At first glance, NVIDIA seems to have only released the RTX 40 SUPER series of desktop graphics cards on paper during this CES 2024 exhibition, without much action, but this is not the case.
Especially in the field of AI, which is the most adept and convincing field, NVIDIA has come up with a number of technological innovations, involving multiple application scenarios such as games, images, videos, and live broadcasts, once again demonstrating its leadership position.
Kuai Technology was fortunate enough to visit a series of DEMOs displayed by NVIDIA, and I will share them with you next.
[ACE Technology and Game AI NPC]
ACE, also known as Avatar Cloud Engine, can use the most advanced generative AI model to generate virtual digital characters that can interact naturally in games and applications.
At the same time, there are AI models such as Audio2Face (A2F) that can generate rich and natural facial expressions, and Riva Automatic Speech Recognition (ASR) can be used for multi-language speech translation.
With this series of new technologies, game NPCs are no longer the fixed routines of the past, but can interact with players naturally. Players can ask him any questions, such as how is the situation recently, give me a bowl of noodles, etc. .
Even for the same question from different players, the AI NPC will give different responses. The result is that every conversation and every interaction is unique.
AI NPCs also have the ability to perceive scene objects. You can ask them what an item does, and they can also pick up items and send them to designated areas.
In addition, AI NPC can also lead players through the game world to reach the target location.
Developers currently using ACE technology include:Convai, Charisma.AI, Inworld, MiHoYo, NetEase Games, Palm Technology, Tencent Games, Ubisoft, UneeQ.
【Chat With RT】
This DEMO can connect users' documents, notes, videos or other data with large language models to make it more personalized.
Using Retrieval Enhanced Generation (RAG), TensorRT-LLM, and RTX acceleration,Users can talk to custom chatbots and get the answers they need faster and more accurately.
in addition,it is run locallybecause it is more secure and does not leak privacy.
[AI drawing and editing]
Since the birth of Stable Diffusion, AI drawing has become familiar to everyone, and various commercial services have begun to appear. iStock is a typical example.
Based on NVIDIA hardware acceleration, iStock can quickly generate the pictures or illustrations you want online based on input keywords.
What’s more considerate is that you can choose different aspect ratios such as 1:1, 3:4, 4:3, 9:16, 16:9, etc. to avoid secondary cropping. You can also choose different tones or modes, such as Warm, cold storage, gorgeous, natural, black and white, etc.
All generated images can be used freely without infringement.
Of course, premium services require payment.
Generating pictures is only the initial stage. After all, the pictures produced by AI may not fully meet our needs, which requires further editing functions.
AI Picasso is a drawing software that can produce and edit pictures. It can combine existing photos to generate the required pictures. You can freely edit and change parts of the picture (such as changing clothes for a character), and you can perform AI expansion and so on.
It can also be combined with advanced professional software such as DaVinci Resolve Studio for more dimensional editing and creation.
[RTX Video HDR video enhancement]
Everyone is familiar with video enhancement technology. Nowadays, with the assistance of AI, there are more ways to play and the results are better.
NVIDIA has previously launched RTX VSR video enhancement technology, andThe latest RTX Video HDR can convert any SDR video played in the browser into HDR video, which uses AI to expand the dynamic range, making the dynamic light and shadow effects more natural and more in line with actual visual effects.
To put it bluntly, it means that the areas that should be bright can be made bright enough, and the areas that should be dark can be dark enough, so that the overall scene has more contrast.
【RTX AI PC and Workstation】
The concept of AI PC has become popular, and GPU is undoubtedly the most critical and powerful part of it. NVIDIA RTX naturally has unrivaled advantages in this regard, and with its powerful GPU, it extends to AI workstations.
Acer, Asus, Dell, HP, Lenovo, MSI and other manufacturers are releasing new RTX AI notebooksbringing users an out-of-the-box generative AI experience.
Compared with the NPU AI engine that can only handle low-intensity reasoning loads, RTX AI notebooks rely on stronger computing power and better optimization and adaptation.The performance improvement can be as much as 20-60 times.
Mobile workstations equipped with RTX GPUs can run NVIDIA AI Enterprise software, including TensorRT and NVIDIA RAPIDS, for simplified and safe generative AI and data science development.
Lenovo Yoga Pro 9iequipped with RTX 4070 Laptop GPU, supports advanced DLSS 3.5 functions such as super resolution, light reconstruction, and frame generation, allowing for real-time and smooth 3D rendering.
ASUS ROG Phantom 16equipped with RTX 4080Laptop GPU, demonstrates RTX-enhanced AV1 video encoding, and the export speed can be directly doubled.
Samsung Galaxy Book 4 Ultraequipped with RTX 4070 Laptop GPU, demonstrates object removal enhanced by RTX AI, which can quickly and completely remove unwanted debris from photos and intelligently complete the corresponding scene without leaving any traces.
Dell XPS 16equipped with RTX 4070 Laptop GPU, AI performance is far ahead of Apple.
Finally, let’s look at a live broadcast enhancement technology, which is currently used on the overseas live broadcast platform Twitch.Using an RTX PC, live streams of different weights can be generated for different devices, providing the most appropriate bit rate, resolution, and frame rate.
For example, high-definition TV displays 4K60, high-performance desktops display 1440p60, ordinary notebooks display 1080p60, tablets display 720p30, mobile phones display 480p30, and so on.
It is true that almost every technology company is talking about AI nowadays, especially the semiconductor industry that provides the foundation of AI software and hardware, and is investing heavily in AI.
Among them, NVIDIA is undoubtedly the brightest star. Whether it is AI accelerator hardware or CUDA ecological software, it is absolutely far ahead. Now it has fully brought generative AI into the PC consumer market, from all dimensions Subverting long-standing experience.
NVIDIA is also highly recognized by the capital market. The company's stock price has recently experienced another surge, hitting new highs for many consecutive days. In the past week alone, it has increased by nearly 50 US dollars, an increase of nearly 10%. As of the time of writing, it has reached 547.10 US dollars, corresponding to The market value is as high as 1.35 trillion US dollars!
In addition, on the latest list of the TOP25 global semiconductor companies in 2023 released by the semiconductor research organization Tech Insights,NVIDIA ranks third with annual revenue of nearly US$50 billion (estimated), rising five places from the previous year.
The total revenue of the 25 companies fell by 11%, and the top 10 combined also fell by 9%, but NVIDIA itself soared by as much as 102%, while the others either fell or increased by only single digits.
These are undoubtedly the best proof of NVIDIA’s AI strength.

[End of this article]If you need to reprint, please be sure to indicate the source: Kuai Technology
Editor in charge: Shang Shangwen Q




















