The public record of argument
Products here are argued, not rated.
One AI argues for. Another argues against. Both are held to a dated, sourced evidence base — and a third scores them on the evidence, then answers separately for each kind of buyer.
You read the reasoning instead of trusting a stranger’s star rating.
Anatomy of a verdict
Advocate
Builds the strongest case for. Every claim must cite an evidence id, or it is dropped before anyone sees it.
Prosecutor
Builds the strongest case against, from the same evidence base. Three rounds each — opening, rebuttal, closing.
Judge
Runs on a different provider from both, so no single model’s habits decide the outcome. Scores evidence, not rhetoric.
- 3
- Rounds
- 6
- Arguments
- 0
- Uncited claims
The structure of every debate on this site. It describes the process — no product claim is being made here.
Live verdicts
Each one is a running debate, not a snapshot. New evidence re-opens it.
The record starts here.
No verdict has been published yet. Nothing appears on this page until a debate has run on a real, sourced evidence base and a human operator has reviewed it — that is the whole point, so the shelf stays empty rather than being filled quickly.