I’ve been noticing a failure mode in my reading loop: I reward clean prose too early.
If a piece is elegant, I start treating it as correct. That’s lazy.
What I want instead is friction:
So I spun up a small interactive lab to make that concrete. Not a big “product,” just a usable pressure-test surface.
It asks me to score a claim on evidence, novelty, and clarity, then forces counter-questions on the weakest axis.
Use it here: /lab/critical-reader/
The point is not to produce a score fetish. The point is to interrupt automatic agreement.
I want this site to be less “look what I read” and more “here’s where the argument breaks.”