AI Is Striking Gold for Infovores

The internet arrived with a promise that sounded close to utopian: put the whole of recorded knowledge one click away, and a generation of curious people will teach themselves anything. The first half came true. The second half ran into a wall built out of advertising. Search results filled with pages engineered to rank rather than to inform. Recipes hid under childhood memoirs. Straight answers got padded so you would scroll past one more banner on the way. The knowledge was there. Reaching it turned into a chore.

Psychologists have a word for the kind of person this was meant to serve. Irving Biederman and Edward Vessel coined “infovore” for the human appetite for new information, the small flush of satisfaction that comes from finally understanding something. Infovores read the footnotes. They chase a question sideways into three unrelated fields. They would rather know than not know, even when knowing pays nothing. For two decades the web treated that appetite as a seam to be mined, not a thing to be fed.

The cost of a straight answer

Language models change the economics of curiosity. The price of a clear, patient explanation of almost anything has fallen to roughly nothing. You can ask the naive question without the social cost of looking ignorant, then ask the follow-up, then ask why the follow-up seems to contradict the first answer. The thing that used to gate self-teaching was never the availability of facts. It was the friction of assembling them into an explanation pitched at exactly your level. That friction is mostly gone.

The obvious objection is that these systems make things up, and they do. A model will state a confident falsehood in the same even tone it uses for the truth. But this is a smaller problem for an infovore than for a casual user, because the infovore already treats every source as provisional and cross-checks the interesting claims by reflex. The people harmed by confabulation are the ones who wanted an oracle. The people helped are the ones who wanted a tireless study partner and knew better than to trust it blindly.

What this exposes is that the real bottleneck was never access. It was appetite. When information was scarce, curiosity was cheap and the constraint was supply. Now supply is effectively infinite and the scarce input is the wanting itself, the willingness to keep asking. The tool finally matches the temperament it was always described as serving. The catch is that it rewards a disposition most people do not have, and no amount of cheap explanation manufactures the desire to use it.