#010 Saturday Special - Your Next Performance Review Will Be By Confirm, US Bans Nvidia’s GPUs to Middle East, Nvidia Aims to Increase Its GPU Production by 300% in 2024
Fresh & Hot curated AI happenings in one snack. Never miss a byte 🍔
This snack byte will take approx 5 minutes to consume.
AI BYTE # 1 📢- Your Next Performance Review Will Be By Confirm (Raised $ 6.2 Million)
⭐ Performance reviews are often dreaded by both managers and employees. They can be subjective, biased, and inefficient.
But what if there was a better way to evaluate and reward talent in your organization?
That’s the vision of Confirm, a San Francisco-based startup that has raised $6.2 million in Series A funding to transform the performance review process from the ground up.
Confirm uses Organizational Network Analysis (ONA) and Generative AI to deliver fairer, more scientific, and more efficient performance reviews.
ONA is a method that visualizes and analyzes the formal and informal relationships in your organization. It measures employee performance by examining how all employees in the company view one another.
ONA reveals not only who the high and low performers are, but also who the real influencers within the company are.
Generative AI is a technology that can create realistic text and images. Confirm uses OpenAI’s GPT-4, to generate customized performance reviews for each employee with input from their peers and managers.
It also auto-generates employee survey results and auto-calibrates ratings for employees to minimize bias from any one particular manager or another.
Confirm was founded in 2019, but it has already attracted clients like Canada Goose, Niantic, and Thoropass. Thoropass, for example, managed to retain all of its top performers during the wave of employee turnover known as “The Great Resignation” thanks to Confirm’s platform.
It has also won several awards and recognitions, such as a “World Changing Ideas Award” from Fast Company, an HR Tech Award for Best Talent Intelligence Solution from Lighthouse Research & Advisory, and a spot in SHRMLabs’ 2023 WorkplaceTech Accelerator program.
Confirm’s mission is to create a world where employees are recognized and rewarded for their hard work and positive impact, not their ability to play office politics.
In a time where remote and hybrid teams are commonplace, data-driven performance reviews matter more than ever.
AI BYTE # 2 📢 - Now US Bans Nvidia’s A100 and H100 GPUs to Middle East (New Export Limits Due to U.S.-China AI Rivalry)
⭐ I always like to keep an eye on the geo-political developments in the field of AI.
The U.S. government has now imposed additional restrictions on exports of Nvidia’s high-performance compute GPUs to certain regions, including some countries in the Middle East.
The reason behind this move is to prevent China from acquiring these GPUs, which are essential for training and running advanced AI models.
China is already restricted from buying Nvidia’s A100 and H100 GPUs, which are among the most powerful chips in the market.
Nvidia has created alternative products, such as A800 and H800, with lower performance to sell in China.
However, the U.S. government is concerned that the high-powered and latest GPUs could be diverted to China from customers in the Middle East, where Saudi Arabia and the UAE have been investing heavily in AI research and development.
These countries have also been collaborating with China on various AI projects, which could pose a threat to the U.S.'s global leadership in AI.
The U.S. government is worried about Chinese companies training AI models overseas due to a domestic shortage of top AI GPUs and then bringing that technology back to China.
This could give China an edge over the U.S. in developing and deploying AI applications in various domains, such as military, healthcare, education, and entertainment.
Nvidia has said that the licensing requirement does not significantly impact its revenue and that it is working with the U.S. government to address the issue.
AI BYTE # 3 📢 - Nvidia Aims to Increase Its GPU Production by 300% in 2024
⭐ Nvidia, the leader in AI and HPC compute GPUs, has just reported a record-breaking revenue of over $10 billion in one quarter.
But the company is not resting on its laurels. According to the Financial Times, Nvidia plans to at least triple its output of data center-oriented compute GPUs in 2024.
The demand for Nvidia’s flagship H100 compute GPU, based on the GH100 processor, is so high that they are sold out well into 2024.
The H100 is a powerful and complex chip that can generate realistic text and images using Nvidia’s CUDA framework. It also supports HBM2E or HBM3 memory and uses TSMC’s CoWoS packaging.
To achieve its ambitious goal, Nvidia will have to overcome several challenges.
First, it will have to secure enough 5nm-class wafer supply from TSMC, which is shared among other major clients like AMD and Apple.
Second, it will have to ensure enough supply of CoWoS packaging and HBM memory packages from TSMC and other vendors.
Third, it will have to coordinate with its partners to increase the production of AI servers that use H100-based devices.
If Nvidia can pull it off and the demand for its AI and HPC products remains strong, this could mean incredible revenue and market share for the company.
It could also cement its position as the dominant player in the AI and HPC fields, where its CUDA framework is widely used and preferred by many applications.
Mental Disorders are a function of too many unnecessary options