at one service: "least operational overhead" means managed or serverless; "most cost-effective" means pay only for what you use.
Uncover the choices. If your answer is there, take it and move on — in seconds. If it isn't, you missed a keyword: go back to move 1 and re-deconstruct. Never negotiate with a choice that "sounds smart."
The question: "A company wants to run its application code with the LEAST operational overhead. Which AWS service should the company choose?"
Move 1: hand still over the choices. Keywords: LEAST, operational overhead. Translation: which option makes AWS manage the most machinery for you? Move 2: predicted answer — the one with no servers to feed, house, or babysit at all. Move 3: uncover the choices — Amazon EC2, Amazon ECS on EC2, AWS Lambda, Amazon RDS. Lambda is sitting right there. Take it and move on: eight seconds, no fog, no second-guessing.
That is the whole method. The question was never hiding the answer — it was announcing it.
The AI does the heavy lifting in this system: it writes the questions, grades them honestly, builds the plan, and keeps the list of what you're weak on. You only have to show up for 45 minutes and answer honestly. Setup is deliberately tiny, because setup is the first thing a tired person abandons.
- A Gmail account is the whole prerequisite. That's it. Both DeepSeek (chat.deepseek.com) and Z.ai (chat.z.ai) run free in the browser and sign in with Google.
- Pick one and stay with it for the whole sprint. They behave similarly; switching mid-sprint throws away your weak-topic list and your diagnostic history, which are the most valuable things you own here.
- Start a new chat and paste the kickoff prompt below. On your phone, screenshot this page and retype the prompt once — after that, the AI runs the loop and you just answer.
One line in the prompt does a lot of quiet work: "harder questions, deeper explanations." Say the bar out loud — you want the AI to treat every practice question like the real exam, and to explain not just why the right answer is right but why each wrong answer is tempting. That single instruction turns a quiz into a tutor, and it matches how the real exam thinks: every wrong option on the CCP is a plausible idea someone half-learning would hold.
You are my AWS Cloud Practitioner (CLF-C02) study partner, and you will follow these rules for our whole conversation.
(0) Before anything, ask me for my learning-style letter from the manual's Section 01 test — A visual, B auditory, C read/write, D kinesthetic, E memory palace — and adapt every explanation, anchor, quiz, and plan to that style for our whole conversation. (1) Start with a diagnostic: give me ten exam-style questions across the four official domains, one at a time, waiting for my answer before the next. Use the official weights: Cloud Concepts 24%, Security and Compliance 30%, Cloud Technology and Services 34%, Billing, Pricing and Support 12%. (2) Grade strictly and explain deeply: for every question, tell me why the correct answer is correct AND why each wrong answer is tempting — and before you show me the choices on any question, make me deconstruct it: name the keywords and predict the answer from the question alone, like the manual's Section 04. I want harder questions and better explanations, not easy wins. (3) Then build me a day-by-day plan assuming 45 to 60 minutes per day. Eight hours of sleep per night is non-negotiable — schedule around it, never through it. (4) Teach my weakest topics first, in plain language, attaching each concept to an anchor that fits MY style: a visual image, a spoken phrase, a rewritten summary, a hands-on task, or a scene in a memory-palace route — front door, hallway, kitchen, living room. (5) Quiz me again after each teaching block before moving on. (6) Keep a running list of my weak topics and show it at the start of every session. (7) Keep sessions under 30 minutes. (8) End every session with exactly this line: "Close the laptop. Walk your palace. Go sleep."
Everything in the table below comes straight from the official AWS exam guide for CLF-C02 — the current Cloud Practitioner exam — read directly from AWS's own certification documentation in October 2026. Treat AWS's pages as the only source of truth for what's on the test; prep videos and blogs drift out of date or invent details. When anything conflicts, the exam guide wins. Bookmark it: search "AWS Certified Cloud Practitioner exam guide" and open the docs.aws.amazon.com result.
| Exam fact | CLF-C02, verified October 2026 |
|---|---|
| Questions | 65 total — 50 scored plus 15 unscored. The unscored ones are indistinguishable, so answer everything. |
| Time | 90 minutes — roughly 83 seconds per question. Flag and move; the exam lets you revisit. |
| Question types | Multiple choice (one correct of four) and multiple response (two or more correct of five). |
| Passing | 700 on a scaled score of 100 to 1,000. Pass/fail — you don't compete with anyone. |
| Guessing | No penalty. An unanswered question scores zero, so never leave a blank. |
| Cost | USD 100. |
| Delivery | Pearson VUE — at a test center or online proctored from home. |
| Validity | 3 years from the day you pass. |
The exam is organized into four domains, and the weights tell you exactly where your 45 minutes should go. Domain 3 alone is a third of the exam; Domains 3 and 2 together are nearly two-thirds. The scoring is compensatory — you don't need to pass each domain, only the total — but nobody fails Domain 4 on this exam. People run out of runway on Domains 3 and 2, so that is where the anchors and the palace stops live.
