Hello! Im Ted Shachtman, I have degrees in cognitive science, education, and computer science, and I just left my job (working for jeremy clarkson) to work on sharing and teaching a novel mnemonic technique. If you find this interesting, I’d be unbelievably happy to chat and I’m free all day.

Okay, so

4 months ago my friend asked me “Do you have any component to your thinking that isn’t imagery of one of the 5 senses, and if so what’s the utility of it?” In my search to answer his question, I found a very interesting answer that offers a whole wealth of very useful mnemonic techniques (that I can’t find on the internet) for people who say yes to my friend’s question (and more specific versions of it). In this post I’ll describe enough of it so that you can guarantee your interest enough to dive into the resources to learn the full thing— and then hopefully we can chat about it!

You can read about the technique at the website i made, thoughtfeel.net, and have published several youtube videos teaching the mnemonic techniques (all on my website).

So, to do that poster-board work, Here’s the description of it from the website:

From the website, not typing this on my phone

“Some people have it and some don’t. It’s like the mind’s eye: some people can picture things vividly, and some, with aphantasia, see nothing at all. If you have thoughtfeel, you can imagine in concepts, not just words or images.

What is thoughtfeel?
thoughtfeel is a sense for concepts. It lets you take in a concept and its properties all at once, the way vision lets you take in what’s in front of you.
You could describe a dog in writing: it has fur, a wet black nose, and it’s about the size of a capybara. Or you could look at it, and vision gives you all of that at once, with no list.

Concepts have properties too, including properties you could never see. Take the economy. You could write out what’s true of it: millions of people, making and trading things, sometimes booming and sometimes in a slump. People with thoughtfeel can still make that list in words. But when they use thoughtfeel, the sense takes in properties like these all at once.

That makes comparing fast. Show someone photos of two dogs, and they can tell right away whether it’s the same dog. If it isn’t, they can look again and point to why: one has a brown tail, the other a white one. With thoughtfeel, the same works for concepts you imagine. If two concepts aren’t the same, that’s clear right away, and paying attention to each one shows which properties differ.

Like vision and hearing, it differs from person to person. For some people, taking in and imagining concepts is the default way they experience the world. For others, it’s much quieter. Just as people can imagine in pictures or in sounds, people with thoughtfeel can imagine in concepts.

Is it really all at once?

Taking in the concept is. Putting its properties into words comes after. When a friend walks into a room, you know it’s them right away, before you could describe a single feature. If someone asks what color their eyes are, you look again and say. thoughtfeel works the same way. Someone with thoughtfeel who thinks of a promise knows right away what it is. Then they can pay attention to it and put its properties into words: it’s made to someone, it’s about the future, and it can be broken.
With practice, people notice finer properties, the way a birdwatcher learns to spot details in a bird that look the same to everyone else.

Mnemonic Techniques (not from website, me typing on phone)

My lesson’s explore a bunch of them, but thoughtfeel is generally really useful. I’d compare it to if 20% of people had a minds eye but society never talked about it. Some people would discover the keyword technique to learn vocabulary, and they’d just be better at learning words than everyone. A rare bloke would discover the memory palace and be quite good at memorizing grocery lists.

That’s where we are with thoughtfeel.

For instance, my friend Henry, who’s brain auto imagines in thoughtfeel, can read a board game rule book back to front and comprehend it. Me and all my friends just thought he was a genius until i interviewed him about thoughtfeel and found it that’s what he was doing. Then, in my head (where thoughtfeel doesn’t auto build, i have to do it on purpose), i tried “visualizing” the rules in thoughtfeel as i read and… boom! I could do it.

Another family friend could remember a whole song from one read of the lyrics, and she had no idea how she did it. I interviewed her, and she was turning each sentence into a thoughtfeel “image”. I tried it, and now i can too!

But, there are a bunch of techniques where all the people i’ve interviewed go “huh… yeah i never thought of doing that.” These, are like the memory palace POA system to remember 1k numbers— crazy unintuitive unless you know what to do and people are talking to each other.

The language learning and prospective memory techniques (you can check out the videos to learn them) are one’s like these… quite unintuitive unless you’re thinking of thoughtfeel as an extra sensory modality where you engineer weird mnemonic techniques.

Anyway, I’m looking for more people to chat about this with. It was interesting enough for me to leave my job and work on this full time, but tbh that was because i bet that people like u guys would find this interesting. If you don’t, then that would be very a good signal that i should get my job back, so please tell them.

A friend asked our group a question a couple of months ago: “Is any component of your thinking non-sensory?”

Beyond your inner voice, mental images, or imagined sounds, is there something else that thinking feels like?

Some friends immediately recognized the experience. Others didn’t. For me, thoughts have always seemed to have something I can best describe as a texture. That conversation made me start paying attention to it. I’ve been calling it Thought Feel.

