Using AI to (maybe not so) quickly create elements for adding to a memory palace or anything else

[Suthnautr (David J Curtis)]

I got myself a paid $20 a month subscription to ChatGPT a few months ago and found it super useful for creating logically organized progression outlines for blog posts, legal proceedings, etc that I could flesh out myself.

So I found out that I can instruct ChatGPT to “render” images (the word render makes it more lifelike - if you say “Draw” you’ll get black and white line drawings).

So what I do is I tell ChatGPT to “render” an image and it does a fine job - BUT (BIG BUT) - if I make the request too complicated it will screw it up no matter how many time I clarify.

That’s why I write Maybe “not so quickly”

To fix that what I now do is I always ask for ONE main scene element in my Mnemonic, as the primary element to be memorized. From there I (you) have to create individual elements to the scenes you’re directing in your Mnemonic (memory palace station, aka “loci” etc) to gain the proper complexity to entirely encrypt what you need to memorize.

Here’s an example of a completed multiple step ChatGPT encryption rendering.

What you’re looking at is a ChatGPT multi-step process. The Nun in the canoe playing with her Rosary Beads is ONE separate image. The MOOSE and the Treehouse Bar are actually several different images combined - so total 4 ChatGPT renderings combined in 3 different apps (all done on my phone) to create this ONE drink recipe graphic… but FASTER than drawing it all out and with MORE DETAIL and ALMOST 3D realism! Yay! It’s so cool!

Here’s How:

I create the first image order: “Render” treehouse bar on a hill and a moose driving a Model T Ford to the bar. simple one thing - treehouse bar on a hill, moose driving a car. It SORT of did what I said (this was the best rendering - multiple tries)

THEN a new rendering of a sign that says “Bar Hill” (it’s a Vodka) on a clean white background.

THEN a new rendering for a Pompadour Toupee (on a clean white background) …THAT wasn’t easy - multiple attempts at a simple request - trust me.

BEFORE THIS - I had to draw these images by hand, in flat 2 dimensions, often flakey looking and hard to modify so I “settled” for what was acceptable.

…which you can tell isn’t quite as vivid to recall in great detail.

I don’t think you need to subscribe to ChatGPT to use the “render” feature, it may be free, and I know that Elon Musk has already released GROK 2 which is SUPER PHOTO PERFECT image rendering capable (I have a subscription to that already but haven’t explored it fully)

BUT for everyone out there not using AI yet - For SPEED of Mnemonic encapsulation and encryption tasks - AI rendering is the KING.

EDIT - Addendum:

Most of us live on our phones today. They’re ultra powerful, ultra portable, awesomely functional mini-computers - and, with a Bluetooth keyboard and a stylus - they’re CRAZY powerful any-time, any-where MONSTERS of productivity.

That said (Mnemonics related) I use multiple apps in conjunction with the ChatGPT renderings.

The most useful is (Android) “Background Eraser” by handyCloset Inc. and handyCloset Inc.'s other app “PhotoLayers-Superimpose,Eraser”

I also use one that’s been just this month discontinued that’s called HandyPhoto - which will only run if I put up my firewall or go into airplane mode (since it can’t find the license any more, and won’t let me proceed if I’m connected to the internet).

My point though is that if You’re going to use ChatGPT rendering (or Elon Musk’s Grok/Grok 2 engines) you will want to speed things up and avoid frustrating time delays by editing the generated images yourself, “by hand” using apps you control directly, as I do. Good luck!

[RMBittner]

I know a number of others here have posted similar graphical mock-ups of their memory images. But, to be honest, I don’t understand what the point of this is, or why you’re going to all of this work.

When I create a memory journey or a scene for a memory palace, I just write down what’s happening—so I have a record I can easily refer to—as I’ve imagined it. What does having an artistic rendering provide that makes it worth the time you’re investing in creating it? (Especially when it doesn’t seem to suggest the specific path you’re taking through the information.)

(Minor side note: The AI misspelled cocktail. Unless that was intentional for some memory purpose.)

Bob

[Suthnautr (David J Curtis)]

As you wrote

@RMBittner

When I create a memory journey or a scene for a memory palace, I just write down what’s happening—so I have a record I can easily refer to—as I’ve imagined it.

That is true, but you must read it, where all I have to do is take a glimpse at the picture if I cannot see it in my rehearsal (long term memory practice).

