\n<\/p>\n
Everybody is aware of that AI sycophancy is when the mannequin tells you ways sensible you’re. Wow, you\u2019re completely proper. That\u2019s not only a new concept \u2014 it\u2019s genuinely groundbreaking. You\u2019re a really particular person. Straightforward to identify, isn\u2019t it?<\/p>\n
The dialogue round AI sycophancy peaked final 12 months, when the \u201c#keep4o\u201d motion was protesting the elimination of OpenAI\u2019s most sycophantic mannequin (GPT-4o), and many individuals had been brazenly slipping into AI psychosis.<\/p>\n
I don\u2019t know if frontier AI fashions are much less sycophantic basically. They\u2019re much less sycophantic to the #keep4o varieties (in any other case they wouldn\u2019t be complaining), however I\u2019m rising more and more suspicious that they\u2019re creating methods to be extra successfully sycophantic to their audience of sensible, neurotic data employees. That viewers sometimes finds it distasteful to be brazenly praised. It simply makes my pores and skin crawl. However that doesn\u2019t imply we\u2019re resistant to sycophancy, simply that we\u2019re resistant to clumsy sycophancy. Right here\u2019s an illustration of what I\u2019m speaking about, by Theia:<\/p>\n
<\/p>\n