The IKEA manual for learning anything

Four ordinary tools, one learning loop

July 30, 2026

In partnership with

I have a language app I abandoned at day nine, a Kindle full of books stuck at 12 percent, and an ESP32 board that has been in its box since March. That is my learning system. There isn't one.

It works out roughly like putting together IKEA furniture with no instructions in the box. You burn an afternoon guessing which screw goes where, instead of picking up the sheet somebody already wrote and skipping to the part where the thing stands up.

That instruction sheet is the whole idea here, and I am slightly annoyed at how long it took me to see it.

I came across a walkthrough from Tina, an ex-Meta data scientist who moved into AI content and now runs an AI education business. She shared the personal system she uses to learn things roughly ten times faster, and what got me was how ordinary the pieces are. Four connected parts, all of them tools I already have open. Every part maps onto the IKEA box, so let me walk you through it the way I walked myself through it.

The New Blueprint for AI Powered Support

Most support teams are experimenting with AI. Few are transforming because of it. The gap is where the most consequential decisions in support are being made, and where most teams get stuck. Hear from industry leaders on moving from pilot to production.

*Ad

The four parts in the box

Every flat-pack box ships the same core pieces. Here is how her system lines up:

The instruction sheet (stand on the shoulders of giants). Instead of guessing, you borrow the proven path somebody already figured out.

The custom fit (adjust it for you). No two rooms are the same, so you tweak the plan to the skill level you have and the goal you want.

The screwdriver (the productivity system). A plan with no tool to build it stays in the box. This is what makes you do the work.

The receipt drawer (the memory system). You keep a record of what you built and how, so you can look back later.

Here is what each of those looks like in practice.

Step 1: Grab the instruction sheet

She starts in Notebook LM, which recently rebranded to Gemini Notebook. Whatever you want to learn, odds are somebody already learned it and left a trail behind them.

She opens a fresh notebook and types a clear goal. Something like:

I want to learn AI coding so I can eventually build a desktop app that tracks my productivity habits.

She names the specific tools she has heard of too, Claude Code, Codex, Cursor, so the notebook knows the territory she is aiming at.

The tool then goes and gathers real sources for her. She imports them, asks it to design a study plan, and in seconds she has a structured four-week roadmap. Her point lands hard: doing that the old way, digging through Reddit threads and other people's abandoned plans, would have eaten days.

10x the context. Half the time.

Speak your prompts into ChatGPT or Claude and get detailed, paste-ready input that actually gives you useful output. Wispr Flow captures what you'd cut when typing. Free on Mac, Windows, and iPhone.

*Ad

Step 2: Make it fit your room

This is where the workflow got clever for me. A generic plan is a starting point, not the finish line.

She copies the study plan out and drops it into her go-to chatbot, in her case Claude, because that is the one that knows the most about her. Then she asks it to customize the plan against her real background.

Here is the part I loved: the AI pushed back. It noticed she already holds a computer science degree and runs coding pipelines daily, so the beginner framing about overcoming your intimidation did not fit her at all. It rebuilt the plan around her actual level and even referenced the hardware she owns.

A few practical moves from this step:

  1. Use the chatbot that knows you best, so the plan reflects where you genuinely start rather than where an average beginner starts.

  2. Ask it to turn the plan into an interactive HTML checklist you can tick off as you go.

  3. Chat with it live while you work. "How do I set this up?" and "what do you mean by this?" are the whole point.

She also had a smart take on picking the model for the job, which I have started copying:

A local model for anything private you would rather not send anywhere.

A visually strong model when it needs to look at pictures and diagrams with you, like wiring up electronics.

A big-context model for the weeks that are mostly heavy reading.

A fast, direct model for the sessions where you just want to power through.

Step 3: Pick up the screwdriver

A plan you never touch is a nice document and nothing else. My abandoned language app is proof. To close that gap, she vibe coded two small desktop widgets for herself.

One is a typewriter-style app where she pastes her study plan and checks off tasks. The other is an avocado-shaped Pomodoro timer where she names a task, sets the minutes, and logs the session when it finishes. Both widgets quietly save everything into her notes, which becomes her accountability layer.

I tried it without building anything: phone timer running, checklist pinned open on the second monitor. A timer going and a list you can see. That is the whole mechanism, and it did more for me in a week than the plan did on its own.

The best voice models, now with full orchestration. Build real-time voice and chat agents on one low-latency stack: any LLM, your tools and knowledge, testing, Guardrails, and omnichannel deployment.

*Ad

Step 4: Keep the receipts

The last piece is a memory system built in Obsidian. Her widgets automatically log every Pomodoro and every study session there, so she ends up with a running record of how she works and what she learns.

Then she points a local AI agent at those logs. It graphs her habits, suggests productivity tweaks, and answers questions like "what was I working on two weeks ago at this time?" That kind of recall is basically impossible from memory alone, and it only exists because the logging happened automatically.

This is the part I have never once done. My notes app is a graveyard of sessions I never wrote down, which means I have no idea when I actually learn well or what makes me quit. She does, because her tools log it for her while she works.

Where this kit works beyond code

The nice thing is that none of this is coding-specific. She mentioned running the exact same flow for:

AI coding, ending in a real desktop app she uses every day.

Electronics projects, with an ESP32 board.

Gardening, of all things.

Basically any skill with a learning curve fits the kit, whether that is a career skill, a new language, a hobby, or prepping for something concrete like a driver's test.

Why I am rebuilding my own loop this week

What I appreciate most is how boring the parts are. A research notebook, a chatbot I already pay for, a timer, a notes app. None of them impressive on their own. She just wired the four together so each one hands off to the next: borrow the proven path, tailor it, do the work, keep the record.

I have had all four of those pieces on my machine for months. Four tabs open, none of them talking to each other.

Tonight, open a notebook and write one sentence about what you want to learn and why. Just the goal, nothing more. That single line is the top of the instruction sheet, and everything else in this kit hangs off it.