When your attention span is shorter than a TikTok video, how do you get through a 30-page paper?
The Problem
I study AI for a living and I still catch myself reaching for my phone mid-paragraph. If you have ever re-read the same sentence four times because your brain drifted off, you are not alone.
A 2019 Nature Communications study found that our collective attention span is getting shorter. On Twitter, hashtags stayed in the top 50 for 17.5 hours in 2013 but only 11.9 hours by 2016. The same pattern showed up in books, films, and search trends. We are wired for 15-second clips now, but papers are still long and dense.
The fix, of course, is to read more papers. But how do you do that when your brain is screaming for a dopamine hit every five minutes?
Why “Just Read It” Does Not Work
When I was an undergrad, the advice was always: “Just read the papers.” That is like telling someone to “just run a marathon” without any training. It assumes you have the time and the focus, which is not true for most of us. We are all dealing with a constant stream of notifications, emails, and the pull of social media. Expecting to read papers in that environment is unrealistic.
And what happens if you start reading a paper and realise halfway through that it is not relevant to your work? Do you finish it anyway? Skim the rest? Give up and move on? The “just read it” advice does not give you any tools to make those decisions.
This is why the Three-Pass Method is so useful.
The Three-Pass Method
This method is not new. S. Keshav laid it out in “How to Read a Paper” (ACM SIGCOMM, 2007). Here is a brief summary of how it works — though I would recommend reading the original for more detail.
Pass 1: The 60-Second Skim
Give every paper just one minute. Read the title, abstract, first and last paragraphs of the introduction, the section headings, and the conclusion. Glance at the figures.
After that, you should be able to answer three questions:
- What problem are they solving?
- What is their approach?
- Does this matter for my work?
If the answer to number three is no, move on. Most papers stop here, and that is perfectly fine.
Pass 2: The Proper Read
Set a timer and read the paper properly — but not front to back. Start with the method, then the results, then the related work. Skip the proofs unless they are the main contribution. As for how long: Andrew Huberman explains in his Focus Toolkit episode that our brains operate in roughly 90-minute focus cycles, with a natural dip in concentration at the end. A single 25-minute Pomodoro fits well inside that window; for a Pass 2 you rarely need more.
The goal is to get the story: what they did, what they found, and why it matters. Mark anything confusing with a question mark in the margin. Do not stop to look things up — that is a rabbit hole you will not climb out of.
Pass 3: The Deep Read
This is reserved for papers directly related to your research. Now you go through the proofs, the experimental details, the assumptions. You ask the hard questions: Is this baseline fair? Would this hold up in practice? What are they leaving out?
Most papers need only Pass 1. Some need Pass 2. Very few need Pass 3. This is not laziness — it is triage.
Staying Focused
Now, I am not here to write yet another article about the Three-Pass Method. You can find that everywhere. What I actually want to talk about is how you adapt it in the age of TikTok. The following bits of advice are based on my own experience and may not work for everyone, but they are worth a try.
Multitasking Is a Myth
“But Ilia, my [insert senior figure here] told me I need to multitask to be efficient! I need to check emails while reading papers, keep an eye on Twitter for new submissions, reply to messages while I read!”
Tell them they are wrong. Your brain can only focus on one thing at a time. When you try to do two things at once, you are really just switching rapidly between them, which tanks your efficiency and increases errors. This is not an opinion; there is a pile of evidence:
- Rubinstein, Meyer and Evans (2001) showed that task switching involves two costly mental stages — goal shifting and rule activation — and that the time lost to switching increases with task complexity (Journal of Experimental Psychology: Human Perception and Performance).
- Ophir, Nass and Wagner (2009) found that heavy media multitaskers are worse at filtering out irrelevant information and worse at switching between tasks than people who multitask less (PNAS).
- Strayer and Johnston (2001) demonstrated that talking on a phone while performing another task impairs performance as much as being legally drunk — even with a hands-free set (Psychological Science).
- Sana, Weston and Cepeda (2013) found that laptop multitasking during lectures harms not only the multitasker but also nearby peers who can see the screen (Computers & Education).
Your Dopamine Baseline Matters
Before we talk about environment, it is worth understanding why sitting down with a paper feels so painful in the first place. Huberman explains in his episode on dopamine that if we regularly expose ourselves to high-dopamine activities — social media, short-form video, endless scrolling — we actually lower our baseline dopamine over time. The result is that low-stimulation activities like reading a paper feel unbearable by comparison. Chris Williamson puts it well: “Dopamine is not about the pursuit of happiness, it is about the happiness of pursuit.” If your brain is trained to expect instant reward, the slow reward of understanding a paper cannot compete.
The practical takeaway is simple: if you are going to block out time for reading, do not spend the 30 minutes beforehand scrolling through TikTok or Twitter. You are sabotaging yourself before you even start.
