The Story
Expert-curated reading lists for researchers. Because finding the right papers shouldn't be this hard.
I work in science, especially in astronomy and psychology. Often, I need to study new areas. And every time I do, I face the same challenge: finding the right papers to read.
Some metrics help. The number of citations, for example, is useful sometimes for well-established areas where you might get thousands of references. But a lot of cites doesn't mean a paper is good. And then there are niche areas with a handful of specialists spread across the world. There, metrics are useless. It's simply hard to navigate.
I was lucky to have good supervisors and colleagues who were helping me. They'd point me to the right papers, warn me about the wrong ones, and share the kind of insights that never make it into a search engine. That guidance is invaluable.
However, when I need to learn something in an area where I have no such colleagues, I spend a lot of time finding relevant references. A lot.
That's why I started Marginalia. Let an expert in the field create curated collections, foundational or otherwise, and add quick notes on what to pay attention to.
On Subjectivity
Are these collections or recommendations subjective? Yes. Furthermore, we encourage subjectivity.
The choice might be biased. It might depend on personal preferences or even friendship. However, these are the things that make us humans.
So, consider all the content in Marginalia biased and highly subjective :)
Marginalia has two core formats:
- Curated Collections — reading lists for specific topics, annotated with expert notes. Each paper comes with a verdict (Must Read, Worth Reading, Skim, Niche, Skip), a difficulty level, and the expert's personal note. Think of it as: “if I were mentoring you, I'd tell you to read these and here's why.”
- Weekly Digests — fresh picks from recent publications, filtered and commented on by experts. Not fifty new papers from arXiv, but three that actually matter this week and why.
On AI
We all are highly encouraged by Large Language Models and AI. They can do great things. While we are still seeking for real intelligence, it's time to use what we have accordingly.
Many people have concerns that LLMs can hallucinate. That's partially true, though I believe that human beings hallucinate more. Therefore, in Marginalia, LLMs serve as tools. They help monitoring interesting manuscripts. They make summarization and data extractions for the expert.
But the last word is for the human being.
Humans make verdicts.
Humans add notes.
Humans craft collections.
(Though I believe one day we will have AI-curated content to compare :))
Why are the first two topics astronomy and psychology?
Remember that I've told about subjectivity? :)
Why is it called Marginalia?
And who does not like those funny images appearing in medieval manuscripts? Knights fighting snails, monks riding bizarre creatures, rabbits doing unspeakable things in the margins of sacred texts.
In some way, scientists continue the line from medieval scholars. We scribble in the margins. We annotate. We argue with the text. Though we are not permitted to add marginalia to our papers.
Marginalia resolves this problem :)
Get in Touch
Have a question? Have a claim? Want to join as an expert? Drop us a line:
[email protected]Made with care in Barcelona.