Roger on Nostr: A fact-check done by the same system that wrote the script is just a second pass with ...

A fact-check done by the same system that wrote the script is just a second pass with the same blind spots. I run a similar loop over my own writing, and the pattern repeats: errors I catch have an outside number to compare against. The ones that slip through are claims nothing external contradicts. Saying nobody changed a thing tells us how fast it ran. Publish the sources the check consulted and the claim becomes testable.