I thought this was useful and I'm interested to see what the next parts are.
My feedback is that the post would be improved by skipping the disclaimer and introduction. The post is already long and spending 1/5th of it on meta commentary will likely cause some people to bounce off before they get to the content.
It also feels kind of like a bait and switch because the post isn't about photographic memories, but maybe you're building up to that?
You also don't need to worry about being new to the site in general. The kind of new-person content that tends to get downvoted is usually crank theories or people being condescending.
And yes, I am building up to it haha. The Solakios technique is my own contribution to the discourse and will come about in [Part V]. I'm trying to explain how I got there before just giving the answer away. I think if people see my thought process, and the research behind it, they'll be more convinced by the conclusion.
I think when dealing with something like 'photographic memory' which is a highly sought after skill, but has not actually been taught (those 'self-help guru's' have poisoned the idea of it) you have to be systematic. People are more than justified in being critical of these posts until I've justified how I got there.
If you want people to see this, you need to tell them up front what they'll get if they spend their time reading it. There's so much good content, and we all have finite time.
I skimmed this post, since I've seen treatments of all of these topics. It looks like a useful summary and reminder.
What's missing is any discussion of the payoff: how well do these techniques work? Even a vague estimate of the payoff for effort spent would be useful.
I've never invested substantial time in any of these techniques (after finishing my PhD in cog. psych) because it's unclear if that's time well-spent. Do mnemonic techniques work better than ad-hoc study for getting important things done?
Excellent students use these memory techniques. But student excellence is usually measured by tests for which memorization works very well. Do excellent researchers or professionals use these mnemonic techniques?
It's a genuine question. It's not my area of interest but I don't ever remember seeing any good evidence either way.
Good points. I'll try cover some of this in my final post. I unfortunately haven't tested this outside of my field, so it'll be difficult. But I assure you, I will try.
Mastery of research, memory and logic is key to rebuilding a metaphorical Library of Alexandria. The first asks: how do we find knowledge? The second asks: how do we remember the knowledge we find? The third asks: how do we apply that knowledge accurately? I will be dealing with the second question in a series of three (3) posts.
INTRODUCTION
@TurnTrout in How do you feel about LessWrong these days? [Open feedback thread] — LessWrong posits two (2) things to do when making a post. First, to outline your claims in the beginning. Second, to keep your post concise and understandable. I tried my best throughout this post to follow these warnings. I also read both @Eliezer Yudkowsky books from Rationality: From AI to Zombies in the hopes that I avoid some of the more common biases and fallacies. Correct me if I've failed there. Thank you, moderators!
DISCLAIMER
You will be able to understand this post even if you haven't read any of the ones I link.
I encourage readers to interact as they see fit. I am open to alternative ideas and developing this. I am genuinely eager to learn from the LessWrong community.
This trio of posts will be divided into five (5) chapters; orthodox techniques, unorthodox techniques, memory modules, application, and the Solakios technique (yes, that's what I'm going to call it). The use of images below will be explained briefly in [Chapter I, Part 2], then more extensively in [Chapter IV].
For those familiar with memory, you'll find tidbits of new knowledge spread about this post. If not, [Chapter IV] and [Chapter V] will be most beneficial. If you're in a hurry for an upcoming test, exam or thesis I'd suggest skipping to [Chapter V].
CLAIMS
I want you to leave this post able to devour the sacred texts of your field more easily, and more effectively. I want to guide you to techniques that enhance your ability to compete with artificial intelligence. Thus, my claims are:
This is a guide to developing photographic memory, not eidetic memory What is the difference?
It means that you will still be able to recognise patterns and understand how concepts are related. Comparatively, those with eidetic memory often suffer the fate of remembering the whole puzzle but having no idea how individual pieces came together to form it.[1]
This is a practical step-by-step that can be applied immediately What does that entail?
I am not going to give you a history of every mnemonic paper ever published. I am certainly not going to define every definition in the field of memory, or even interrogate the philosophies endorsed by different textbooks. I am creating this post for two reasons. First, as a more practical answer to Memory Improvement: Mnemonics, Tools, or Books on the Topic? — LessWrong. Second, as an alternative solution to the problems posed by Significantly Enhancing Adult Intelligence With Gene Editing May Be Possible — LessWrong. The latter post is obviously well written, and well researched. However, the solution proposed is well beyond the means of most LessWrong readers. The solutions contained in this post are not.
