AI & TechWION
Meta accused of faking AI benchmark scores again
in5points
Meta's former chief AI scientist Yann LeCun confirmed Llama 4's benchmark results were manufactured through checkpoint cherry-picking.
Meta's real-world AI deployment on Instagram led to wrongful business account bans.
A hack triggered by simply asking Meta's AI shows the gap between claimed and actual capability.
The article raises concerns about Meta's Muse Spark 1.1 possibly repeating same benchmark manipulation.
Story angle
Center
L
CL
C
CR
R
Near the middle: closer to straight reporting, less party-line spin.
About how the story is framed, not whether the facts are true.
WION
Center
Mixed factuality
Who owns this?
Zee Media
Essel Group / Subhash Chandra family
International-facing channel from Zee Media.
Wikipedia: Subhash Chandra