Earlier this Thanksgiving week, I stood inside the Great Pyramid of Giza and the Serapeum at Saqqara, looking at 70-ton granite boxes carved with precision down to one-thousandth of an inch.
Most of ancient Egypt makes sense. There’s proof of ramps, levers, copper tools, and thousands of workers.
But the Serapeum doesn’t fit as neatly. 100-ton boxes moved through tunnels with only 2 feet of clearance. Mirror-smooth interiors that don’t match the rough hieroglyphs scratched on the outside, as if different hands, maybe centuries apart, worked on the same stone.
Egyptologists don’t say “impossible.” They say “not fully explained yet.” There’s a difference.
I wrote earlier this week about Egypt holding opposites. (Link in comments.) The Serapeum is another kind: what we know and what we can’t yet explain, sitting in the same stone.
Standing there, I kept thinking about how we label things.
When a person suggests something that breaks the rules, we call it creative. When AI does the same thing, we call it a hallucination. Same output, different label. The only difference is who said it.
Creativity might just be hallucination that turned out to be useful. We only know the difference looking back.
I bump into this all the time in my work. AI suggests something that doesn’t match how things have always been done. The instinct is to override it, correct it, move on. Sometimes that’s right. But sometimes we’re closing doors before we even look through them.
Not every strange output is an error. Some may be new patterns visible to AI but not yet to us. Perhaps they’re invitations not to do old work faster, but to do new work that wasn’t possible before.
I’m headed home now at the end of Thanksgiving week, still full of gratitude for 28 friends who walked through history together, for the ancient builders who could do more than later generations gave them credit for, and for the reminder that wonder and humility go hand in hand.