Dear Reader,

This week's newsletter is about Obsidian mistakes that most of you (and I) are likely guilty of. I talked to the community, looked back at my own journey, and identified the 13 mistakes you need to avoid to build a healthy information management system for your research.

But why do you need an information management system, Obsidian and all that?

I did a personal literature review of the exploding number of AI-powered publications in recent years and found that most of the growth comes with clumsy use of AI and poor quality. To put it cynically: Most people don't know how to use AI in any way other than to produce slop.

There is a huge opportunity if you know how to use AI not only to generate text but to enhance your thinking, and I don't think many people have figured it out yet.

I explain this below and in the webinar with Avi Staiman on Nov 22nd. Please join us if you have time and want a solution to academic AI slop.

Let me know what you find useful and interesting,
Ilya

EA Tutorial: 13 Mistakes to avoid as an Obsidian user

My original article, The Absolute Obsidian Beginner Tutorial, turned out to be one of the most visited pages on the site. So today I want to contribute something in the same vein: The top mistakes people do, when starting with Obsidian.

For this article, I analysed what the EffortlessAcaedmic community was saying and looked at the top questions, as well as the numerous conversations I had with you over the courses and my personal mistakes over the years. Here is the result:

The 13 mistakes are:

1. Overengineering folders, properties or tags.
2. Installing too many plugins and not checking if they are actively supported
3. Copying vault structure that you don't understand (and often need)
4. Putting notes on one paper into one note
5. Highlighting PDFs and copy-pasting from papers
6. Writing down notes without mentioning the source
7. Creating one note for every small thought
8. Using AI to generate notes or summaries
9. Using folders to find notes
10. Relying on memory to find notes
11. Quitting too early: At ~50 connected notes, something magical happens.
12. Not re-reading or modifying your notes
13. Importing and maintaining old notes or files

The article guides you through why they are so problematic, how to avoid them and how to fix them if you have made them already.

AI is flooding journals with slop: How do you stay relevant?

You have likely heard that AI is exploding the number of publications. The real problem, though, is declining quality and an overloaded peer review system. Journals have reacted by leaning into the preprint model (src) or by punishing careless AI use, such as hallucinations, more harshly (src). These policies have so far not proven effective (src).

The problem of ever more publications, however, has existed for centuries, and technology has kept pace. In 1869, journals replaced books as an easier way to access new findings (src). Then, in the 1970s, decades before the internet, digital library systems made searching for ideas easier (src). In the 1990s, the internet and Google made access faster still (src).

The way we access and process information has kept pace with how fast we generate it. Writing, ideating and thinking, however, have remained largely unchanged until now.

Today's problem shows up in surprisingly clumsy AI mistakes, such as hallucinated references (src) or predictable sentence structures that we now recognise as AI (src).

» In other words, many people try to cheat with AI, while nobody has yet figured out how to use it productively in research, even though we all know we need to adapt somehow.

I believe the solution is to use AI to enhance thinking and ideation, not just writing, by using AI as a curator of our knowledge. Just like in a museum, the curator is responsible for presenting the most relevant exhibits for a given goal. AI can do this for our knowledge and ideas, giving us instant access to relevant parts of our thinking, helping us work on bigger problems than before, instead of producing slop.

I will discuss how this works in my upcoming webinar with Avi Staiman. Sign up if you haven't yet (over 320 people are joining). Everyone who joins will also see how to get started implementing AI in this way.

Want to leave Effortless Academic forever: Unsubscribe from everything. I am sorry to see you go, and farewell.