The AI Setup Guides ยท One free section
One full section of 25, free. The other 24 are paid.
This is one complete section of the Claude Setup Guide, published in full with nothing trimmed and nothing gated. It is here so you can judge the writing before you spend anything. The rest of the Claude guide, all of the ChatGPT guide, and every future edition are the paid product, $79 CAD plus tax, one time.
30-day satisfaction guarantee. Email within 30 days for a full refund and include a brief note about what did not work. A full refund ends the licence and guide access.
From Edition 2026.4 of the Claude Setup Guide, version 3.7, last updated 5 September 2026. The current model lineup and Fable plan access changed in this version were checked against Anthropic's official pages on 5 September 2026. The broader product review remains dated 12 August 2026. Where your screen and the guide disagree, the live product and pricing pages win.
Verified as of 11 August 2026. The volatile claims changed in this release were checked against Anthropic's and OpenAI's own published pages on that date.
Published in full
Privacy and data defaults.
Everything below this line is free, in full, and identical to what buyers read.
Free section
Section 10
Privacy and data defaults 4 min
You have seen what Claude does. Now spend nine minutes across this section and the next making sure it never costs you anything. This half is the training toggle. The next half is what is safe to put in.
The default changed in 2026. On the consumer plans (Free, Pro, and Max), your new chats are used to help train Anthropic's models unless you turn that off. This is a reversal of how it used to work, so most people who signed up before are affected too.
How to opt out
Open Settings, then Privacy
You are looking for the training toggle.
Find "Help Improve Claude" or the equivalent model-improvement setting
Some rollout cohorts may still show "Help improve our AI models." Match the meaning and confirm the description says whether new chats may be used to improve models.
Turn it OFF
That stops new and resumed chats from being used for future model training. It does not delete your chat history or create a 30-day deletion timer.
New signups, check the toggle during setup
You may be asked about this when you first sign up, with a big "Accept" button and a smaller toggle. Slow down for a second, find the toggle, and set it the way you want before you click through.
Training and retention are two different controls
Deletion is an action, not a toggle
When you delete a conversation, it disappears from your history immediately and Anthropic says it is deleted from back-end storage within 30 days. Content flagged by safety systems may be retained longer: inputs and outputs for up to 2 years, and safety classification scores for up to 7 years. Legal and dispute holds can also apply. Do not use any privacy mode as permission to paste confidential material.
PersonalYou use a normal chat to plan a holiday. Turning training off keeps that chat out of future training, but the chat remains in your history until you delete it.
BusinessA no-training workspace is useful, but it does not grant permission to upload client material. Your agreement, workspace policy, and the client's authorization still decide what can go in.
Do this now
Go to Settings, Privacy, turn model improvement off if you do not want future chats used for training, and note your plan. Then open one harmless test chat, delete it, and confirm it disappears from your history. Remember the distinction: the toggle controls training; Delete starts deletion.
NextSection 11 is the other five minutes: a three-colour rule for what you can safely paste, and the incognito escape hatch.
That is the end of the free section. Everything past here is paid.
Paid from here
The paid editions are checkable
Tick one and see.
Every "Do this now" step in both guides, including the one at the end of the section you just read, carries a box like these. Tick it and it stays ticked: in your browser the moment you click, then on your purchase a couple of seconds later, so opening the same guide on your phone shows the same boxes already ticked. Your library page counts how many steps are done, and on any device that has already opened the guide it points the link straight at the first step you have not ticked. When every box in a guide is ticked, a certificate of completion unlocks with your name on it. The three below are the real control, wired to this page only, and they are the three steps you have just read.
These three are stored in this browser and nothing is sent anywhere. Reload the page and they will still be ticked. In the paid guides the same ticks are saved to your purchase, which is what lets a second device pick them up.
Also included in the purchase
The five-level Beginner to AI Architect course.
The two guides show the current products. The course turns that setup into lasting capability. It starts with one safe result, then progresses through secure operation, Projects and context, Skills and agents, coding and automation, and a governed personal AI operating system.
38 practical lessons
Every lesson ends with an exercise, proof criteria, recall check, and an optional advanced layer.
5 hands-on final projects
Build a personal assistant, work assistant, power-user system, development environment, and personal AI operating system. Inside the course, these projects are called capstones.
44 reusable workspaces
Download and adapt security rules, instruction files, context packs, Skill files, evals, recovery records, and capstone checklists.
Private AI Command Center
See your next session, progress, permanent milestones, setup profile, spaced reviews, and the artifact to build next.
