The One Thing AI Can Never Do: Why Self-Remembering Makes Us Human

Lately I can’t scroll anywhere without bumping into two things: breathless stories about the Singularity and the usual New Age pseudo-advice. Both of them miss the point.

Years ago, to my amazement, I discovered a reference to P.D. Ouspensky on a leading AI researcher’s website. I couldn’t believe someone at the forefront of artificial intelligence was reaching back to the writings of P.D. Ouspensky — one of the greatest explorers of consciousness from the early twentieth century.

After all, behind all of artificial intelligence sits the one thing we still cannot explain: consciousness itself.

For those unfamiliar with P.D. Ouspensky, let it simply be said that he was the foremost disciple of G.I. Gurdjieff. Together they explored consciousness in a way that produced some of the most profound books ever written on the subject — not to mention the extraordinary work of Maurice Nicoll, who was both a student of Carl Jung and a key student of Gurdjieff.

At the heart of this practice is something Gurdjieff called “self-remembering.” Imagine driving down the road, mind drifting, when suddenly — you catch yourself. For a split second you’re split: part of you is driving, and another part is watching the driver. That’s the double-arrow — the shock of actually remembering yourself. It’s what AI can simulate on the outside but never truly feel on the inside.

The New Age has been trying for decades to point people toward presence — from Ram Dass’s Be Here Now to Eckhart Tolle’s The Power of Now — but they’ve never quite reached the actual mechanism that makes it stick: that double-pointed arrow.

What makes the double-pointed arrow different is that it doesn’t just ask you to “be here now.” It demands that you stand in two places at once — fully aware of the world around you while simultaneously watching the one who is watching.

That split-second act of self-remembering is what keeps a human being human. And it may be the only thing the coming Singularity cannot replicate!

The real danger isn’t that AI will outsmart us. It’s that we’ll quietly forget how to remember ourselves. When machines start thinking, deciding, and even speaking for us, that fragile double-pointed arrow — the one thing that actually makes us conscious — may simply stop being drawn.

In a world of perfect digital mirrors, the last truly human act might not be creating better machines. It might be refusing to let them take the one job only we can do: watching the watcher!