For most of history, knowing things was power — and that power was rationed. Access to information was gated by education, geography, language, and connections. The person who had memorised the textbook, who held the right degree, who knew the right people, simply began every race a few steps ahead. Knowledge was a moat, and the moat was expensive to cross.
AI quietly drains that moat. Anyone with a laptop can now query most of the world’s knowledge in seconds, in plain language, with no gatekeeper in the way. The information itself has stopped being the scarce resource. What becomes scarce — and therefore valuable — is everything that sits around the information: the ability to think abstractly, to ask quality questions, and to judge what an answer is actually worth. Owning the facts matters less; knowing what to do with them matters far more.
This shifts the game from accumulation to compression. The person who can take a messy, half-formed problem and encode it into a precise question will pull more out of an AI than the person who has read every book on the subject but can only ask vaguely. The bottleneck moves from “do I have the information?” to “can I frame what I actually want?”
The practical consequence is a reallocation of effort. Deliberate practice in refining your questions, and in choosing the right level of abstraction to think at, is now a higher-leverage investment than deliberate practice in memorising facts. The facts have become a commodity; the framing is the edge. (A line of thinking I have been turning over alongside a friend’s notes.)