going-beyond-computer-tools-to-ai-skills-|-nextbigfuture.com

Going Beyond Computer Tools to AI Skills | NextBigFuture.com

Nvidia CEO Jensen Huang describes the AI revolution as going beyond computer tools to AI skills.

They are already revolutionizing the data center with compute accelerating GPUs.
The $2 million AI GPU racks have far denser compute than regular CPU powered servers. The cables connecting the old larger racks cost more than the new denser and smaller $2 million GPU servers.

The GPU servers are small enough to be liquid cooled at affordable cost.

Converting the $1 trillion of existing CPU data centers is already a multi-year valuable business.

Computers with AI can write software.

AI can understand data and not just work with data.

Generative AI enables conversion of all kinds of data from text to images and text video or large text to small text and the reverse for each.
It can also convert data into proteins.

They enable AI skills but what are AI skills.

Skills are AI chauffeurs, AI lawyers, AI customer service etc…

For every $1 spent on AI Infrastructure, AI providers are getting $5 worth of rentals.

Benefits of NVIDIA-Accelerated Spark
The GPU acceleration provided by NVIDIA for Apache Spark offers several advantages:

Performance Improvement: Companies like Adobe have reported up to 7x performance improvement in initial tests
Cost Savings: GPU acceleration can lead to significant cost reductions, with Adobe reporting 90% cost savings in their tests
Enhanced Model Accuracy: The performance gains allow data scientists to train models with larger datasets and retrain them more frequently
Unified Infrastructure: AI model training can be processed on the same Spark cluster as data processing tasks, eliminating the need for separate infrastructure

Nvidia Blackwell systems are actually 7 chips and many systems. Customers using Blackwell will get three times more revenue versus using systems using the previous Hopper chips and systems.

Jensen describes the value of their systems and how keeping pace will make it difficult to beat the Nvidia moat

Article Source




Information contained on this page is provided by an independent third-party content provider. This website makes no warranties or representations in connection therewith. If you are affiliated with this page and would like it removed please contact editor @riverton.business

Similar Posts