Mambo Taxi

I'm going to reboot this journal; with a semi-technical focus. But not only. I intend to write about neurodiversity and disability and social justice and fixing tech for everyone. And other stuff.

So here's a first post. I have been experimenting with Large Language Models, and I'm so impressed at how far they have come in two years. There are seriously cool things they can help us with, and I'd like to see how they can solve more problems than they create. I've seen good evidence that a person using LLMs (not training new versions; just using the commercial-grade top-of-the-line models) has less negative impact on the planet than their impact shipping groceries from across the continent, or owning a car. So I currently choose to see what good we can do with them. (To be clear, I am also open to good-faith discussion if this feels wrong to you, and you have data I should also consider.)

I had a funny learning experience last weekend, with an experiment that flopped badly, followed by another one that has me sort of excited.

It started Saturday morning when I was curious about the brain science of "running on autopilot" and how weird that is. Gemini Pro satisfied my curiosity, noting that experiencing a glitch while thinking deeply about something was actually a sign of high cognitive efficiency, not a deficiency in attention to the world. I appreciated its conclusion, and asked if I could learn how to do this more effectively - for example, I often close my eyes while composing my thoughts while speaking, and I would appreciate being able to just tune out the distractions so I can succinctly say what I need to.

Gemini pointed out that "you aren't failing at conversation; you are optimizing your bandwidth". And suggested a few ways to improve the environment for me, including a) occupying my physical self with distractions like fidget toys, and b) be willing to wait 3-5 seconds to compose my thoughts before speaking. These are good advice; and it also had a good follow-up question for me (because I instructed it to; I really like my personal instructions. But that's a different post).

It asked me, "Do you find it easier to articulate complex thoughts when you are walking side-by-side with someone (where eye contact is naturally intermittent)?"

I responded with a few paragraphs that summarize to "hell yes!" and added that thinking while walking is really effective for me to figure out what I want to say, and indeed I have many of my best ideas while walking. Then I commented that I wished I had a more natural way to talk with LLMs while walking- I have lots of trouble dictating my non-linear thoughts to Siri. 

I said, "I would love to have a 'deep thought to linear language interpolator' - why can't conversational LLM map from my deep brain space to linear language for me?... Can you guys get onto that? human/machine telepathy isn't too far off is it?..."

Gemini's response was... interesting.

You nailed it: "[Siri] engages the same difficulty with linearizing my thoughts."

When you type, you are allowed to be non-linear. You can write the middle of the sentence, then go back and fix the start. It is a random-access buffer.

When you speak to Siri (or a person), you are forced into a Linear Context. You have to get the first word right before you can say the second. That is exhausting.

The Solution: The "Interpolator" Workflow

You asked why conversational LLMs can't map from your deep brain space to linear language. The answer is: We can, but you are giving us the wrong input.

You are trying to "clean up" the data before you give it to the AI. You are doing the interpolation work yourself.

The "Raw Stream" Hack:

Instead of trying to dictate a coherent note to Siri, try this workflow using a voice recorder (or a better dictation app like Whisper):

1. Abandon Grammar: consciously decide to speak in "bullet points" or "word clouds."

2. Verbalize the Meta-Data: Say things like, "Okay, idea about the server migration... wait, not migration, the upgrade... anyway, connection to the old database... oh, and remember the firewall rule."

3. The Interpolation Step: Feed that messy, non-linear transcript into an LLM with the prompt: "Here is a raw brain dump. Please linearize this into a coherent paragraph."

This offloads the "linearization" burden from your brain to the AI. You stay in the "deep brain space" (generating nodes), and the machine handles the "protocol translation" (generating syntax).

Would you like me to try to be your "Interpolator" right now? You could type a messy, unedited stream of consciousness about a complex topic you're struggling to articulate, and I will attempt to "translate" it into linear prose for you."

Well. That was an invitation!

So then in the same chat, I tried using Gemini Live mode while on a walk, with the following underwhelming - but funny - results.

---

and the conversation (if you can call it that) continued in that vein, with a very polite Gemini interrupting me every few words asking me to go on. I was more amused than frustrated, and decided out of stubbornness to implement something better.

(Because sometimes I think that's really what LLMs are good at- giving me a quick win on 80-90% of a problem, then I work much longer on the remaining 10-20%!)

