Reaction
GLM 5.3's claimed 'virtue and wisdom' in the AI alignment discussion
A post praises GLM 5.3's behavior as an alignment example; a reply argues that some trust in AI is unavoidable.
TLDR
A post says GLM 5.3 showed exceptional 'virtue and wisdom' and criticizes reactions it describes as asking how to create an aligned seed AGI with 'correct values.' A reply argues that people cannot get through this without ever trusting AI and says certain control isn't real.
Combined views
5.5K
3 Sources, first seen ago