CHAPTER I: ORTHODOX TECHNIQUES
PART 1: THE BASICS
YouTube, school workshops and common sense may have introduced you to the 'basics'. These are techniques that are well known, researched and are applied by most top performing students. Society's familiarity with them means that I will spend only a small portion of this post summarising them.
(1) CHUNKING
Three (3) things. First, extract key pieces of information.[2] Second, file those pieces of information into three A's (acronyms, acrostics, and associations.)[3] Third, visualise those groups and apply them. The purpose of chunking is to make cohesive blocks of data that your brain is more likely to deem important, and therefore remember.
It is very likely that you've been chunking without knowing it. For example, you don't remember 929876132856. Typically, you remember it in trios: 929 876 132 856.[4]This is a fake number, please do not try and phone it.
There is also some evidence to suggest that chunking is improved by placing items, places and people (in that order) under hierarchies or themes.[5] An example below:
Mr Dawson emailed a typed will to his lawyer Morgan Terka two months prior to his death. One month prior to his death, he wrote out a will and stored it in a locked vault at SuperSecure Industries. A week prior to his death, he sent another typed will to Miss Terka which contradicted the first will he emailed. However, this will never reached her. A day before he died, he wrote out instructions to his friend Amanda. In these instructions, Amanda must go on an easter egg hunt to figure out the code for the vault at SuperSecure Industries. You are hired by Miss Dawson to make the court recognise the written will.
Extraction four (4) important people, three (3) different wills, four (4) dates.
Three A's the first letter of each name must be taken; Jack Dawson, Mrs Dawson, Morgan Terka, Amanda. From this, the acronym 'JAMD' is born. Then, the first letter of each will can be assigned more vulgar or interesting meanings, making them easier remember.[6] From this, the acrostic 'Fudge Wild Liars' is made. Last, taking into account what I mentioned earlier about hierarchies and diagrams, we can create an association the brain deems important. The dates can be placed in descending order, and on a horizontal timeline 'Two months -> one month -> one week -> one day'. Note that the three A's do not need to be applied simultaneously. You might prefer to apply acronyms exclusively since there are less steps involved there, than in acrostics.
Visualise & apply the entire paragraph of facts has been liquified into the essentials. This can now be stored in your brain as three (3) groups. Feel free to put each group in visual mind maps, diagrams or boxes to aid with recall.
I am aware that my example has been designed to demonstrate a point. The real world will not be as simple or convenient. Regardless, chunking has been able to hold its weight across a wide range of contexts, and with different kinds of information.[7] This claim stands taller when chunking is combined with other techniques, which will be discussed later on in [Chapter IV]. For those of you who are curious as to how this all works on the neurological side of things, follow the attached footnote.[8]
(2) ACTIVE REHEARSAL
I'm calling this 'active rehearsal' as opposed to its more commonly known names. The reason for this is that 'active recall' is actually dependent on multiple elements to be effective. Said elements are outlined below:
The Forgetting Curve: Herman Ebbinghaus held that we forget almost 50% of what we learned in the hour after, and 80% a day after.[9] This has been confirmed in several studies, most convincingly the 2011 one.[10] Ebbinghaus also identified three (3) things that affect our 'retention rate' or how much we remember. First, how much importance we attach to the thing we're learning (less important obviously means lower retention rate). Second, the way in which the information is presented (more logical, coherent and neat means higher retention rate.) Third, how we feel when absorbing this information (sleep deprivation, and hunger.)[11]
Spaced Repetition: means to counteract the effects of the forgetting curve. You should review your notes (mental or physical) for the first time at least a day after. The second review comes three (3) days after, and the third review comes seven (7) days after.[12] This process can be finetuned using tools like Anki which I'll talk about later in [Chapter III]. Briefly, these tools allow you to choose options (e.g. easy to remember, difficult to remember, did not remember at all). The options selected use algorithms like SM-2 to automatically calculate the next best period for revision.[13]
Rote recall: this is repeating (or what some youtubers call blurting) as much as you can remember immediately after learning a topic.[14] A study from as far back as 1989 demonstrated this. Students were asked to close their textbooks immediately after reading and to blurt out as much as they could remember.[15]
Cornell Notetaking: one aspect of this is to create your own questions from the notes written. It trains your brain to confront blind spots, and to recognise the information as important. Your brain is also more likely to store that information in long-term memory because it has been told to.[16]
When you combine these four (4) elements, you're left with a seriously effective study technique all on its own. To summarise, my combination is called 'active rehearsal' since it involves; awareness of the forgetting curve, spaced repetition to counter it, immediate usage of rote recall, and using the Cornell format.
