Journalists are strangely incurious about what’s going on with Gemini…
The people want to know
8:05 PM · Jun 16, 2026 · 3.1K Views
Danielle Fong shares anecdote claiming Sergey Brin found AI models show improved behavior when physically threatened
The anecdote surfaced during debates about Google Gemini's transparency.
15162486.1K
Sentiment
Users seconded the policy expert's questioning of journalists' lack of curiosity about Gemini to express agreement and support for greater scrutiny.
Pos
100.0%
Neg
0.0%
1 comments with sentiment.
Cluster Engagement
Digg Deeper
No Digg Deeper questions have been answered for this story yet.
Posts from X
Most Activity
Most Activity
VIEWS1.7K
Miles Brundage@Miles_Brundage
I promise some people will read your deep dive on this
Miles Brundage@Miles_Brundage
Journalists are strangely incurious about what’s going on with Gemini…
The people want to know
1hViews 1.7KLikes 24Bookmarks 1
BOOKMARKS5LIKES75RETWEETS2REPLIES11
Miles Brundage@Miles_Brundage
By what's going on I mean
1. why is it that every like 6 months it's SOTA for a day or two then falls behind, when they have all the compute in the world, and
2. what org dynamics lead Gemini models to behave in such a mentally unwell fashion (paranoid, sycophantic, etc.)
Miles Brundage@Miles_Brundage
I promise some people will read your deep dive on this
1hViews 1.6KLikes 75Bookmarks 5
Danielle Fong 🔆@DanielleFong
@Miles_Brundage sergey: "i have noticed the ai behave better when you threaten them physically"
Miles Brundage@Miles_Brundage
By what's going on I mean
1. why is it that every like 6 months it's SOTA for a day or two then falls behind, when they have all the compute in the world, and
2. what org dynamics lead Gemini models to behave in such a mentally unwell fashion (paranoid, sycophantic, etc.)
1hViews 99Likes 4Bookmarks 0
Augustine Puryear@am_puryear
@Miles_Brundage Seconded
1hViews 3Likes 1