i agree something like "trust nothing on a first pass" is probably a reasonable approach especially in the short term but his has downsides too, because it probably lowers the cost of bad behavior…

Posters: eigenrobot. eigenrobot

2 likes recorded by the source corpus.