Wikipedia relied on a web that sent readers back

In The Economist's essay on digital culture, the premise is that humanity will never assemble another volunteer project on the scale of Wikipedia. The reason comes down to an implicit bargain that the web has largely abandoned. Wikipedia succeeded because contributors saw their work read by millions, direct links drove traffic to primary sources, and editors earned standing within an active public community.

That loop broke when consumption shifted from browsing pages to querying models. When chat interfaces and summary engines answer questions by scraping open archives, users get their answers without ever visiting the source. The resulting collapse in web traffic cuts off the pipeline of curious readers who eventually turn into volunteer editors. At the same time, the flood of synthetic text forces the remaining volunteers into defensive triage, spending their hours validating citations against automated hallucinations instead of writing new entries.

For anyone relying on public data to train or ground intelligent systems, this dynamic is an operational risk. High-quality human consensus does not regenerate itself automatically under unilateral extraction. As an industry, we cannot treat the open web as an infinite quarry for model weights while eroding the incentives that kept the information accurate in the first place. If we want dependable training data, our platforms need to route attention, attribution, and economic value back to the communities curating it.