Skip to main content

AI Training

From zero to a working AI system in 60 minutes.

A paint-by-numbers guide to AI infrastructure for absolute beginners.

Most AI advice skips the most important part: where your files go. This guide teaches you to build a personal system where prompts, conversations, and project files live in one organised place — so every AI session builds on the last.

Audience

Anyone who has never used AI tools before

Outcome

A personal workspace where your AI work lives in one organised place

Time

About 60 minutes for the full setup

Why most AI investments fail

The bottleneck was never the model. It’s the system around it.

0%

of companies struggle to scale AI value

BCG

0%

of organisations see no measurable ROI

MIT Media Lab

0%

abandoned AI initiatives in 2025

S&P Global

0%

say effort feels disproportionate to value

ClickUp

The 24% seeing real returns aren’t smarter prompters — they’re better file managers.

Two ways to work with AI

Same tools. Same models. Different infrastructure.

Most people
Ask AI to generate something
Get a great response
Paste it somewhere… anywhere
WordGoogle DocsNotionDesktop
Need it two weeks later
Can’t find it. Wrong version. No context.
Start over from scratch
“AI doesn’t really help that much”

Repeat forever

What actually works
Ask AI to generate something
Get a great response
Save it in Obsidian as a Markdown noteFree app. Your files. Your folders. No lock-in.
Organise in a simple folder structureProjects, references, conversations.
Need it again? Search, find, reuse instantly.
Feed past work back as context for new sessionsVS Code for power editing. GitHub for version history.
Every session builds on the last

Context compounds

The bottleneck was never the model. It’s the system around it.

What you’ll learn

Ten parts, zero code required. Start with the first three today.

Quick start
01

What you need

AI accounts, file management tools, and the basics of Markdown.

15 min

02

Set up your workspace

Install Obsidian or VS Code, create your folder structure, write your personal profile.

15 min

03

Your first AI conversation

Custom instructions, file uploads, and your first reusable prompt.

15 min

04

Your prompt library

Six tested prompts for research, planning, summarising, review, teaching, and exploration.

20 min

05

Getting unstuck

Six diagnostic prompts for common blockers — from “I don’t know what I don’t know” to “I’ve been going in circles.”

15 min

06

What AI gets right and wrong

Three failure modes every user should understand: hallucination, sycophancy, and user bias.

10 min

07

Building your rhythm

A sustainable weekly workflow. Minutes, not hours. Paste profile, paste prompt, save response.

5 min

08

Where this can take you

Case studies, adversarial analysis, and a quality assurance framework for your AI output.

10 min

09

How projects grow

Three stages: prototype (minutes), experiment (days), production (weeks). Know which stage you’re in.

10 min

10

Where to go from here

Personal AI Infrastructure philosophy. Building a system that scales with you.

5 min

Key concepts

Three failure modes to watch for, and the one principle that changes everything.

Hallucination

The factory produces defective output.

AI generates confident, well-structured responses that are factually wrong. There is no red warning light — hallucinated text looks identical to accurate text.

Defense:Verify claims independently. Cross-reference. Never cite AI output as a primary source.

Sycophancy

The factory tells you what you want to hear.

AI tends to agree with you, validate your assumptions, and avoid pushing back on flawed reasoning. It is optimised for helpfulness, not honesty.

Defense:Explicitly ask for counterarguments. Use adversarial prompts. Push back on agreement.

User bias

You control what goes into the factory.

The frame you bring shapes the output you get. If you only ask for evidence supporting your view, that is exactly what you will receive.

Defense:Provide multiple perspectives in your prompt. Ask for the strongest counterargument to your position.

Context compounding

The 24% of organisations seeing real AI returns aren’t using better models. They build context systems — organised, reusable knowledge that makes every session more valuable than the last.

Infrastructure beats intelligence. Every time.

Ready to start?

Download the full guide — all ten parts, prompt templates, appendices, and glossary. It’s a plain Markdown file. Open it in any text editor, or feed it to an AI assistant for a personalised walkthrough.