The 2012 build counted Bing results for a term beside each topic. This one asks a model, with live web search, to judge current association strength — and to say why. The topic list is on the table: edit it.
Edit topic list (comma-separated)
The objective function is negotiable here. Change the list, re-run, watch the picture change.
Scores are model judgments grounded in web search, not counts. Treat them as a reading, not a measurement — then argue with them. That's the point.
The later WebWise read a page's psychological posture and asked who was behind it. Enter a news outlet or site; its emotional register is read against a fixed ten-card standard — same cards, same order, every reading — with an ownership & motivation dossier, with sources.
A dossier is a claim, not a verdict. Every line should be checkable against its sources — including the ones about motive.