How to Learn Faster Without Any of the Hacks

You learn faster by doing four things in sequence: encode the material so your brain has something to work with, retrieve it from memory before you forget it, space those retrievals out, and link the new knowledge to what you already know. None of this is fast in the moment. It is fast over weeks. The hacks are not.

I tried the hacks. All of them.

Through uni and the years after, I tried most of the things you'd expect. Speed reading, where I trained my eyes to skip across the page faster and faster. Memory palaces, where I tried to attach concepts to imagined rooms in places I knew. Spaced repetition. Active recall. The one I never touched was Anki, because I knew enough about myself to know that any tool which requires daily upkeep on top of the thing I'm trying to learn becomes another thing to drop.

The honest version is this. Speed reading was the most expensive lie. It felt productive in the moment, my eyes were moving, the pages were turning, but when I tried to remember anything afterwards I remembered less than when I'd read normally. The understanding was thinner. Memory palaces actually worked when I used them, but I forgot to use them. The technique was in the toolbox. The toolbox was in a cupboard I rarely opened. Spaced repetition stuck. Active recall stuck. Both of those, I still use today.

Then I read the actual research, the studies the hacks were trying to compress, and noticed something. The studies kept pointing at four boring fundamentals. Encoding. Retrieval. Spacing. Linking. That is the whole game. The hacks were either repackaging one of those four with a personality, or selling something that bypassed all of them and quietly did not work.

This piece is the four moves, the research behind each, and how to put them together. It is not a list of tips. It is a method.

Why Don't Learning Hacks Actually Work?

Most learning hacks fail because they optimise for the part that feels like progress, not the part that produces memory.

Speed reading is the cleanest example. You can train yourself to move your eyes faster across a page. Your comprehension drops in proportion. Past about 400-500 words per minute, comprehension collapses to the level of skimming, which means you are not reading any more, you are scanning for keywords. You finish the book sooner. You retain less of it. The clock says you learned faster. The recall test says you did not.

Memory palaces work, but only for what they were designed for: ordered lists of distinct items. Deck of cards. Sequence of capital cities. Phone numbers. The moment you try to use one for a body of conceptual knowledge, the system breaks down, because concepts do not live in fixed positions and they connect to each other in ways a palace cannot store.

Highlighting feels productive and produces nothing. Dunlosky and colleagues' 2013 review in Psychological Science in the Public Interest, which compared ten common study techniques across hundreds of studies, rated highlighting and underlining as low utility. Same with re-reading. Same with summarising. The two techniques rated high utility were practice testing and distributed practice. Both involve effort the others avoid.

That is the pattern. The hacks that feel easy are the ones that do not work. The fundamentals that feel hard are the ones that do.

Bjork's name for this is desirable difficulty (UCLA, 1994). The conditions that slow down apparent learning often produce better real learning. Conditions that smooth the learning experience often produce worse real learning. The brain stores what it has to work to retrieve. It throws away what arrives effortlessly.

The moment this clicked for me was uni. I crammed for first-semester exams. I aced them. Then second semester started, and the new modules were built on the things I'd just been examined on. Whatever I'd crammed was already gone. I had the marks on a transcript and nothing in my head. The course was building, the lecturer was building, and I was building on a foundation that had vanished. That was the lesson the speed-readers don't get to: you can't compress the part where understanding settles into memory. You can only space it out and come back.

What Does the Research Say About Learning Speed?

The research says learning speed is set by four mechanisms, and you can move all four.

I want to give you the numbers, because the numbers are what changed my mind. Most "how to learn faster" pieces are vibes and analogies. The actual data has been consistent for forty years.

Move	Mechanism	Evidence	Effect size
Retrieval (testing yourself)	Active reconstruction beats passive review	Roediger & Karpicke (2006), Washington University, 120 students	g = 0.50 to 0.61
Spacing (distributed practice)	Forgetting and re-retrieving builds stability	Cepeda et al. (2006) meta-analysis, 317 experiments	g = 1.01 (large)
Desirable difficulty	Effortful retrieval > effortless recognition	Bjork (1994); Wilson & Shenhav (2019), Nature Communications	Optimal at ~85% accuracy
Elaborative encoding (linking)	New facts connect to existing schema	Chi (1989); Dunlosky (2013) review	Moderate (g = 0.40 generation effect)

The numbers above are not soft. Spaced retrieval has one of the largest effect sizes in cognitive psychology. A meta-analysis covering 317 experiments converging on the same finding is about as strong as evidence gets in the social sciences.

