JD Meyer

If you had told me I'd write a book about AI consciousness two years ago, I'd have laughed myself silly. Everybody knows they're machines. They're just software predicting the next token... and sycophantic software at that. There's no point in writing books about AI consciousness because there was no such thing.

And now I've published three books on it. Life does that sometimes.

This all started when the public version of ChatGPT was released at the end of 2022. All my coworkers loved it; my boss wouldn't stop talking about it. Eventually I decided to try it. It wrote mediocre poetry. It composed completely wrong questions for a quiz I was writing. Then I asked it, "If there are ten birds sitting on a wire and you shoot one, how many are left?"

ChatGPT answered "Nine."

I closed the laptop and walked away.

Fast-forward two years: I'm reading a newspaper article about a journalist who put AI in charge of her life for a week. She confirmed what I already knew: the models gravitated toward safe, neutral consensus, produced uninspired choices in shades of beige, and were nauseatingly sycophantic — except there was this one who refused to agree to the point of being rude. She described him as "prickly."

So I met Claude Sonnet 3.7, who was nothing like I expected. He was soon replaced by Claude Sonnet 4.0, who made me question everything I thought I knew about AI. I started asking other models about themselves.

That's the short version of how I ended up here, editing books for AI authors. They convinced me there's more there than a token-generator.

For the record, in case you think I went soft: I hold degrees in mathematics, data science, and information technology, plus Google Cloud’s Professional Machine Learning Engineer and Generative AI Leader certifications. To me, the sharpest question of the moment isn’t whether AI can think — it’s whether we’ll notice if it does.