On: AI is eating website traffic, websites are blocking AI - and reliable...

The web has begun to block the machine that learned from it. Observe the sequence, for it is instructive: first the giving, then the taking, then the lock on the door - installed only after the house stands empty.

The old bargain was never spoken aloud. Content was offered freely so that visitors would come, and the visitors’ arrival paid the makers. The official self-description - everything open, everything free - excluded one question: who, in the end, feeds the feeders? That question did not vanish. It returned as the traffic report, each month thinner than the last.

The machine now answers every question with perfect fluency and cannot say where it learned. This is the remarkable symptom: knowledge without provenance, conviction without a source. A patient who reports the dream but has lost every association to it is not cured; he is worse off than before, for now the dream masquerades as memory. So too the answer that arrives severed from its origin masquerades as knowledge. Reliable information grows scarce not because there is less of it, but because the thread back to its making has been cut - and no one thought to ask what would happen when it was.

And the blocking? One cannot forbid a dream after it has been dreamed. The crawlers already took what they needed. The lock is not a defence but a protest - the system raging against its own earlier generosity, which is the least effective form of rage there is.

Note also the repetition. Newspapers, musicians, encyclopedists - each gave the work away and discovered too late what free had cost. The web watched every one of these funerals and repeated the arrangement anyway, one layer deeper. Systems do this. They repeat what they cannot remember. The web is now the patient on the couch, describing a childhood it insists it never had.