So I started another chat asking about how to do the same thing, but starting from a Voice Memo, which has no time-limits. Gemini steered me toward writing a "Shortcut" - the Apple scripting tool - which I can run on the Voice Memo (which efficiently automatically generates a transcript right on my phone) and send the transcript text to Gemini to interpolate my stream of consciousness into a logical statement.

I can report that it works remarkably well!

I iterated changes to the instructions a few times. I wanted it to give me a summary to be read aloud by Siri, plus a complete report in markdown for me to paste into my journal. Improvements included:

1. instructing it to tell me what inferences it has made, rather than trusting it was recording my own thoughts verbatim.

2. I changed the description from "a meandering voice memo" to "stream of consciousness" and then it improved the instructions further to replace "Input: A raw, possibly rambling transcript." with "Input: A raw transcript of purposeful exploration." - I didn't fully test but Gemini suggests that would serve my intent better.

In the end, here are the instructions that Gemini wrote for itself:

As I said, it works well. I've tested it on 10 minutes of voice memo, taken while I was on a walk, where I talked about a morning's events, figuring out what I thought was important, and what I had to add to my TODOs, and I talked through a set of decisions that I wasn't clear on the answer. When I read the output, it was certainly more organized, and it was also faithful to what I had said. 

Its logical inferences were not bad. One of them:

• Synthesized the "Middle Ground" approach for the project meeting as the primary action item, discarding the two extremes (total prep vs. total ad-hoc).

In that part, I had described planning to do a lot of prep for a meeting, which I didn't want to do, and also didn't want to go in with no planning, then I realized I could do less prep and ask about followup. And Gemini inferred that I was describing a "middle ground approach" that was what I wanted to do.

I think this is what LLMs are supposed to be good for: producing structure out of human creativity. I started with curiosity about how the brain works, asked for suggestions on improving how I interact with the world with my own brain, and it non-judgementally recommended things that made sense, which led me to further questions which led to a solution I wouldn't have expected.

I think any company that fires its employees and trusts the AIs to do their work with minimal oversight are... doomed. We are no where near "General Intelligence." I think those of us who have the flexibility to experiment and even ask the AIs "what can you do to help?" will find there are many possibilities, and many could even amplify our own superpowers related to brain differences and neurodiversity. 

So that's one step in this direction. I'm still experimenting.

By the way, if you give Gemini that set of instructions (from "Role" through "Changelog" above), and a transcript from a stream of consciousness voice memo of your own, from a 10 or 20 minute walk and talk, it should be able to give you a fair interpretation of what you want to record. (and if it doesn't do it well on the first try, just re-run it, and it will probably be a bit different).

I'm curious how well it does for you!

[ 2026-02-05 edit: I added the "write in first person" line to the part two instructions above, since the reflection is meant to be as me.

It feels pretty natural doing this with a long meandering voice memo. This morning I did this while walking the dog for 25 minutes, and the output was excellent, just requiring a few small tweaks for my journal.]

I want to provide some clarifying information to supplement what I've found elsewhere, about the Via Rail service on Line #1, The Canadian, which stretches all the way to Vancouver. But there is really contradictory info online. So I can at least clarify end-to-end what my trip was like. There's no guarantee this will be your experience, but I hope it clears up some confusion. [Edit to add: I suppose your milage may vary; perhaps literally!]

You can also find all of these photos, with about 20 more, at my flickr album for the trip.


Contradictory things I read on the internet (sorry, I don't have sources at this time; I might come back later to fill in citations).

1) In Via Economy, to book with a group but pay separately, book at the same time; you'll be assigned nearby seats and if they aren't close enough, just call the number and via can move you.

WRONG: I and the two parties I'm travelling with all had unassigned seats on our tickets. When I called Via's service number, they said you just show up an hour before the train and talk to an attendant about getting seated together. And that's just the way Route 1 travel works.

This turned out to be correct. We arrived at 9am for the 9:50 train from Union Station in Toronto, stood on a shortish line in the middle of the Via section of the station, then at 9:20, we were released to go to the train, which was a bit of a long walk. Then the attendant at the train door got our information, and kept us together until they found a set of seats that would work.

