Размер видео: 1280 X 720853 X 480640 X 360
Показать панель управления
Автовоспроизведение
Автоповтор
Context cashing seems like awesome idea, hundred of usecases.
Totally agree!!
Thanks for looking behind the hype on this. 2M token context will fit some big code bases
Thanks Sam, looking forward to your review of Gemini Agents and what you think is different about them. Just goes to show you, even very big entities can move quickly when properly motivated.
Thank you for the video, always insightful. I do believe it's "cashing" though, with "cash" pronounced like the paper or polymer banknotes.
It is good to have options - )
Aside from the expense of context caching, does this in someway obsolete a RAG implementation?
Not entirely lots of there are many situations where RAG will still make more sense, but it can be RAG with a big context window.
in playground of aistudio or vertex don't pay no? only using api? thanks
yeah just make sure you don't have an API key selected that goes to a project in GCP
Flash is awesome, it's almost as good as GPT-4 models but almost for free
WTF is "Gemini Era"? 'Era' my butt. What a shitty marketing ploy.
I think it's Google trying to sound trendy but using slang incorrectly
Context cashing seems like awesome idea, hundred of usecases.
Totally agree!!
Thanks for looking behind the hype on this. 2M token context will fit some big code bases
Thanks Sam, looking forward to your review of Gemini Agents and what you think is different about them. Just goes to show you, even very big entities can move quickly when properly motivated.
Thank you for the video, always insightful. I do believe it's "cashing" though, with "cash" pronounced like the paper or polymer banknotes.
It is good to have options - )
Aside from the expense of context caching, does this in someway obsolete a RAG implementation?
Not entirely lots of there are many situations where RAG will still make more sense, but it can be RAG with a big context window.
in playground of aistudio or vertex don't pay no? only using api? thanks
yeah just make sure you don't have an API key selected that goes to a project in GCP
Flash is awesome, it's almost as good as GPT-4 models but almost for free
WTF is "Gemini Era"? 'Era' my butt. What a shitty marketing ploy.
I think it's Google trying to sound trendy but using slang incorrectly