NVIDIA will Kill Them All

The chip war between NVIDIA and Intel has become fierce. With the latest announcements at GTC, NVIDIA has taken on Intel with its AI compute platforms such as NVIDIA H100 NVL, L4, L40, and Grace Hopper.?

The chips have been designed specifically for the increasing demand in AI, with each model tailored for distinct inferencing tasks. For example, L4 is an accelerator chip for enhanced video decoding, streaming, and generating AI video. L40 provides the same functionality for graphics and AI image generation in 2D and 3D.?

The H100 NVL chip can be viewed as a duo of H100 GPUs that have been interconnected to provide a greater memory capacity of approximately 188GB of HBM2 VRAM. On the other hand, the final addition to this series, Grace Hopper, combines the processing capabilities of CPU (Grace) and GPU (Hopper) technologies, with the aim of handling massive datasets required for recommender systems and LLMs.

The latest range of these chips released by NVIDIA suggests that the company is shifting its focus away from general-purpose GPUs (GPGPUs) and instead, is prioritising the burgeoning field of AI. This move can be seen as a strategic step towards achieving the goal of diversification.

Also, it poses a threat to Intel’s specialised compute business, which includes products like Intel Xeon Scalable CPUs and Habana.?

For real-time inference, if Intel had Xeon Sapphire Rapids, NVIDIA has launched Grace Hopper in response. Besides, various tests also prove that Grace Hopper performs 2.5x more with 3.5x more efficiently than its competitors but is, as of yet, untested against Xeon’s ‘hybrid’ approach to AI compute.

NVIDIA's strategy of offering tailored compute platforms for different AI workloads has not only challenged Intel's market approach, but also presented a superior option, backed by a wide network of partners eager to adopt it.

Read the full story?here.?

OpenAI vs CohereAI

Although OpenAI's ChatGPT has recently gained attention for its enterprise offerings, Cohere AI has been providing similar solutions for some time now. What distinguishes Cohere is its approach to AI, which is centred around meeting the technological needs of its customers, rather than aiming for Artificial General Intelligence (AGI).?

The company's mission is to make its constantly improving technology accessible to all developers, with a low barrier to entry, as the process of training models requires a significant amount of computer data and coordination that many developers may not have access to.

Read the full story?here.

Redis the Lifeline of OpenAI?

OpenAI's ChatGPT was recently affected by a bug in an open-source library, causing some users to be able to view other users' conversation history and payment details. Fortunately, the Redis team was able to assist OpenAI in resolving the issue. Interestingly, Redis was unaware that OpenAI was using its software to scale ChatGPT until they were contacted to help fix the bug, which they were happy to do.?

Yiftach Shoolman, Redis' co-founder and CTO, noted that a bug report is one way to discover the usage of open-source software.

Read the full story?here.?

Tech Bros

Elon Musk, Sam Altman, and Sundar Pichai are some of the biggest names in the tech industry who have revolutionised the way we interact with technology today. However, not many know that they have equally successful siblings who are also making significant contributions to the industry. These lesser-known tech titans are making waves with their innovative startups and established ventures, proving that excellence truly runs in their blood.

Read the full story?here.