Because 3295 people worked on Gemini. From the paper (page 61):
> The development of Gemini is a large-scale collaborative effort involving over 3000 individuals across Google, including researchers, engineers, and operations staff. These individuals contributed their hard work and expertise across diverse areas, from foundational research and the development of model architecture, data, training, and infrastructure, through to evaluation and ensuring safety and security. We gratefully acknowledge the dedication and hard work of each contributor in making Gemini a reality.
tzury•6mo ago
that is obvious, however one must wonder what is means when 14 pages of a paper are simply credit names...
I was never fond of the "Hollywood credit list", and it seems like this culture entered our trade.
If a paper is about
Advanced Reasoning
Multimodality
Long Context
Next Generation Agentic Capabilities
then perhaps we should limit the "authors" to those who contributed directly to the topic.
I am convinced a large amount of those in the list are from DevOps, DevEx and DevRel (let alone "Digital" and "Product")
asimpleusecase•6mo ago
Could also be a way to obscure the core team to make it harder to poach - right now the key scientists are worth 8 figures
tzury•6mo ago
sounds about right.
the irony is you can use the very tool (Gemini) to extract LinkedIn profiles of all names and understand well who's doing what.
bhaney•6mo ago
> The development of Gemini is a large-scale collaborative effort involving over 3000 individuals across Google, including researchers, engineers, and operations staff. These individuals contributed their hard work and expertise across diverse areas, from foundational research and the development of model architecture, data, training, and infrastructure, through to evaluation and ensuring safety and security. We gratefully acknowledge the dedication and hard work of each contributor in making Gemini a reality.
tzury•6mo ago
I was never fond of the "Hollywood credit list", and it seems like this culture entered our trade.
If a paper is about
then perhaps we should limit the "authors" to those who contributed directly to the topic.I am convinced a large amount of those in the list are from DevOps, DevEx and DevRel (let alone "Digital" and "Product")