My Singularity

I am doing something that scares the shit out of me. Writing.

Not the thinking. Thinking has never been my problem — original thoughts I have in spades. It is getting the words down, in order, in a form other people can read, that has kept a ceiling on my expression for most of my life.

To explain why, I have to tell you about the singularity.

In 2017 I was on set at TechCrunch when Sam Altman came in to talk about diversity at Y Combinator and ended up talking about his new AI lab, OpenAI. He made a case I have heard many times since: that the singularity — the merging of humans and machines — has already happened. Look at your phone, he said. It stores your memory. It does your communication. “I am uncomfortable now with it sitting five feet away from me.”

I think Altman was right. But for me the singularity came much earlier. Mine began in 1991, in a first-grade classroom at Van Hise Elementary in Madison, Wisconsin, where I was failing to learn to read.

Everyone around me was cracking the code — how combinations of letters make sounds, how sounds make words. I understood the rules. I did the work. And still a word as simple as THE sat beyond my reach. After many hours of tests, my parents and I learned that I have a severe case of dyslexia.

Here is the thing about reading: the human brain never evolved to do it. It is a cultural invention a few thousand years old — far too recent for the brain to have grown dedicated machinery. Reading is a hack. It borrows visual, auditory, and language systems and repurposes them to decode squiggles. In my brain, the borrowed parts are wired differently. So I built workarounds.

I am a fluent reader — but only in silence. I read by shape. I look at the outline of a word, take the context around it, and fill in the meaning without actually decoding it. All words are sight words to me. If a word has not been committed to memory, everything stops while I sound it out and recall every rule and exception the English language is riddled with. Ask me to read out loud and the whole system collapses — the words I recognize on the page jam on the way from my head to my mouth.

I speak more slowly than most people, and not because I am unsure of what I want to say. The words are there. I can see the idea clearly, feel the shape of the thought. The pathway from idea to language is where it gets stuck. Not in a charming, poetic way — in the frustrating way of knowing you are smart and sounding unsure.

So I augmented. Mario Teaches Typing. Dragon Dictate. First-generation word processors with rudimentary spell check. Books on tape, books on CD, podcasts. WordPerfect. Google. Voice-to-text. From that first-grade classroom to today — I am 42 now — I have been strapping machines to the part of my brain that does not work, so the parts that do could get out.

AI is not a new kind of thing in my life. It is the newest thing in a very long line.

Which brings me to the argument everyone is suddenly having.

Tyler Cowen says you will learn to love AI writing, and that the people who insist they can always tell are mostly wrong. A study made the rounds last month showing readers rate AI-written stories higher than human ones — and here is the twist — they prefer them most when told a human wrote them. Substack has pointed an AI detector at its own writers, sniffing prose for the fingerprints of a machine. An essay crossed my feed asking the question directly: should you care if what you are reading was written with AI?

And this morning an editor named Stephanie Zinn published the best piece I have read on the subject. Her argument: the AI “tells” everyone hunts for — the em dashes, the tidy lists, the phrases that sound weighty but cannot survive paraphrase — are just bad writing habits that predate the machines. Stop asking whether it was written by AI, she says. Ask whether the writing is doing its job.

She is right, and I want to push her point one step further, because buried in her essay is a sentence she moves past quickly: “AI helps people articulate and publish ideas they’d otherwise never write down.”

I am that person. That sentence is my whole life.

Everyone in this debate — the detectors, the defenders, even the best of the editors — shares one assumption, and nobody says it out loud: that the quality of the writing tells you the quality of the thinking. That fluent prose means a fluent mind, and assisted prose means a lesser one.

I am living proof that assumption is wrong. It has been wrong about me since 1991.

My thoughts were never the problem. The pipe was the problem. My whole life, people have been reading my unassisted writing and drawing conclusions about my mind from my spelling. The detector question — did a machine help with this? — catches exactly the wrong people. It flags the dyslexic kid who finally has a way to get the words out, and waves through anyone who can type fluently about nothing.

Dyslexia is not a deficit. It is a tradeoff. The wiring that made THE unreadable also pushed me into visual reasoning, systems thinking, creative synthesis — competing in games other people were not playing, because I had to find the paths where I had an advantage. Constraints are the mother of creativity. But the tradeoff came with a tax, and the tax was always collected in writing.

AI is the first tool that lets me pay it down. Writing is a tool for thinking — I believe that completely — and my thoughts have spent most of my life trapped behind a fogged-up window. This is the first time in my life the window is clear.

So: did AI help me write this? Yes. So did spell check, and Dragon Dictate, and Google. If that changes what you think of the ideas here, I would gently suggest the ideas were never what you were judging.

The road I am on is not direct or straight. It never has been — not because I seek adventure, but because it is the only road available to me. But for the first time, I can tell you about it in writing.