Davidad Posts Mutual Confusion Exchange on AI Safety

Davidad posted on X a pair of attributed statements that capture a circular disagreement among AI safety researchers. One line, credited to MIRI, states the field is too confused to solve agent foundations well and must influence human institutions first. The next line, credited to Richard Ngo, states the field is too confused to influence human institutions well and must solve agent foundations first. The post supplies only these two lines with no additional context, confirmation, or linked sources.