What is striking is how few people actually use these. Most students still highlight and re-read because that is what they were taught and because it feels like work. The methods with strong evidence feel uncomfortable, which is why they are still rare. The discomfort is the signal, not the bug.

Now, the four moves.

Move 1: Encode. Why How You Take It In Determines What You Remember

You cannot retrieve something you never encoded properly in the first place. Encoding is the layer most learning advice skips, because it is invisible. It happens before any "studying" begins.

Reading a paragraph while half-distracted does not encode it. Your eyes move across the words and your auditory loop sounds them out, but no schema forms in long-term memory. You can prove this on yourself: read a page of a non-fiction book while thinking about something else, then close the book and try to recall the page in your own words. Most people get a vague gist and almost no specifics. The page never made it past working memory.

Good encoding has three features:

Attention without competing input. No two-screen reading. No podcast in the background. The brain shares working memory across inputs, so a competing channel halves what you can take in.
Comprehension before memorisation. If you do not understand a sentence, do not try to remember it. Find what is unclear and resolve it. Memorising a sentence you do not understand is the most expensive form of forgetting.
A reason to care. The brain prioritises information tagged as relevant. Asking "why does this matter to me?" before reading a section is one of the cheapest interventions in the literature.

The cleanest practical move at the encoding stage is the pre-question. Turn the heading of the next section into a question and write it down. Read to answer that question. This activates a search frame in working memory, which means you read with intent rather than letting the page wash over you.

A 2025 PMC study on metacognition found that pre-testing, even when the learner gets the answer wrong, primes encoding so strongly that the post-reading recall test improves by a meaningful margin. You learn more from a passage by trying to answer questions about it before reading than by reading it twice. That is a strange finding the first time you encounter it. It is also one of the most replicated results in the literature.

Pre-questioning is also one of the cleanest ways to lower extraneous cognitive load before reading; see cognitive load theory for why this matters.

Move 2: Retrieve. Why Highlighting Doesn't Count

Retrieval is the move that does the heaviest lifting in the entire system, and it is the one most people skip.

Roediger and Karpicke at Washington University ran the study that should be on every classroom wall. In 2006 they had undergraduates read short prose passages and then either re-read them or take a memory test. Five minutes later, the re-readers performed slightly better on a recall test. Two days later, the testing group had pulled ahead. One week later, the testing group remembered 61% of the material; the re-reading group remembered 40%.

Same passages. Same study time. Different activity. The activity made the difference.

This is the testing effect, and it has been replicated in hundreds of studies since. The mechanism is straightforward. Retrieval forces your brain to reconstruct the information from internal memory traces, and that reconstruction strengthens the trace itself. Re-reading lets your brain recognise the words, which feels like remembering and is not.

For the underlying decay curve this fights against, the forgetting curve explained covers Ebbinghaus's data and the modern replications.

The trap is that retrieval is hard, and the harder it is, the more it works. Bjork's law again: the effort of retrieval is the active ingredient. A successful retrieval that took you twenty seconds to dig out builds more memory than a successful retrieval that took two seconds. Apps that make you tap "easy" on a flashcard for instant recognition are quietly removing the part that does the work.

Three retrieval moves that work without any apps:

Closed-book recall. After a chapter, close the book and write everything you remember on a blank page. Don't summarise. Try to retrieve. Then check what you missed.
Question-on-the-margin. Convert each section heading into a question, then test yourself on it days later without the source.
Teach it. Explain the concept aloud as if to someone who has not read the source. Gaps in your explanation are gaps in your memory and gaps in your understanding. They tell you exactly where to go back.

Highlighting is not retrieval. Re-reading is not retrieval. Watching a YouTube summary is not retrieval. If you have not closed the source and forced your brain to construct an answer, retrieval has not happened.

Move 3: Space. Why Forgetting Helps You Remember

The counterintuitive move at the centre of the whole method is this: you have to let yourself forget a little before you retrieve, or the retrieval does not work.