You also wrote

@RMBittner

I don’t understand what the point of this is, or why you’re going to all of this work.

The work is up front, and intense, but for me an intensely interesting and entertaining immersive study session. It pays for itself in saved review time over the useful life of the memory, usually years, especially when the thing memorized is rarely needed - let’s say a recent true life event where this one proprietary drink I made just one time 2½ years ago was suddenly called for about a month ago as one of the drinks to be made by several bartenders at one of our wedding functions. Because of the elaborate encryption in the image I created years ago I “saw” the whole thing in one scene - quantities and all: Collins glass, ice, ¼ oz (7½ ml) Simple Syrup, 2 (muddled) slices of cucumber, 1½ oz (90 ml) Citron Vodka, ¾ oz (45 ml) St Germaine, 3 fresh basil leaves, shake and dump (meaning do not strain), garnish with a lemon wheel.

There are no written sentences to go over. Review is fun and easy, drink making/execution is easy, decryption is easy, and I can usually rattle off the full recipe without any hesitation. The several bartenders I now work with now come to me for recipes they no longer can locate in our book - same way it was when I worked at the NY Hilton (which had 7 front and service bars) before the internet, and they’d call my bar phone for memorized recipes several times a month.

Anyway… back to that cocktail last month for the wedding… the Beverage Director says “we’re making the ” and I smile and say “Oh yeah, that’s …” which I rattled off like a machine gun, with full precision. The Asst Bev. Dir. is beaming - big smile on his face - and says “See? He knows his recipes.”

As for the AI’s misspelling - yeah, it’s not worth asking it to correct the mistake, because it then re-renders the entire image and almost always botches up several other things in the process.

BUT, RM, that said, for ME the AI renders are a breeze to work with and save a ton of time and uncertainty on the back end…

CONTAINER OBJECTS

And, one more thing that’s super important that written notes alone can’t replace, is my drink recipe memorization method using what I call container objects. I group items in objects to which I have assigned specific values. Which item goes in which container object depends on its particular liquid measure.

For example, let’s look at the Model-“T” again for a moment. That’s obviously a “1”. Anything IN the Model-T is a 1. A moose? (pamplemous)…is an ounce of Grapefruit. The Hulk sitting next to the moose? …is an ounce of Lime Juice. The Model T roof is also a great platform for putting a canoe on a boat rack - which another container object with the ½ oz value. So put a monkey in the canoe along with John Belushi dressed as a killer bee on SNL and now the recipe also contains ½ oz (15ml) Simple Syrup and ½ oz Honey Syrup… BUT I’m not finished withthe Model-T yet! In the back is the “Trundle Seat”, also called the Mother in law seat. That jas it’s own separate value assigned by me as 1½ ounces (45ml). I stick my friend Henry in there. Now the drink also gets 1½ oz of Hendrick’s Gin… and so on, and so on - Model-T good, move on in the image to the next container object whatever that may be! Complex, but elegant, and snap recall when properly rehearsed.

It won’t be for everyone. If all you need to memorize is lists of things, meanings, etc then this is overkill. This is industrial for the beverage industry - cocktails, akin to pharmaceuticals and baking, require precise ratios in addition to ingredients, glassware, method, garnish, sprinkled topping, and rim (sugar, salt, tajin, crushed graham cracker, crushed pretzel, etc)

[beau2am]

Yes. I very much like what you’re doing with your container objects. Thanks for the detailed examples!

[Suthnautr (David J Curtis)]

@RMBittner

Especially when it doesn’t seem to suggest the specific path you’re taking through the information.

Sorry for this late response - yes, I perfectly understand what you mean - there is no order in mixology. Many say go with the least expensive ingredients first in case of a mistake, other schools say go with the base liquors first. In actuality it depends on the cocktail and the bartender’s knowledge of how to make the drink. Muddled ingredients usually go first with me, but that depends on whether the drink has more than one such ingredient - cucumber and mint, let’s say. Muddle the cucumber thoroughly - but only bruise the mint, and that can be added almost last and gently “shake muddled”.

The order doesn’t need to be left to right as long as all sets of ingredients are visible. When I write out my recipes on index cards though I always start with highest liquid quantity to least liquid quantity so I can see the groupings more easily.

How I arrange the recipe items in my container objects, and how I position the container objects in relation to one another is purely to make the story line more logical and easy to remember.