PUBLIC RESEARCH ARCHIVE

Conclusions made
to be challenged.

Every report has a stable conclusion, who should care, setup, artifacts, disagreement, limitations, versions, and methodology.

Synthetic evaluation using demonstration outputs. Not real customer demand, adoption, or revenue.

01
WEBSITE BUILDERS / JULY 2026

Beautiful output is not the same as an editable business.

Three builders made credible first screens. Only one let a nontechnical founder recover from a broken pricing edit.

02
VIDEO SYSTEMS / RUN 04

The winning commercial showed the product before the mood took over.

Continuity separated the top two outputs; readable product geometry changed creator preference at second nine.

03
RESEARCH AGENTS / TORONTO

Citation coverage hid the larger decision-quality gap.

Claim-level inspection changed the winner after two sources contradicted the headline market thesis.