This morning, outside my daughter’s school, we may have accidentally reduced the meaning of life to two symbols.
When Nadia walked past the window, I drew a heart. It’s one of our little rituals. I’m on one side, she walks past on the other, and the heart means what hearts usually mean without either of us having to say anything. Then a slightly older boy appeared and added a smiling face.
A heart and a smile.
Love and happiness.
That’s it, isn’t it? Or at least, if I had to compress a good life into two symbols, those would be pretty strong candidates.
Funny enough, compression is a good place to start today because my mind seems to have the opposite problem. There’s too much of it.
Yesterday I had technical problems with my recording. Somewhere toward the end, the audio simply disappeared, probably because my microphone ran out of battery. Then this morning I discovered that the little list of subjects I was absolutely convinced I’d written down in To Do had disappeared too.
Or maybe I never wrote it down.
Which is worse.
Because this morning, while showering, I had one of those wonderful streams of thought where one idea connects to another and another, until suddenly you feel like you’ve discovered something genuinely interesting. There’s only one problem with profound shower thoughts: showers are remarkably badly designed for note-taking.
By the time I was getting dressed, I was already trying to reconstruct what I’d been thinking.
And that gave me another thought, naturally.
Maybe I don’t actually have to remember every thought. Maybe I just have to remember the thought that started the route. Like finding the first breadcrumb and walking the trail again.
In this case, the breadcrumb was a LinkedIn post about One Flew Over the Cuckoo’s Nest.
My father watched that movie dozens of times. I watched it quite a few times too, probably because he did. Jack Nicholson, the mental institution, the whole question of who is actually sane and who gets to decide that in the first place.
That brought me to The Armageddon Conspiracy by Mike Hockney, a so-called coded novel I read years ago. I don’t remember the complete plot anymore, which is rather fitting in a post that started with me worrying about memory. What I do remember is the underlying idea that people society dismisses as crazy may sometimes see something everybody else is unable, or perhaps unwilling, to see.
Careful there, Mike.
That’s exactly the territory where I think context matters.
Mike’s Mind is an experiment. This is me thinking out loud, wandering into ideas, testing associations and sometimes saying things before I have properly decided what I think about them. That’s very different from me presenting professionally, developing a product, talking to a customer or simply being Dad.
We all have different roles. Technology sometimes has an annoying habit of flattening all of them into a single searchable person.
Maybe that’s why I increasingly like pulling words apart.
Dissecting Language
The Dutch word I was looking for was ontleden. Dissecting.
I love doing that with language. Breakfast is the obvious English example: break-fast. You are literally breaking your fast. Once you see it, the word almost becomes a tiny sentence.
Dutch does this beautifully with prefixes. Vergeten, for example, means to forget. Verongelukken is much harder to translate neatly. In everyday Dutch, when the news reports that somebody is verongelukt, you normally fear the worst. English has “to have an accident,” but that definitely doesn’t carry the same meaning.
And apparently this is where my brain went from Jack Nicholson to Dutch prefixes.
Welcome to Mike’s Mind.
Large language models made me more interested in these little pieces of language because they too break language into smaller pieces, tokens, and somehow reconstruct meaning from the relationships between them. Not quite the way humans dissect a word, obviously, but close enough to make the association interesting.
It also brought me back to this idea I encountered that language might behave almost like a parasite in the mind. That sounds unnecessarily sinister, but the underlying observation fascinates me: language doesn’t merely help us describe reality. It may influence what we notice in reality in the first place.
I recognise some of that simply by speaking two languages. Dutch Mike isn’t exactly English Mike. Same person, obviously, but there’s a slightly different rhythm, humor and perhaps even a different way of arranging thoughts.
Then you find much more striking examples.
There are languages that encode events differently from English or Dutch. Who knocked over the vase? Or did the vase simply fall? That looks like grammar, but underneath it sits attention and perhaps even blame.
And then there are languages where people orient themselves using cardinal directions rather than constantly thinking in left and right. North, south, east and west aren’t occasional navigation terms but part of ordinary communication.
Think about what that might do to your brain.