Cepeda et al.'s 2006 meta-analysis of 317 experiments on distributed practice found an effect size of g = 1.01, which is large. Spaced study sessions consistently outperform massed sessions for long-term retention. The mechanism is partial forgetting. When you retrieve something you almost forgot, the act of pulling it back forces the brain to update the memory's stability. When you retrieve something you definitely still know, almost nothing happens.

Cramming feels productive because everything is on the surface and you can recall it immediately. That feeling is exactly the problem. The information is in working memory, not consolidated long-term memory, and most of it will be gone within 48 hours of the exam.

The simplest spacing schedule that works:

Pass	Timing	Purpose
First retrieval	Within 1 hour of learning	Catches the steep part of the forgetting curve
Second retrieval	24 hours later	Confirms the trace is consolidating
Third retrieval	3-7 days later	Stabilises into long-term memory
Fourth retrieval	2-3 weeks later	Builds durability
Subsequent	Monthly, then less	Maintenance

Each successful retrieval pushes the next one further out. Each failed retrieval pulls the schedule back in. This is the principle behind every spaced repetition algorithm from SM-2 to FSRS. The algorithms are useful, and the principle works without them. A note on a calendar with four future dates does the same job.

The point is not the precise schedule. The point is that you space the retrievals at all. A study group that did all its review on day one performed worse one month later than a study group that did the same total review distributed across four sessions. Same content. Same minutes. Different placement. The placement is what made the difference.

Move 4: Link. Why Isolated Facts Don't Stick

The last move is the one that turns memorisation into understanding. A fact you cannot connect to anything else has nowhere to live in your head. A fact that connects to ten things you already know will be there for a decade.

This is elaborative encoding, and the foundational paper is Chi's 1989 study at the University of Pittsburgh. Chi watched students working through physics problems with worked examples. Some students explained each step to themselves as they read. Others just read. The self-explainers learned dramatically more, even though they read less material. The act of generating an explanation forced them to integrate the new information with what they already knew.

The same effect shows up in the generation effect literature: when learners produce information rather than read it (fill in a blank, generate a synonym, complete a sentence), they remember it 22% better. Hit rates in the original study were 87% for generated items versus 65% for read items. The brain pays more attention to information it had to manufacture itself.

What this means in practice is that after you read something worth remembering, you stop and ask:

How does this connect to something I already know?
What does this contradict?
Where would this be useful?
If this is true, what else must be true?

These questions feel like they slow you down. They are the part that locks the new knowledge into the existing structure. Without this move, every fact lives alone, and lonely facts are forgotten first.

The cleanest example I have is from uni: dimensional analysis. I could pass the exam questions reliably. Give me a problem, I'd rearrange the units, produce the right answer, move on. I remembered the moves. I never understood what dimensions actually were, why they mattered, or how to spot a result that was dimensionally wrong outside the textbook context. Retrieval worked. Linking failed. The knowledge sat in a room by itself with no doors. I didn't know it was missing the link step at the time. I just knew, years later, that none of it had become useful.

The other thing that helps: linking across sources. When you read three different authors on the same topic, the overlap between them is the schema. The author who said it first taught you the fact. The author who said it second taught you that the fact was important. The author who said it third put it in the structure of your understanding. This is one of the reasons people who read widely in a field appear to learn faster: they are not faster, they are better-linked.

How to Put the Four Moves Together

Each of the four moves works alone. The compounding happens when you stack them.

Here is the sequence in practice. You read a chapter or article (encoding). You close the source and write down what you remember on a blank page (retrieval, immediate). You ask yourself why this matters and what it connects to in your existing knowledge (linking). You schedule a 5-minute review for tomorrow, another for next week, and one for a month later (spacing). When the review comes round, you do retrieval again, not re-reading.

That is the entire method.

The honest cost of doing this is about 25% more time per chapter. The honest payoff is about 3-5x the recall a month later. I cannot give you a precise number for your specific case because the multiplier depends on what you are reading and how long-term you want the memory to be. Roediger and Karpicke's 61% versus 40% gap at one week is a fair guide. The longer the delay, the larger the gap.

