Update (December 2025)

Someone left a comment regarding when the book will come out, and instead of leaving a long response in the comments section, I figured it warranted a proper post:

I'm really sorry about the delay, but there is one really important hangup. To summarize what has happened over the last many months: I over-invested in a mnemonic technique that was based SOLELY on leveraging the power of AI image generation, and the results were not what I had expected.

The core idea is that I thought that using a good mnemonic system (Image style/vibe = mnemonic meaning), you could use bespoke AI generated images and that ALONE would guarantee strong memory traces. After all, aren't pictures worth a thousand words? I thought so, but it turns out that this probably should not be the CORE of a memory system. I REALLY thought it was the "silver bullet," to be honest. Keeping that in mind, AI images can be a very powerful SUPPLEMENT to a good mnemonic system.

Since I basically rewrote my "global mnemonic system" to accommodate for AI images, after the failure of this system, I had to decide what to do. I had two options: 1. Use the same mnemonic system that I was using previously (Which I had been using for many years) or 2. rewrite the system while making adjustments in light of what I had learned. Although it is time consuming, I opted for option 2. Think of it as when Microsoft failed with Windows Vista or Windows 8 so they course corrected with Windows 7 or Windows 10.

When I first started creating an organized mnemonic system years and years ago (Which at the time was mainly for Japanese character readings), I didn't TRULY know what I was doing, I was just learning as I went along. Essentially this entire blog is an exercise in self-experimentation, and I simply post my results here. Over the years I have learned what mnemonic systems are easier to work with compared to other systems. For example, source material WITH linear scenarios and progression seem to be more effective/faster to associate mnemonic meaning than source material WITHOUT linear scenarios and progression. So for example, a Tarantino movie or a TV show that features many flashbacks will not work as well as something with a straightforward beginning, middle and end. Not that stories with non-linear source material are IMPOSSIBLE to use with an organized mnemonic system, but it seems like it requires more work to create and maintain.

These are the kinds of useful insights I want to articulate and put out onto the internet so others in the same position as me can benefit from my failures. I like to think of myself as a cartographer that is exploring a mostly unexplored and windy cave. Many passages lead to dead ends, but it's still worth it to KNOW that it is a dead end. I don't want to say "This works, I recommend you do this too" unless I know that FOR A CERTAINTY it ACTUALLY works. When I started experimenting with AI images, I thought FOR SURE this was going to change the world of mnemonics, but instead it ended up being a great SUPPLEMENT somewhere in the middle. I have a pattern of doing this with many things, actually. Heck, this same cycle happened to me with SuperMemo itself. I started using it, became WAY too obsessed with it, and eventually found a balance in between the two extremes of NOT using it and using it TOO MUCH. I think I have found this balance when it comes to the application of AI images in mnemonic systems, and I'm in the process of proving this to myself. I'm not sure how much work is remaining until I feel like I can put my "stamp of a approval" on the system, but I don't want to recommend it too early because I don't want to mislead anyone.

I don't want to say "I did this for 2 weeks and it was useful!" I would rather say "having tried system x, y and z thoroughly, I can say absolutely system x is the way to go."

So again, sorry for the delay! But as a famous nerd once said:

Update 2025 (Book is nearing completion, AI is useful, Possibly figured out IR for iPad/iPhone)

Hello! How are you doing? Hopefully well! Here are some updates on my end:

MY LIFE

My mom is doing great! She recovered physically and her speech is pretty much back to normal! Her ability to enunciate certain syllables was the one thing that was notably lagging behind, but we did daily speech therapy for about 10 minutes over FaceTime.

I used ChatGPT to come up with tongue twister stories that featured syllables that she had a hard time saying. I was also able to choose the direction of the story so that it included family inside jokes! I would recommend using ChatGPT (Or any other compatible AI tool) to come up with sentence exercises, especially for pronunciation. Rather than reading the same stories day after day (Which we did for a couple of weeks), by choosing the subject matter, you can keep the sentences fresh! 

Outside of that, I can’t think of much happening in my life. Oh yeah, when I went back to the United States to help out my mom, I was able to go to my first concert in like 10 years! I saw Alvvays, a really happy and catchy indie shoegaze-y rock band. Their discography is small but each album is filled with great songs! Definitely check them out!

BOOK

The book itself is pretty much done, but there are a few holdouts that I want to MAKE SURE work before I say “hey, try this!” I hate when I see people talk about the system they have been using for only a few days, I want to know what system has been working for MONTHS or YEARS.

My two main holdouts are: 1. Various applications of AI in Incremental Reading and visual mnemonics and 2. Incremental Reading substitute / supplement for the iPhone / iPad

AI IS USEFUL FOR LEARNING

AI continues to be a big source of hype; much of it is unfounded, but in terms of learning, it is quite the game changer. For me this has been mainly in two ways:

1. You basically have a tutor that you can contact at any time of the day and can explain things to you that you do not understand. You can also ask if the flashcards you make are accurate (This has been a big help)! There are many various applications of AI with SuperMemo (and specially incremental reading), but I will get into that a bit more in the book.

2. AI Generated visual mnemonics. Although it sounds like I’m beating a dead horse, the quality of images that various AI models can output continues to astound me. Aside from the many real world (And meme) applications that pop up across the internet, highly detailed mnemonic images makes “encoding” memories SO MUCH FUN! For the purposes of SuperMemo, I have made about 7,000 images so far, and I can recall them mostly with very little effort.

IPHONE AND IPAD INCREMENTAL READING

My other project has been trying to figure out a low friction way to make a simple imitation of incremental reading on my iPhone or iPad.

While I would love for there to be an official app that allowed for incremental reading, unfortunately there isn’t one. I have spent the last few weeks playing around with a simple system that imitates the most useful aspects of incremental reading. It is very simple in terms of features (I don’t want to pay a subscription fee for something) and it seems to be working very well. I have incrementally read and analyzed a number of articles and books. I also wanted a way to “capture” my conversations with ChatGPT, since the responses are often dense with useful information that I would like to “break apart” like I would a topic in SuperMemo.

Assuming that it keeps being useful for me, it will occupy another section of the book.

If you have any questions or stuff you want an answer to before the book comes out, let me know and I’ll try to answer it in the comments section!

Enjoy your day, keep on learning!