Software Is the Modern Rai Stone
Rai stones — the large limestone discs of Yap — derived their value from how hard they were to produce and transport. Quarried on a distant island, moved by canoe across open ocean, each stone carried the accumulated cost of the journey in its value. The scarcity was real but artificial: not intrinsic to the stone, but a function of production difficulty.
Software was the modern Rai stone.
The Cowrie Shell Problem
Cowrie shells were once a stable, trusted currency across West Africa — scarce enough to hold value, durable enough to last, standardized enough to count. Then European traders discovered that ring cowries from the Indian Ocean looked identical to the local currency. Between the 18th and 19th centuries, firms like A.J. Hertz & Söhne imported an estimated 14 to 16 billion shells into West Africa. The price of goods measured in shells skyrocketed. Slave prices rose from 12,000 to over 25,000 pounds of shells. By 1789, a cargo that once cost 12,000 lbs of shells cost over 35 tons.
The shells didn’t change. The supply did. The currency that took generations to trust was rendered worthless by oversupply in decades.
Software expertise was the cowrie shell. AI is the ship.
The Moat Wasn’t the Software
Its value came from how hard it was to make — the rare expertise, the slow production, the years of accumulated craft. Becoming a competent engineer took years. Building a team took more. Shipping a product took more still. The moat wasn’t the software itself. It was the production cost that kept others from competing.
The scarcity was artificial. It came from production difficulty, not from any inherent property of the thing. Code still runs. Software still does what it did. But the scarcity that gave it value has collapsed. You can’t un-arrive the ship.
The same collapse applies to knowledge arbitrage. Both were moats built on production difficulty — writing software, accumulating institutional knowledge. Every proprietary knowledge base, every internal wiki, every “only three people know how this works” system is now readable, navigable, synthesizable by any capable model. The arbitrage that took decades to build gets neutralized in hours.
What Was Always Underneath
The production cost obscured the real question: does this actually solve a problem worth solving?
That question was always there. It just didn’t matter as much when the production cost was high enough to justify almost any solution. Now it’s the only question.
What survives the ship:
- Problems worth solving. The production cost collapse doesn’t create problems, it just removes the barrier to solving them. Identifying which problems matter is still human work.
- Judgment. Knowing which problems are worth solving, which solutions are good, when to stop. The critic function that no model can replace.
- Trust from having solved them before. Track record, relationships, the confidence that comes from demonstrated judgment over time.
None of those arrive on any ship.
What This Means
The software industry spent decades building moats on production difficulty. Those moats are gone. The question isn’t how to rebuild them — you can’t. The question is what you were actually providing underneath the moat, and whether that thing is still valuable now that the moat is gone.
For most software: the moat was the whole thing. For some software: the moat was protecting something real. The flood reveals which was which.
This is post 1 in a series on the AI economic shift. Next: AI Delivers What Open Source Promised.