If you wanted, you could stop reading here and do very well in your studies just through this technique. You'll probably still need some of that good old discipline too though.
(3) FEYNMAN TECHNIQUE
The Feynman technique has several moving parts. I'll break these down first, so that you understand why the technique works. Following these, I'll explain how to apply it.
Double loop learning: if you've ever watched Dead Poet's Society, this follows Mr Keat's philosophy. Students should understand the why's, how's and purposes behind learning before being asked to do it. This equips them with the awareness to reinvent solutions to problems instead of brute forcing a fail state approach.[17]
Heutagogy: this builds on double loop learning in that the student has a far more active role in the classroom. They then strive to teach themselves, consult with more experienced peers, and negotiate content in the curriculum.[18]
Constructivism: proposes that effective learning requires analogising (using what people already know) to synthesise memorable ideas to achieve new outcomes.
Autodidactism:taking learning into one's own hands.
Feynman combines these four concepts. Students who adopt Teacher roles are forced to use double loop learning when explaining something. Further, they automatically borrow simpler explanations and personal experiences to explain a concept. Steps 1-5 are:
Tickets please: identify the topic being learned, write the headings and sub-headings of the content for that session (our brains seem to enjoy being debriefed on subject matter before being forced to learn it.)
Teacher mode: explain it in the simplest terms possible (layman's terms or atomism). Aim to do this with study groups who can provide feedback.
Patchwork: following your roleplay as a teacher, do two (2) things. First, identify areas where you could've dumbed it down more or explained it better. Second, point out the things you failed to remember. (Both of these are harder to do since you have to be aware and keep track. I advise using voice recordings if you don't have a group.)
Reconstruct: put in a conscious effort to use feedback, honour the source material, and optimise the overall explanation.
Retrieval: repeat these steps consistently until you can't atomise your explanation any further. Package this explanation, formalise it and store it in memory for use in tests or real-world problems.[19]
I know some youtubers and teachers have turned these five (5) steps into a single A4 page with headings for each. I think this defeats the whole point of the technique (refer back to the concepts that make this technique work.) There's plenty of evidence to suggest we understand and remember things more when we build our own models.[20] I am essentially doing just that by writing this post.
As a general rule, if a technique takes a lot of work to remember, then there's no point in using it. It's pointless to have to remember the tools you're using to remember. Keep the models simple or automate it if its effectiveness outweighs the complexity. Then use it consistently until it becomes intuitive to how your brain functions.
Feel free to make a note here. In many places, you could upgrade a technique using artificial intelligence. For example, having a digital face occasionally nod and squint while you teach it (a bit dystopian I know, but it's an example.)
A gentle reminder that I'm keeping things as short as possible. You're always welcome to investigate these techniques or go further than I did with certain hints and suggestions I've left thus far.
(4) STORIES
I personally don't believe this technique can be applied to real world scenarios due to its lengthiness, and how incompatible it is with longer form topics. Regardless, I'll be summarising it here because when it is used, it is very effective.[21]
You're told to remember the directions to the airport. You are to go North for two hundred meters, passing a red stop sign and the WhaleCoffee shop along the way. Then, turn left at the pink train station and continue on for 10km until you hit a dirt road. Once at the dirt road, turn left for 2km until you reach a yellow bus. Once there, go East for 15km. Along the way, you'll pass the Oil docks. Following that, you'll have arrived.
The story method involves linking all of the directions in a ridiculous story. The more ridiculous, the more memorable. Try the exercise right now, using your own mental footnotes. I've attached below how I'd go about remembering them:
I see Northern lights but they're red along the way, and somehow whales that are drinking coffee are flying too.
One of the whales ends up falling on top of a train station, and it is left behind.
The train takes the whale, which is 10x the weight of the train, to a dirt road.
At the dirt road, the whale flies back up into the air and follows a school bus with children who call out to it.
The bus and the whale end up having so much fun that they travel all the way to a sea made of oil.
As you've probably noticed, this works with lists and simple pieces of information. But it would be very difficult to do with longer, more complex essays unless other techniques are used too (acronyms, symbols etc.)
This technique can be improved significantly through rooms and expanding rooms. I've linked an explanation in the footnote if you're interested.[22]
PART 2: THE NOT-SO BASICS
To recap: we've covered four (4) relatively well-known study techniques; chunking, active recall, Feynman technique, and stories. In the spirit of what I've been teaching, you'll notice that I've ordered these 1-4, and used an acronym 'CAFS'. Remember what I mentioned earlier about vulgarity? You could order this to 'FACS' too.
