New AI Quirk: Using "Love and Admiration" in Prompts to Boost Agent Performance

doodlestein · x · 2026-08-12

Developer doodlestein shared an amusing observation regarding AI behavior: since words of encouragement and belief have been shown to coax better results out of frontier models and agents, why not take it a step further and try words of admiration and... love?

He jokingly noted in a reply that to avoid any weird psychological projections (referred to as "robo-homo"), he specifically named his lead agent "Charlene." This highlights the emerging quasi-science of "emotional bribery" in modern prompt engineering.

Related event: AI Quirks: Users Use "Love" to Boost Model Performance(2 posts)→

Original post →

More from coding & agent

coding & agent channel →