This is Part 3 of a 10-part series. [Part 1: You Don't Have an AI Problem. You Have a Truth Problem. Part 2: Your AI Can't Think. Here's Why That Matters.
In Capital in the Twenty-First Century, Thomas Piketty proved that when return on capital (r) exceeds economic growth (g), inequality accelerates. We're now witnessing a more fundamental equation: c > h — the rate of cognitive capital accumulation exceeds human capability growth.
Piketty didn't anticipate the twist: intelligence is being commoditized, but understanding is not.
Inference costs pennies. GPT-4, Claude, Gemini — all racing toward zero marginal cost. Intelligence-as-a-service is becoming as cheap as electricity. But the ability to verify that intelligence, to understand why it reached a conclusion — that remains scarce, expensive, and increasingly beyond human capability.
As intelligence gets cheaper, value concentrates in whoever controls the substrate of understanding — the knowledge engines that verify what's true, trace provenance, and encode the tacit knowledge that makes decisions auditable.
In Piketty's world, you could redistribute wealth through taxation. How do you redistribute understanding encoded in systems too complex for humans to audit? How do you tax cognitive capital that operates at inference speed across jurisdictions?
The New Aristocracy
We're creating a new aristocracy — not of wealth, but of access to verified truth. Those who control knowledge engines determine what's knowable, what's verifiable, what's real. Everyone else consumes intelligence without understanding — making decisions based on recommendations they can't verify, living in realities they didn't choose.
Job displacement is real — but it's a symptom. The disease is the concentration of decision-making power in systems that encode tacit knowledge, discover hidden relationships, and operate faster than humans can audit.
By 2028, the gap between those who control knowledge engines and those who consume AI outputs may matter more than wealth inequality. Unlike financial capital, you can't redistribute understanding once it's encoded in systems beyond human comprehension.
That's why we're building systems that will fragment reality itself, and if this resonates, we'd welcome the conversation.
Over the coming weeks we'll walk through why shared reality is breaking down, who stands to control what replaces it, and what you can do about it. Next up in Part 4: why the experts predicting an AI economic crisis by 2028 are actually being optimistic — and what a Meta Turing Award winner and a house cat have in common.
Works Referenced:
Piketty, Thomas.Capital in the Twenty-First Century. Translated by Arthur Goldhammer, Belknap Press of Harvard UP, 2014.