points by sroussey 1 day ago | 0 comments
Deepmind creates these, Google Cloud productizes them.
If Google Cloud was in charge, Gemini 5 would be public, and they wouldn't be messing with Gemini 4 nor 1M context when they have 10M internal. C'est la vie.
[–] guelo 1 day ago
A 10M window would require dozens of GPUs (or TPUs) dedicated to each query. That is something that Google could afford internally but wouldn't be able to scale to an external product.
[–] anticensor 1 day ago
More like one huge waferscaler such as Cerebras.
[–] bushbaba 1 day ago
[–] mixtureoftakes 1 day ago