A response · July 2026
After the descent, the only moat is memory
Chamath Palihapitiya says the price of intelligence is collapsing toward zero, and he's right. Here is the question he leaves for the rest of us: when everyone can rent the same mind, what's left to own?
文章以英文發佈。
Fifteen years after Marc Andreessen told us software would eat the world, Chamath Palihapitiya has named the sequel: intelligence is descending the same cost curve the smartphone rode from a $500 luxury to six billion pockets — except faster, because it compounds two discounts at once. Cheaper silicon, like every technology before it. And more efficient models, a second curve stacked on the first. He calls it the great descent. We think it's the most useful frame anyone has put on this moment.
He's right about the deeper thing, too. The internet made information abundant and left expertise exactly where it had always been — scarce, expensive, locked in the heads of trained people. What's falling toward zero now isn't access to facts. It's judgment: the reasoning of the analyst, the intuition of the clinician, the drafting of the lawyer. The oldest bottleneck in the economy is breaking.
But look carefully at what actually arrives at the bottom of the curve.
Brilliant, and amnesiac
The expertise you can rent today has a property no human expert ever had: it forgets you completely. Every conversation is its first day on the job. It will solve the same problem for you a thousand times and never once notice it has seen it before. It reasons like a veteran and remembers like a stranger.
That is not a small caveat. It's the difference between an expert and an encyclopedia of expertise. A consultant who forgot every prior engagement the moment it ended would not be a cheap consultant — they would be an expensive one, because you'd pay back in repetition everything you saved in salary. Rented intelligence answers questions. It does not accumulate.
The trap has a second floor
Chamath names the first trap precisely: if you consume the same off-the-shelf intelligence as your competitors, piped into the same generic workflows, you haven't built an edge — you've erased one. We'd add a second floor to that trap. Even careful, proprietary use of stateless intelligence doesn't compound. Prompts can be copied. Workflows converge on best practice. Fine-tuning is a snapshot that starts aging the day it ships.
What cannot be copied is what your systems have learned inside your operation — from your exceptions, your customers, your mistakes, your decisions and their outcomes. That learning is the edge he's describing: the tacit knowledge that today walks out the door in retiring heads. Encoding it isn't an act of writing it down once. It's building systems that keep learning it, every day, in production.
When everyone can rent the same mind, the only thing left to own is what yours has learned.
The loop is the moat
This is the wager behind our research program, Learning-Augmented Generation: the next differentiating layer isn't better answers, but remembered ones. Systems where every action becomes memory, every outcome becomes learning, and the answer you get in March is measurably better than the one you got in January — because of what happened in February.
Compounding has a clock. A learning loop that starts this quarter has banked two quarters of your reality by the one after next — and none of it transfers to a competitor who starts later. They can rent the same model. They cannot rent your memory. The descent makes intelligence cheap for everyone at once; it makes accumulated learning more valuable, not less, because it becomes the only scarce thing left.
Where the descent lands first
The end of the curve is ubiquity, and we happen to work where ubiquity lands hardest: Southeast Asia and India, the geographies that took the smartphone from a luxury object to six billion hands. The same arc is now running for expertise — and the companies here won't adopt it the way the last era's incumbents did, carefully and a decade late. They'll do what they did with mobile: skip the intermediate step entirely.
Every company is somewhere on this curve — curious, experimenting, scaling, or already learning. Wherever you are, the descent asks the same question: when intelligence is a commodity, what will you have encoded that is yours? That's a question worth answering early. The water is moving.