Learning and research: from sixty saved papers to one argument
A folder of unread PDFs is not knowledge. A notebook where every source has a question next to it is.
Ira is in the second year of her PhD. The chapter on how remote teams document decisions is due in six weeks. She has sixty papers saved as PDFs, twenty of them with highlights, and five she could summarise from memory. There is also a folder from a course two years ago that she has not opened since, and a feeling that something in it was relevant.
She knows the real problem is not reading. It is that every paper she reads replaces the last one in her head, and at the end she has sixty impressions and no argument.
Three questions change what she reads
Before opening a single PDF, Ira writes three questions in the notebook, each with a line about what would count as an answer. It feels like procrastination. It is the opposite: from that evening on, a paper that answers none of the three gets a note saying so and goes back on the pile, and she stops reading it out of guilt.
Twenty papers get real reading notes, with the quotes she will cite kept visibly apart from what she thinks about them. In six months she will not remember which words were hers. The notes will.
The course folder was relevant after all
In week three Ira asks Second brain to look at her whole library together with the chapter notebook. The graph puts a reading note from that two-year-old course next to her second question, with the reason on the link: both discuss the same 2019 survey, and they read it differently. She had forgotten she ever disagreed with it.
That disagreement becomes the most interesting paragraph of the chapter. The analysis cost her some points; the link she then draws by hand to a third note costs nothing. Her supervisor’s shared group notebook is in the scope too, and her own messy notes stay her own.
A draft that can show its work
The chapter is written by Ira, not by the model. Where a sentence has no source behind it, AI says so and points at the reading note that could support it, or at the two notes that contradict it. She accepts about half of those remarks. The March draft she abandoned stays in version history, and in May one of its paragraphs comes back.
Her supervisor comments on the paragraphs themselves, in the notebook, and the next draft starts from those comments instead of from an email that says “see attached”.
You are writing a thesis chapter, a literature review or just trying to understand a field properly. You have saved sixty papers, highlighted twenty and remember five. Below is how a research notebook is organised in Kartoteq, so that reading turns into notes, notes turn into questions and questions turn into a text with citations you can defend.
How the notebook is laid out
Start with three questions, not sixty files
Create a notebook for the topic and switch on AI. In the goal write what the text at the end should prove or explain: “Why do remote teams document decisions less than co-located ones, and what changes that.” The interview asks what you already believe, what evidence you have, what the reader expects and when the draft is due.
Write your questions as separate notes in the Questions section, three to five of them. Each gets a line: what would answer it. Everything you read from now on is read against those notes, and a paper that answers none of them can wait.
If a question splits in two while reading, make a second note and link them. The section grows with your understanding, which is the point.
One paper, one note, the quote apart from the thought
Each source is a note in Sources: the PDF attached, the citation details on the first line, the link to where it came from. The text of the PDF is searchable, so a half-remembered phrase finds the page.
Reading notes go in a separate section. For each paper you actually read, one note: the passages you want to cite, in quote blocks, with page numbers; below them, what the paper changes in your thinking and what it leaves unanswered. The two parts must look different on the screen, because in six months you will not remember which words were yours.
Link the reading note to the question it speaks to. Tag by method or by school if that helps across notebooks. Then ask AI across Reading notes: which sources support question two, which contradict each other, which question has no evidence at all yet.
Let the graph show what you forgot you knew
After a few months the library holds notes from three courses and two earlier projects. Ask Second brain to analyse it together with this notebook. The graph shows your reading note from last year sitting next to this month’s question, with the reason on the link: both discuss the same survey from 2019.
It also shows contradictions: two notes in your own library that reach opposite conclusions about the same effect. That is usually where the interesting paragraph of the chapter is.
Links you draw yourself cost nothing and are worth as much as the found ones. The AI analysis shows a points estimate before it runs. If your supervisor shares a group notebook, add it to the scope; your private notes stay private, and the graph only shows what you have the right to see.
Write the argument as a document that keeps its sources
The result document is the chapter draft. Write it yourself; ask AI to propose edits where a claim has no source behind it or where two reading notes disagree with the sentence. Each proposal comes with the note it rests on. Accept or reject, one at a time.
Earlier versions stay in history. The draft you abandoned in March is still there in May, when it turns out the abandoned paragraph was right.
Export to Markdown or PDF for the supervisor. If they comment in the notebook instead, their remarks sit on the paragraph they concern, and the next draft starts from them.
What is left when the project ends
A notebook you can open a year later and still understand: the sources, the discussions and the result, with the reason next to every decision. If the next project is similar, you start from here instead of from an empty page.
Where to go next
See how your notes are connectedStart with one notebook.
The Personal plan is free: notes, files and search. AI and shared notebooks can be added later, when a project needs them.
Create account