For example, consider a line of people stuck behind someone who has stopped moving, then boxes backing up behind a jammed box on a conveyor belt. Once I understand both situations, I can attend to a similar felt texture in the two thoughts.

Recognizing that both situations involve blocking is easy to describe in words. What interests me is that, for me, the shared relationship also has a recognizable feel, beyond the words and images I use to represent it. “Texture” is my way of describing that felt quality.

My friend’s next question was: “Can you actually do anything useful with that?”

I didn’t have a clear answer. I’ve spent the last two months reading related work and experimenting with deliberately attending to these textures.

Related work includes Hurlburt and Akhter’s research on unsymbolized thinking, which describes thoughts experienced without words, images, or other symbols, and Eugene Gendlin’s Thinking at the Edge, which develops ideas by working with felt meaning. Thought Feel, as I’m describing it, can also accompany words and images.

My practical interest is in developing a palate for the textures of abstract relationships: distinguishing them within complicated thoughts, recognizing them in unfamiliar situations, and using them to bring analogies to mind.

The chef analogy helps explain the practice. You can know tomato sauce is in lasagna while struggling to pick out its flavor. Comparing lasagna with pizza gives you another way to attend to the shared ingredient. I’m exploring the same operation with thoughts: compare concrete situations, attend to a shared texture, then return to the original thought and recognize that component within it.

I’m putting these exercises into a practical user’s guide. I’ve made two animated videos:

1. Start here: Introduction to Thought Feel — what I mean by a thought’s texture, and how to attend to it so you can recognize it again.
2. Distinguishing the ingredients of a thought — comparing different situations to isolate a shared texture within a more complex thought.

I’ve found these practices useful in my own thinking. I’m sharing them as exercises to explore; their effectiveness hasn’t been established through controlled research.

Does this description match anything in your experience? Do you already attend to it deliberately? I’d especially like to hear concrete examples, differences from your experience, and relevant work I should read.

I think modern AI in ttrpgs is a slam dunk, and its awesome to see you building something here. Gonna look into it more!

Throughline is an AI tool that helps human GMs run tabletop RPG sessions. It heavily uses modern AI (hence posting here), and does not replace any humans.

While you're at the table running the game, Throughline listens to your session live and generates scene-beat storyboards (small grids of images showing what the players would encounter if they make a choice) that you can glance at and parse quickly. It also tracks campaign canon across sessions, plants and tracks callbacks, and proposes opening narration when you start a new arc. Throughline does not narrate to your players, run combat, or appear at the table. Players never see anything it produces. The GM does all the live performance: voicing NPCs, improvising, reading the room. The job of Throughline is to handle the long-horizon planning so the GM can focus on running the table. We're at pre-alpha. We've done 6 live playtests plus a lot of internal testing. One-shots have been reliable. Multi-session campaigns are less proven, so we'd suggest starting with a one-shot.

We're opening access to about 10 outside GMs to use it for their own sessions and give us feedback. The fit we're looking for is GMs who are strong on the social side of the table (improv, NPC voices, table feel, in-the-moment narration) but who either don't have time to prep extensively or don't have years of practice at long-term narrative planning. If you're already a great GM who enjoys prep and does it well, Throughline probably isn't for you.

The product is a web app. You sign in with Google. There's no GitHub or terminal setup. You can run a homebrew world by giving it the lore, or a setting you already love from commonly known books, games, or shows. You'll need a payment method on file because we forward LLM API costs at cost (no markup during alpha). In practice that works out to about $0.50 per hour of live play, so a weekly three-hour session runs around $6 to $10 per month. There's a trial for $5 that should get you a beefy 1-shot. If you can’t pay the API prices, i can cover you, but it’s out of my pocket.

There will be bugs. We want testers who find that interesting rather than frustrating, and who are willing to be in active conversation with us. Design feedback is the main thing we want; we're not looking for early customers or business partners. If you have an eye for game design, that's especially welcome.

About the developer: I'm Ted Shachtman, an educator and software engineer. I play Fabula Ultima and D&D, and GM both. The reason I'm building Throughline: a friend of mine, Ben, is a math PhD and the best GM I've played with. He preps three hours per session, voices a dozen NPCs, plans coherent arcs in large worlds, and adapts brilliantly on the fly when the players do something he wasn't planning for. He moved away, and the next best GM in our group is me, and I'm not very practiced nor have the time to prep. I built Throughline so I could be a better GM. We're trying to raise the floor for people who can't prep the way Ben does, so they can still run a session worth playing.

If you're interested, you can read more about the system at our website (link in comments) and sign up for the waitlist. The site has a longer writeup of how the system works and the design behind it. I'll respond to everyone within a few days.