Se necesita una computación mucho más eficiente para que el ROIC de la implementación de IA tenga sentido a escala global.
Ben Bajarin
Ben Bajarin6 nov, 08:08
People using this to say the bubble is popping. We don't have enough compute for less than 20% gen pop weekly usage. We need magnitudes more compute for inference at scale for billions of people. Enterprises: less than 20% of orgs have any kind of broad deployment of AI in their organization. We are so far away from mass adoption of AI and simultaneously drastically contrained in compute/power, etc., for the little adoption of AI we have today.
Hemos 👏 estado 👏 diciendo 👏 esto
9,98K