What is behind the line
Every section of both guides.
Titles only, so you can see exactly what you would be getting before you pay for it. The privacy and data defaults section of the Claude guide is the one you just read in full. The Claude guide runs twenty-five numbered sections, the ChatGPT guide twenty-eight, along the same spine, written for that product rather than translated across.
Claude Setup Guide
- 01 Start here and how to use this Paid
- 02 Before you start Paid
- 03 AI basics in plain English Paid
- 04 What Claude is and which model to use Paid
- 05 It can be confidently wrong Paid
- 06 Know when you have gone past what Claude can do Paid
- 07 Pick the right plan Paid
- 08 Install, sign in, and secure your account Paid
- 09 Get your first 5 wins Paid
- 10 Privacy and data defaults Free above
- 11 What is safe to put into Claude Paid
- 12 Talk to it well Paid
- 13 How to tell it is wrong Paid
- 14 Personalize Claude Paid
- 15 Build reusable workspaces (Projects) Paid
- 16 Files, images, voice, and Artifacts Paid
- 17 Connect your apps safely Paid
- 18 Claude Cowork, the agent that does the work Paid
- 19 Claude Code, the third tab in the app Paid
- 20 What it costs, and how not to overpay Paid
- 21 Business workflow recipes Paid
- 22 Prompt recipe library Paid
- 23 Free official courses Paid
- 24 Troubleshooting and FAQ Paid
- 25 Your 30-day plan Paid
- Ref Where this stops working Paid
- Ref Go deeper (advanced) Paid
- Ref Final checklist and scorecard Paid
- Ref Plain-English glossary Paid
ChatGPT Setup Guide
- 01 Before you start Paid
- 02 AI basics in plain English Paid
- 03 Catch it being confidently wrong Paid
- 04 What ChatGPT is and the current models Paid
- 05 Pick the right plan Paid
- 06 Install, sign in, and secure your account Paid
- 07 Your first win, before anything else Paid
- 08 Privacy and data defaults Paid
- 09 What is safe to paste or upload Paid
- 10 Four more wins Paid
- 11 Talk to it well Paid
- 12 Check the answer before you send it Paid
- Practice Build a brief, one slot at a time Paid
- 13 Spot the edge of what it knows Paid
- 14 Personalize ChatGPT Paid
- 15 Build reusable workspaces (Projects) Paid
- 16 Build your own custom GPT Paid
- 17 Files, images, voice and writing Paid
- 18 Connect apps safely Paid
- 19 ChatGPT Work and the desktop app Paid
- 20 Codex, the coding agent in the same app Paid
- 21 Scheduled Tasks Paid
- 22 What it costs to run Paid
- 23 Business workflow recipes Paid
- 24 Prompt recipe library Paid
- 25 Free official courses Paid
- 26 Troubleshooting and FAQ Paid
- 27 Your 30-day plan Paid
- Ref Where this stops working Paid
- 28 Go deeper (advanced) Paid
- Ref Final checklist and scorecard Paid
- Ref Plain-English glossary Paid
The privacy section you read above is the Claude guide's privacy and data defaults section. The ChatGPT guide covers the same ground in its own privacy and data defaults section, Section 08 in the current edition, written against OpenAI's own settings and defaults.
The question a single-tool course cannot answer
Which one do you open?
Anthropic and OpenAI both publish a free course on their own tool, and both are worth doing. Neither one covers the other tool, so neither one can tell you which of the two to open for the job in front of you. The library page your purchase opens carries a routing table that answers that job by job, with the reason behind each call. It sits there rather than inside either guide, because it is the one part of this product that belongs to both. Two rows of it, published here in full:
Two rows of a longer table. The rest of it is on the library page your purchase opens, and it is re-checked against both vendors' own pages every edition, because the answer moves when the products move.
The rest of it
Fifty-three numbered sections. Twenty-five here, twenty-eight there.
You just read the privacy and data defaults section of the Claude guide. There are twenty-four more numbered sections there, and twenty-eight in the ChatGPT guide, plus the reference material at the back of both. Plan selection, account security, first wins, prompting, custom instructions, Projects, connected apps, the agent modes, running cost, workflow recipes, the prompt library, a 30-day plan, troubleshooting, a final checklist, and a plain-English glossary. Same voice, same level of detail, for Claude and for ChatGPT.
$79 CAD plus tax, one time. Every future edition included while your licence remains active. The 30-day satisfaction guarantee ends the licence and guide access when a full refund succeeds.
Already bought them and lost the email? Get a fresh access link.