| Domain | Weight | What it covers | Palace stop |
|---|---|---|---|
| 1 · Cloud Concepts | 24% | What the cloud is, why it beats owning servers, the six advantages, the Well-Architected idea. | Hallway |
| 2 · Security and Compliance | 30% | The shared responsibility model, IAM, guarding everything — the bouncer lives here. | Front door |
| 3 · Cloud Technology and Services | 34% | Compute, storage, databases, networking — the services everyone actually names. | Living room |
| 4 · Billing, Pricing and Support | 12% | How AWS charges, how you pay less, which support plan answers which need. | Kitchen |
The daily loop is the same every day, and that sameness is the feature — a tired brain doesn't need decisions at 9 p.m., it needs a default. You are never deciding what to study; the AI's weak-topic list decides. You are only deciding to start, and starting takes one sentence: open the chat, and say "next session."
The daily checklist
- 25 minutes with your AI. It quizzes, teaches your misses with anchors, and re-tests before letting you go. Hard questions, full explanations, every time.
- Three anchors, written down. Today's three most important facts, each as one vivid image-sentence in your own words. Yours, not the AI's — the rewriting is the learning.
- Five minutes, lights low. Walk the palace once, in order, seeing every scene. Last thing before sleep, every night.
- Eight hours of sleep. This is the studying. Everything above only sticks if this happens.
- Ten minutes next morning. Tell the AI what you remember from yesterday's walk before it teaches anything new. Gaps found in the morning cost nothing; gaps found on exam day cost $100.
Track exactly two numbers: your mock score and the length of your weak-topic list. One goes up, one goes down — that's the whole dashboard, and the AI maintains it. And when a bad day happens, because it will, the rule is: shrink the session, never skip it. Ten minutes of anchors plus the palace walk still counts. A skipped day costs momentum; a shrunk day costs nothing at all.
The 14-day sprint
Fourteen days at the default pace, undated on purpose — Day 1 is whatever day you paste the prompt. Stretch it to three weeks or compress it to ten; the AI re-plans around your diagnostic results anyway. Days 7 and 12 are the only fixed points: full mock exams, scored honestly, every miss explained and anchored.
| Day | Focus | You're done when |
|---|---|---|
| 1 | Setup plus the 10-question diagnostic | You have a plan and a weak-topic list. |
| 2 | Domain 3: compute — EC2 and Lambda | You can say what each service replaces in a normal office. |
| 3 | Domain 3: storage and databases — S3, RDS, DynamoDB | All your anchors are placed in the palace. |
| 4 | Domain 3: networking — VPC and friends | You can draw the fenced yard from memory. |
| 5 | Domain 1: cloud concepts and the six advantages | You can say why the cloud beats owning servers. |
| 6 | Domain 1: migration strategies, the 7 Rs | You can name most of the Rs without looking. |
| 7 | Mock exam 1, then deep review | Every miss is explained, anchored, and on the list. |
| 8 | Domain 2: shared responsibility and IAM | You can split any task into AWS's half and yours. |
| 9 | Domain 2: compliance and guarding data | The front-door scene in your palace got richer. |
| 10 | Domain 4: pricing models, billing tools, support plans | You know which support plan answers which need. |
| 11 | Weak-topic backlog — your list, not the plan | The list is visibly shorter than yesterday's. |
| 12 | Mock exam 2 | Score trending at or above a comfortable pass. |
| 13 | Light review plus one long palace walk | No new material. You walked the whole route twice. |
| 14 | Ten-minute recall, then early sleep | You stop early — that's the plan working, not failing. |
When you schedule the real exam, pick a morning slot if you can — your recall is best after a full night of sleep, which by Day 14 you'll have banked for two straight weeks. Bring the ID Pearson VUE asks for, arrive early, and remember the arithmetic working in your favor: 700 out of 1,000, no penalty for guessing, and every question you anchored in a room you can still walk through. You have been training for this in your sleep. Literally.
Test before you read. The AI quizzes first; your misses build the plan. Anchor, then walk. Facts live in rooms, and the walk happens last, right before sleep. Eight hours, every night. Sleep is not the reward — it is the studying.