Eliminate Distractions
If you want to read papers properly, you need an environment that supports it. Everyone reads in different places and at different times, so I am not going to tell you to go to a library if you hate libraries. I work in an open-plan office and I read there. But here are a few things that have helped me:
- Noise-cancelling headphones. A game-changer. But be careful with music — a 2022 study in the Journal of Cognition found that music with lyrics interferes with cognitive tasks, particularly reading. Instrumental music is less harmful, but the evidence is mixed. What I have found helpful is white noise or 40 Hz binaural beats — Huberman recommends these as a way to support the neurochemical conditions for focus (mainly acetylcholine and dopamine). Five minutes before a reading session can genuinely help you lock in. But do not use headphones as an excuse to put on your favourite playlist.
- Full screen. When you open a paper, make it full screen. No email, no Twitter, no other tabs open.
- Phone on Do Not Disturb. Face down, or better yet, in a drawer. Ward et al. showed in their “Brain Drain” study (Journal of the Association for Consumer Research, 2017) that the mere presence of your smartphone reduces available cognitive capacity — even if you never touch it. Out of sight, out of mind.
Dealing with the Sheer Volume
Let us be honest: there are too many papers to keep up with. arXiv gets over 20,000 submissions a month now. You need help filtering.
My go-to is AlphaXiv. It works directly on arXiv papers — you can ask it about specific sections, get explanations of equations, or get a quick summary. Unlike ChatGPT or Claude, it is grounded in the actual paper text, so it does not fabricate things. I use it during Pass 1 to decide if a paper is worth reading, and during Pass 2 to quickly understand tricky sections.
Semantic Scholar’s TLDR is also solid for one-line summaries when triaging.
But here is the thing: AI is for filtering and clarifying, not for replacing your reading. If you only ever read AI summaries, you will end up with a shallow understanding of your field. You will miss the details that actually matter. Use AI to decide what to read, then read it yourself.
Management and Note-Taking
It is not enough to read papers — you need to remember them. Otherwise, what is the point? Ebbinghaus’s forgetting curve — replicated and confirmed in 2015 — tells us we lose roughly half of newly learned information within an hour if we do nothing to retain it. Without a system, most of what you read will vanish within days.
Every paper that passes Pass 1 goes into a reference manager. I use Zotero, but there are plenty of good alternatives. That is just the first step, though. You also need to take notes.
While reading, I annotate the PDF with highlights and comments inside Zotero. However, my main notes live separately. I keep a digital notebook in Obsidian where I write down the key points of each paper I do a Pass 2 or Pass 3 on. This is where the real thinking happens — connecting ideas across papers, comparing approaches, and reflecting on how it all relates to my own work.
I have tried every note-taking app going. The best system is whichever one you will actually use. That insight came after countless wasted hours experimenting with different tools, methods, and formats. So I will not tell you to “start simple.” Get your hands dirty and figure out what sticks.
Building the Habit
Start with one paper a day. That is a manageable goal. It is not overwhelming, but it is enough to build momentum. You can do Pass 1 on one paper, and if it is worth it, carry on to Pass 2 — or do Pass 1 on another one instead.
Reset between sessions. If you have just finished a draining Pass 3 or your focus has collapsed, do not reach for your phone. Go for a walk, stare out the window, or just sit quietly for ten minutes. The point is to give your brain a genuine break — not to swap one screen for another. As Chris Williamson says: “Most people do not need more information. They need more execution.” You already know what to do. The hard part is actually doing it instead of doomscrolling between sessions.
Find your best hours. Our cognitive performance is not constant throughout the day. A systematic review in Sleep and Breathing found that short-term memory and selective attention tend to peak in the morning, while long-term memory and processing speed are better in the afternoon. Crucially, it also depends on your chronotype — a Frontiers in Human Neuroscience study showed that night owls are significantly impaired during morning hours compared to early risers. So the answer is not simply “read in the morning.” The answer is: work out when you are most alert, and protect that window for reading.
Be like David Goggins. If you do not know who David is, close this page and go find out. He talks about “callousing your mind” by doing hard things regularly. Reading papers in the TikTok era is hard. But if you make it a habit, it gets easier. Your brain adapts. The discomfort fades. And one day you will sit down with a dense 30-page paper and realise you actually want to read it.
Wrap Up
I am not going to pretend that social media has been entirely good for the way we think. There is something quietly corrosive about training your brain on 15-second loops for hours a day and then expecting it to sit still for a 30-page paper. But we cannot wish these platforms away, and I am not sure we would want to — they are also how we discover new work, connect with collaborators, and stay in touch with our fields.
The point is not to fight the reality we live in. It is to adapt to it. The Three-Pass Method gives you a framework for deciding what deserves your attention. The rest — creating an environment that supports focus, using AI to cut through the noise, building a system for what you read — is about protecting that attention once you have decided to spend it.
You do not need to read every paper cover to cover. You do not need to understand every equation on the first pass. You just need a system: a way to filter, a way to focus, and a way to remember.