If your language constantly forces you to know where north is, you probably pay attention to direction rather differently than I do. Not because you’ve been given a biological compass upgrade, but because your language keeps exercising that way of seeing the world.
Which naturally made me wonder about AI.
Because if language can change how humans organise perception, what happens when an intelligence is essentially immersed in hundreds of languages?
And what if we deliberately taught an AI, or an embodied robot, to organise its world through an absolute directional system instead of the body-centered left-right system most of us naturally use? Would it merely produce different words, or would it develop a genuinely different internal representation of space?
I asked Gemini.
It gave me a very long answer.
So naturally I asked the artificial intelligence for the most human thing possible:
Can you please summarize that in one paragraph?
The interesting part of its answer was the suggestion that a persistent global frame might let a system maintain orientation without constantly translating everything relative to itself.
Interesting. But the bigger question for me isn’t whether that particular idea works. It’s what language does to intelligence in the first place.
If language changes us, and AI is trained on extraordinary amounts of human language, then we haven’t merely given machines our information. We may also be handing them our categories, metaphors, assumptions, jokes, contradictions and all the strange little linguistic shortcuts through which we interpret reality.
Language might therefore be more than training data.
It could be inheritance.
And yes, I realise that sounds rather grand for a train of thought that began in the shower.
But that’s the fun part.
The Language We Use on Ourselves
Language doesn’t only operate between people. There’s also the language running inside your own head.
That became another connection this morning because I’ve been thinking about getting healthier and losing weight. If the people around you keep telling you that you’re overweight, and you keep telling yourself the same thing, does that language eventually become part of the condition?
I don’t mean that magically. If everybody starts calling me Superman, I don’t wake up tomorrow looking like Henry Cavill. Unfortunately.
But belief does matter. If your internal language continuously tells you that you cannot change, that has consequences for whether you keep trying.
And speaking of Superman, there was another little synchronicity this morning.
I was thinking that maybe I should get home and do some strength exercises. Superman. Become stronger. And almost at that exact moment, Maarten came down the stairs wearing a mask and looking pretty damn bulky these days.
So we started joking around.
“Who’s Superman, me or Maarten?”
“Maarten,” Nadia said immediately.
Thanks.
“And what are you?”
“Supergirl.”
“And Daddy?”
“Super poop.”
There goes the entire positive-affirmation program.
A little later she asked me something I initially heard as:
“Daddy, are you an app?”
An app?
Why would I be an app?
“Because you’re such a good dad.”
So naturally I started explaining that you don’t need an app to be good at something, which apparently was a beautiful answer to a question nobody had asked.
She hadn’t said app.
She said ape.
Daddy, are you an ape?
Ah.
And there it was again. Language. One tiny sound separating two completely different concepts.
Ape.
App.
From ape to app.
Honestly, I think my daughter accidentally named this entire reflection.
Because isn’t that roughly the strange trajectory we’re living through? Apes developed language. Language allowed increasingly complex ideas to survive beyond a single brain. We turned those ideas into writing, mathematics, software and eventually apps. Then some of those apps became intelligent enough to help the apes remember what they had forgotten.
Including, apparently, their shower thoughts.
We now use technology to extend memory while simultaneously worrying that technology is damaging our ability to remember. We build apps for focus on devices designed to compete for our attention. We create AI to structure our thoughts because our thoughts have become overwhelmed partly by the information our technologies continuously put in front of us.
Wonderful.
Ridiculous.
Very human.
And after wandering through Jack Nicholson, mental institutions, coded novels, Dutch prefixes, language as a possible parasite, tokens, cardinal directions, AI cognition, weight loss, Superman and super poop, I end up back where this morning actually started.
At a school window.
A heart.
A smile.
Love and happiness.
Perhaps we need language to travel enormous distances through thought. Maybe that is one of its greatest gifts. It lets something inside my head travel into yours, sometimes centuries after the person who first thought it is gone.
But occasionally language also gets in the way. We dissect it, mistranslate it, overthink it, tokenize it and ask machines to summarize it.
And then a child walks past a window, you draw a heart, another child adds a smile, and everybody understands.
No prompt required.
Maybe the ape already had the important part figured out before we invented the app.
What's on your mind?