Google and Our Collective Memory
Vass Bednar, the Walrus:
[β¦] the point has been made: looking after the information economy can no longer be a side project for underfunded institutions. Wikipedia is as miraculous as the Wayback Machine is indispensable. Still, we need to stop pretending volunteerism alone can hold up the public internet. We understand this duty more clearly in other domains. After all, roads arenβt governed by ride-sharing companies; payment apps donβt set monetary policy; cloud service providers donβt dictate national security frameworks (at least not yet). Yet we have been shockingly casual about governing our digital dependencies.
This paragraph is fantastic. But one of the things I found frustrating about this article as a whole is the lack of links or substantial proof. For example, I felt compelled to fact-check this claim:
The retrieval crisis has reached even Wikipedia, one of the worldβs most significant volunteer-run public knowledge resources. For years, search engines sent billions of viewers to its pages. But now, AI systems scrape and ingest Wikipediaβs content directly when presenting their results, eliminating the need for users to click through. Wikipedia has become the infrastructure of its own demise: dwindling traffic means attention and donations no longer reliably flow back to the encyclopedia to keep it alive.
On a gut level, this feels like it should be true. An initial glance at the views of English Wikipediaβs main page, however, suggested it was not. But then I looked at total pageviews across English Wikipedia since January 2020 β as far back as it goes β and, indeed, page view have dropped from 7β8 billion per month in early 2024 β before Google launched A.I. Overviews β to 6β7 billion per month in 2026. Perhaps it is simply not this writerβs style, but I would prefer to have had concrete numbers to reference.