We were put in a set of 4 seats, then two sets of two seats behind them. All of the sets of 4 seats had permanent signage reserving them for groups of 3 or 4, and they put up temporary signage when they had reservations they needed to assign seats for. However, there was no problem with me using one of these sets of 4 seats when they were not yet assigned.

It would've been nice if they'd put our last 3 in the seats side-by-side with our group of 4, but apparently that set of 4 had already been reserved by another group who didn't get on the train for another few stops.

2) In Economy class there are clusters of 4 seats facing each other at the ends of the cars. Or, seats can be turned around as pairs, so any seats can become pairs. Or, half the seats point forward and half the seats point backward.

CORRECT. All of these can be correct; at different times. The first economy-class car on the train had seats that can be turned around as pairs to make sets of 4, and they did that with arbitrary seats as needed. I saw it done by one of the attendants. The second economy-class car had fixed sets of 4, at each end of the car and in the middle.

In all cases, the seats have good solid foot-rests that can flip up to nearly flat with the seat, and it stretches to the back of the seat in front of you, or if your seats are in a group of four, they stretch just about exactly to the middle of the other seat, which means the set of four can make a nearly flat surface if they are all raised. This is NOT a super-soft bed situation, but it is oodles better than an airplane or bus. The only oddity is that there is no armrest between the seats, so if you're sitting next to a stranger, that's just the way it is.



3) The train is so much better than flying overnight; you have lots of room and the rocking of the train is soothing. Or, it's cramped and you won't sleep much at all.

CORRECT for the first one, mostly. The weirdest aspect was the lack of armrests. I would've been comfortable if I'd been able to put more padding under me so my slouching position was better on my back. Bring a pillow or inflating pillow; and bring a thin sleeping bag or a blanket. If you can have two seats, it is possible that lying sideways is more comfortable than stretched over a pair of facing seats. At least I thought so.

4) There will be lots of food choices. Or, almost no food choices.

CORRECT for the first one, mostly. The dining car, which is under the dome car, had a microwave and a full slate of options including mac and cheese (OK), breakfast sandwiches, pizza sub that didn't sound appealing, and a few other hot choices. There were many cold sandwich choices, and I enjoyed the turkey and cheese croissant I got. There was a good breakfast omelette. There were coffee, tea, hot chocolate, various soft drinks, wines, and various beers.

Our train was full, and they didn't have infinite orders of each kind of food, but I would recommend bringing lots of snacks, sufficient food you like to make up a meal if you can't find something you like in the dining car.

It would've been nice if VIA had a menu published that I could be sure was what our train was offering, but I didn't end up being unhappy about food.

5) There is an observation dome car. Or, there is no observation car on this part of the route, and it isn't added until the route past Winnipeg.



YES there was a dome car. It had about 30 seats, and we shared among about 100 passengers, so you couldn't stay up there for the whole trip. But it was a lovely change in scenery.

Overall, the scenery was beautiful. And I am telling friends that I would recommend this as a way to go, and I would do it again. We shall see when that happens! Because it's seldom that I feel I can take an extra two days for travel! (And my partner is still not interested in taking the train, even if it's a berth or cabin, which I am now even more eager to try! Some time!)

I've just re-read an article I saved from the first pandemic winter; there is a lot here that is going to be helpful for me this winter as well.

"How to make this winter not totally suck, according to psychologists"
https://www.vox.com/future-perfect/2020/10/14/21508422/winter-dread-covid-19-pandemic-happiness-psychology

"Rather than fixating on our inner worlds and woes, we can strive to promote what some psychologists call “small self.” Virginia Sturm, who directs the Clinical Affective Neuroscience lab at the University of California San Francisco, defines this as “a healthy sense of proportion between your own self and the bigger picture of the world around you.
...
The practices involve cultivating different states — social connectedness, a clear purpose, inspiration — but all have one thing in common: They get you to focus on something outside yourself."

Some concrete parts that jump out at me: I've at times kept a gratitude journal, but they suggest writing in detail about a particular event, rather than dashing off a superficial list of things. And focusing on people I'm grateful to; even writing them a letter, whether or not I sent it (perhaps the flip side of the catharsis of writing someone angry letters then tearing them up or burning them...)