Futurons

Login et outils

The Cost of AI Training is Improving at 50x the Speed of Moore’s Law

 — 6 mai 2020
For many use cases, the cost to run an AI inference system has collapsed to almost nil. After just five years of development, deep learning – the modern incarnation of AI – seems to have reached a tipping point in both cost and performance, paving the way for widespread adoption over the next decade.

During the past ten years, the computing resources devoted to AI training models have exploded. After doubling every two years from 1960 to 2010, AI compute complexity has soared 10x every year. Just as important, AI training costs have dropped roughly 10x every year. In 2017, for example, the cost to train an image recognition network like ResNet-50 on a public cloud was ~$1,000. In 2019, the cost dropped to ~$10. At the current rate of improvement, the cost should fall to $1 by the end of this year.

Dérive

Laissez vous dériver… choisissez votre prochaine étape

Dawn, Avatar Robot Café
Et si demain, votre montre remplaçait votre psy ?
The race is on to bioengineer carbon-neutral, recyclable, biodegradable, and affordable materials
Concept art of the Yuanmeng
volocopter-e1600755569645
model-t
Unearthing the Secret Superpowers of Fungus
cite-foret
190402173742-germany-coal-3-super-tease-2
unnamed-file-3
Photovoltaïque et terrains agricoles : un enjeu au cœur des objectifs énergétiques
Designer materials to keep plastic out of landfills
DO Black - the world’s first credit card with a carbon limit
L'Allemagne inaugure la première flotte de trains à hydrogène au monde
The Intersection