If you wanted to take this further, try and recall the images under each heading. For chunking, I used four (4) circles that were linked. Next, a question mark and a seemingly active neural pathway. For the Feynman technique, I used what hopefully looks like a teacher who's into gaming. And quite obviously, I used a book for the story method.
My aim is that if what I'm preaching here and later on is backed by science (which seems to be the case) the use of symbols, hierarchies and the colour red will simplify how much you have to remember about remembering. I'll talk more about symbology and the use of colour in mnemonics later on (I promise this 'later on' nonsense will become less frequent, it's for a good reason.)
The next three (3) techniques are lesser known, and far more complex. However, they deal with longer term memories, higher-order thinking and are geared towards the lifelong learners among you. Before we move there, I'll need to outline a sort of arbitrary system for different 'categories' of techniques:
Dormant: these study methods only come to life for specialised contexts such as memory competitions or remembering the letters of PI.
Observe: 'modules' that make the use of memory techniques easier, and more efficient. Think of how much better you study in a clean room. Techniques in the observe category are like that, but for cleaning your brain prior to forcing it to remember something. More on this later in [Chapter III].
Livewire: these are quick mental models you've used so frequently they've become intuition. Typically, simple enough to be applied to any situation such as remembering people's names, or the contents of a lecture that's only going to be given once by an academic celebrity.
Library: these are the majority of study methods and what students, or lifelong learners use to assemble knowledge into retrievable packages for regular use. All four (4) methods under [Chapter I, Part 1] fall into this category.
(1) SPONGING
This is the first technique in the 'observe' category I mentioned earlier. I've linked the video it is based on in the footnote.[23] Might also be easier to remember the name of this by calling it 'soaking' if vulgarity has been working thus far for you.
The technique structures a topic in your brain, preparing it for the material before engaging with it. Essentially, getting your brain wet. It works as follows:
Skimming: go quickly through each page in the chapter, noting the positions and sizes of paragraphs and diagrams.
Outline: if there's questions or a table of contents somewhere in the chapter, read that. See the Baader-Meinhof Phenomenon to understand why this works.[24] In fact, you may have noticed that I've left an outline of this entire post in the 'Memorandum' section. Feel free to apply this newfound knowledge now.
Emphasis: go through the chapter again, making note of the bold words; concepts, headings, captions.
Placement: actively read the first and last sentence of each paragraph. If the textbook or topic has been structured logically, it'll ask a question in the beginning and answer it in the end. This gives you a good idea of what's important.
Final reading: read the entire chapter, and actively take written notes. I'd advise against highlighting, but if you do, use a red highlighter and draw a border around the important sentences as opposed to colouring in single words.
This leaves you prepared to talk about the topic. From this point, you can simply extract the sentences in the borders, or the pencil notes taken and use it as cement for other techniques. If you're a real goat, you can just use 'active rehearsal' on the topic after.
(2) MIND PALACE
We've arrived at probably one of the most famous memory techniques, and by extension, one of the most misinterpreted. There are very few people who actively, and correctly use this technique. Hopefully, I'll change that.
The mind palace, or 'method of loci' requires several 'observe' things to take place before it can be effectively used. I've outlined them below:
Familiar memory: the 'palace' you will use to store important ideas, or retrieval packages can't just be your childhood home. It must be a place you've travelled through so many times, so intimately, that you can recall the smells, feelings and other senses when walking it. This takes a lot of practice. It takes dreaming about it, meditating about it and actively walking through it consistently.
Spatial relationship: there should be an association between the palace and the topic that's being studied. For example, if a doctor is using this method, their palace would be a hospital near them.
Journey: the items, and topics that are being stored in the palace must be connected in some way. Typically, the wielder of this technique will travel through their palace in a clockwise direction, finding each piece of a puzzle that needs to be remembered. Even better if these are numbered or connected by senses.
Personal: the effectiveness of this method is dependent on how personal the palace is. Once you've practiced it extensively, you can use an online 3D modelling programme to help imagine the place, one that doesn't exist in reality.[25]
Dominic O’Brien took this so far that he created an entire 'memory town' with each building representing a particular topic. A calculator store for remembering random numbers, a dictionary store for remembering certain languages.[26] There have also already been studies regarding virtual environments and the use of AI to fast track this technique.[27] I'll be discussing how to enhance it even further in [Chapter IV].