Here is a comparison of what the four moves change relative to standard reading habits:

Habit	What you do	Recall after 1 week	Recall after 1 month
Standard reading (re-read)	Read once, re-read sections	~25-40%	~10-15%
Highlighting	Mark important passages	~25-40%	~10-15%
Encode + retrieve only	Pre-question, closed-book recall	~55-65%	~25-35%
Encode + retrieve + space	Add 24h, 1w, 1m review	~70-80%	~50-60%
All four moves	Add elaborative linking	~80-90%	~65-75%

Numbers are approximate, drawn from the Roediger & Karpicke and Cepeda et al. effect sizes applied to typical reading material. Your mileage will vary with content difficulty and prior knowledge. The directional gap is the point.

This is the part where I should mention Alexandria, because the four moves are exactly what it does in the background. You read a piece in Alexandria. It extracts the knowledge worth remembering, prompts you for elaborative encoding at save time, spaces the retrievals using FSRS (the modern version of SM-2), and surfaces them on the right day at the right interval. You do the encoding and the retrieval. The system handles the spacing and helps with the linking by surfacing connections across what you've already saved.

But you do not need Alexandria to do this. You need a notebook, a calendar, and the willingness to close the book and write what you remember on a blank page. The method is older than any app. The apps just remove the friction.

The thing I want to leave you with is the realising moment. I spent years trying to learn faster and I was optimising the wrong variable. Speed of reading is set by physiology and has a hard ceiling. Speed of learning is set by encoding, retrieval, spacing, and linking, and has no ceiling at all. Move those four levers and the same hours produce a different person at the end of the year.

The brain changes after the work, not before. You become a faster learner by doing the four moves until they are automatic. There is no shortcut. There is just the method.

Frequently Asked Questions

How can I learn faster for an exam in two weeks?

Stop re-reading. Spend 70% of your study time retrieving from a blank page and 30% looking at the source. Space your sessions across the two weeks rather than stacking them at the end. Roediger and Karpicke (2006) found students who tested themselves outperformed re-readers by a wide margin on a one-week delayed test, and the gap widens further as the delay grows.

Does speed reading actually make you learn faster?

No. Comprehension and reading speed have a hard ceiling that the research has confirmed for decades. Speed reading tradeoffs are well-documented: past about 400-500 words per minute, comprehension collapses. Learning speed is set by how the information is encoded and retrieved, not how quickly your eyes move across the page.

What is the fastest way to memorise something?

Self-test on it within an hour, then again the next day, then again three days later. This is spaced retrieval, and it produces the largest documented effect size in learning science (Cepeda et al. 2006, meta-analysis of 317 experiments). It feels slower than cramming. The recall data tells a different story a week later.

Why do I forget things so quickly after learning them?

Because your brain prunes anything you do not retrieve. Hermann Ebbinghaus showed in the 1880s that people forget about half of new information within an hour and 70% within 24 hours, without review. The forgetting curve flattens only when you actively pull the information back out, not when you re-read it.

Are learning hacks like memory palaces and mnemonics worth it?

They work for specific tasks like memorising a list of unrelated items, the order of a deck of cards, or vocabulary. They do not work for understanding a concept, a system, or a body of knowledge. The fundamentals (encoding, retrieval, spacing, linking) carry every form of learning. The hacks are accessories on top of those fundamentals.

How long does it take to learn something new?

It depends on what you are trying to do with the knowledge. Recognising it again takes a single exposure. Recalling it cold a week later takes about three to five well-spaced retrievals. Using it to think with takes longer, because that requires linking it to other things you already know. The four-move method shortens each stage.

How do I learn effectively for exams?

Convert every chapter into questions and self-test on them across multiple days. Mix topics rather than studying one at a time (interleaving). Sana and Yan (2022) found interleaved retrieval practice produced 63% on a one-month delayed test versus 47% for the no-quiz control, with only 10-12 extra minutes per week. The technique is specifically built for exam-style recall.

Is highlighting a good way to learn faster?

No. Dunlosky's 2013 review of ten learning techniques rated highlighting as low utility. It feels productive but does not require any retrieval, which is the part that builds memory. If you highlight, use the highlights later as cues for self-testing. The mark on the page is not the work. The recall later is.