NVIDIA Went HARD at CES 2025 - RTX 5090 is a Beast

In the video, the presenter shares their experience at CES 2025, highlighting NVIDIA’s RTX 590 graphics card, which showcases impressive capabilities, including running the DeepS R1 AI model locally and integrating generative AI in ray tracing for enhanced gaming efficiency. They also discuss various innovative technologies at the event, such as autonomous farming equipment and robotics, while emphasizing the importance of open-source models for user privacy and security.

In the video, the presenter shares their experience at CES 2025, highlighting the overwhelming presence of AI technology and innovative gadgets. They express gratitude towards NVIDIA for inviting them, covering their hotel expenses, and providing them with an RTX 590 graphics card. The presenter is excited to showcase the capabilities of this new card, particularly its ability to run the DeepS R1 thinking model locally, which they plan to demonstrate later in the video.

The presenter describes some of the impressive technology they encountered at CES, including autonomous farming equipment from John Deere, which utilizes advanced AI for operations. They also mention the presence of autonomous vehicles, robotics companies like Unit, and various innovative robots on display. NVIDIA hosted a special event for a select group of attendees to preview upcoming technologies, including the RTX 590 and new open-source initiatives that allow developers to build upon NVIDIA’s frameworks.

One of the standout features of the RTX 590 is its integration of generative AI in the ray tracing process. The presenter explains that the card can render a frame traditionally and then generate subsequent frames predictively, enhancing efficiency in video games. This approach signifies a shift towards a future where AI plays a central role in gaming and software development, moving away from traditional game engines.

The video showcases the RTX 590 in action, where the presenter runs two versions of the DeepS R1 model, demonstrating its impressive performance. They provide a prompt to the AI to write a game in Python, showcasing the speed and efficiency of the 8 billion parameter model, which operates at around 90 tokens per second. The presenter then tests the larger 32 billion parameter model, which, while slightly slower, still performs remarkably well, emphasizing the card’s capability to handle extensive AI tasks locally.

In conclusion, the presenter expresses their appreciation for NVIDIA’s support and the opportunity to experience CES. They highlight the importance of open-source models that can be run locally, ensuring privacy and security for users. The video wraps up with a call to action for viewers to like and subscribe, as the presenter looks forward to sharing more content in the future.