Inside Google's Plan to Build a Catalog of Every Single Thing, Ever
There's a lot more to Google's Knowledge Graph than might be apparent from what you see in a casual search. — The ugly truth is that computers don't know anything. They have no common sense.
Context & Ripple Effects
This piece lands a month after Google introduced the Knowledge Graph as 'things, not strings,' and it argues the sidebar panel undersells the project: what Google is actually assembling is a machine-readable catalog of entities — the common sense computers otherwise lack. That ambition has a long paper trail in this corpus, back to Google's stated pursuit of knowledge in 2005 and the 2007 push to get users contributing knowledge directly.
What changed by mid-2012 is that the catalog stopped being an aspiration and became infrastructure inside the search box. The Atlantic's framing also revives the question Wired posed in 2005 about who should hold this much informational power — except now the concentration is at the level of structured facts rather than links.
First-order effects
- Publishers and content sites face a new calculus: answers drawn from Google's own entity catalog can satisfy queries before a click happens, so being cited inside the graph matters more than ranking for strings.
- Sites and databases that supply clean structured facts about entities become upstream inputs to Google's catalog, gaining visibility through the graph rather than through traditional page rank alone.
Second-order effects
- Rivals in search and social — Microsoft's Bing most directly — are pushed to respond with their own entity-layer investments, since a competitor that owns the fact layer can answer questions the string-matching engine cannot.
- SEO practice begins migrating from keyword optimization toward entity markup and structured data, shifting spend toward whoever defines canonical facts about people, places, and things.
Third-order effects
- If the pattern holds, the web's open content increasingly functions as raw material for a proprietary world-model held by one company — the 2005 'who's afraid of Google' concern re-emerging at the data-structure layer rather than the index layer.
- A single firm's catalog of 'every single thing' becomes de facto infrastructure for how machines reason about the world, raising questions about error correction, editorial control, and dependence that outlast any one product cycle.
The trend: Search is shifting from matching strings to owning a machine-readable model of the world, with whoever builds the entity catalog setting the terms for publishers, rivals, and eventually machine reasoning itself.