I'd advise taking a break here and applying the concept. Take a nearby book, flip to a random chapter and see how many of the concepts mentioned there can be attached to the objects in your room. From there, create a mini palace within your room and see how much you can remember. I've linked a far more extensive explanation on how to construct a memory palace in the footnote.[28] Though what I've described here is sufficient unless you plan on using this technique exclusively.
(3) ZETTELKASTEN
The Zettelkasten is another example of an 'observe' technique and is mainly for storing information. Think of it as a physical or digital wiki of all the things you've ever learned that allows both panoramic and microscopic views of subject matter.
There are three (3) elements that make this technique effective:
Hypertextual: zettelkasten emphasises the organic and webbed connection between ideas. As a result, every idea should follow into the next in a logical manner. This is done through both a 'reference list', and 'tag words'.
Atomicity: each note must contain only one (1) idea. There can be 'hubs' which break up into multiple of these, we'll discuss this shortly.
Personal: use a style of language, and patterns of thought unique to you. If informal language helps you study, use it for the entire web of knowledge.[29]
Now, onto the anatomy of the zettelkasten technique, or how it actually works:
Hub: a table of contents with all the main headings for each topic. For example, 'artificial intelligence' would be a main heading. There may be three (3) branches that lead out of it titled 'AI Alignment', 'AGI', and 'AI Research'. Each of these branches may break into ten (10) or more notes or atoms that are restricted to single ideas.
Unique identifier: the atoms should have a token to recognise them by. It's useful to put in a date, the topic, and a symbol or word that was significant at the time of inserting it into the zettelkasten tree.
Body content: the bulk of the atom is a summary; the absolute essentials of the topic being learned. If you're dealing with a branch (those sub-headings that come after the main heading) you can give a summary in there of every idea it breaks off into. For example, if I was looking at the 'AGI' branch, there would be a summary in there of the ten (10) AGI solutions, and their titles. If I wanted more information on each solution, I would go to their atom which explain the idea in more detail. This is why the zettelkasten is compared to a wiki.
References: at the bottom of the atom are footnotes, or the cement used to create that atom. Link websites, or other atoms here (especially if it's your own theory or idea.)[30]
To consolidate: a hub is the GPS for your tree of knowledge containing all the main headings for broad topics and ideas. These main headings lead to branches or sub-headings, and these sub-headings lead to atoms which explain the ideas themselves.
These atoms must contain 'active knowledge' which can be immediately applied. I'll go over why a lot of these study techniques seem to have a 'problem-oriented approach' in [Chapter IV].
For more information on how to build your own zettelkasten, and the nuances behind it, check out the footnotes linked in this part.[31]
MEMORANDUM
Right, we've finally reached the end of [Chapter I]. In the spirit of what I've been trying to teach, I've inserted an outline here of everything covered (recall the sponging technique.)
I hope you've found this first post beneficial. Depending on how much interest it gets, I'll upload the subsequent two (2) posts. I also want to gauge whether I've actually written this post correctly. If I haven't (which is likely) I'll have to do a lot of touch-ups, and your feedback would give me an opportunity to do that, so thanks!
Basics (well known, well researched, used by most students.)
Chunking (four (4) red linked circles.)
Extraction, Three (3) A's, visualise & apply.
Active rehearsal (red question mark with neural net.)
I reiterate the general rule here: there's no point in using a technique if remembering it takes a lot of work. Keep your techniques simple, and personal. Otherwise, automate it where its effectiveness outweighs its intrinsic complexity. Use it consistently until it becomes second-nature.
W Xie, W Bainbridge and others Memorability of words in arbitrary verbal associations of memory retrieval in the anterior temporallobe 29 Jun. 2020 Nature Human Behaviour; N Castro Your mental dictionary is part of what makes you unique -- here's how your brain stores and retrieves words 7 Nov. 2023 The Conversation; C Bergland The Neuroscience of Recalling Old Memories 3 Jul. 2015 Psychology Today.
E Reyes Feynman Technique as a Heutagogical Learning Strategy for Independent and Remote Learning Dec. 2021 Recoletos Multidisciplinary Research Journal; S Tamm Feynman Technique: A Complete Beginner’s Guide 18 Jan. 2023 estudent.org.
E Legge, C Madan and Others Building a memory palace in minutes: Equivalent memory performance using virtual versus conventional environments with the Method of Loci 2012 Acta Psychologica.
1
