Nvidia, a leading global AI chip manufacturer, has launched an AI chatbot application that runs locally on PCs. This application, named Chat with RTX, features interesting capabilities that can assist users in various ways.
Summarizing Documents and Videos
One of the main features of Chat with RTX is its ability to summarize documents uploaded by users in .txt, .pdf, .doc/.docx, and .xml formats. Users can request a summary of these documents by sending them to the application, and the application will provide a brief and concise summary.
In addition to documents, Chat with RTX can also summarize YouTube videos by embedding their links. This application can provide important information from the video, such as the title, duration, description, and key points. The application can also provide recommendations for tourist spots or cake-making tutorials from YouTube creator videos by analyzing the content and keywords of the video.
Advanced Technology and High Specifications
Chat with RTX uses retrieval augmented generation (RAG) technology and Nvidia TensorRT-LLM software that harnesses the power of AI. This technology allows the application to search for and generate relevant and accurate information from the documents or videos provided by the user.
To run this application, users need a minimum device specification that is quite high, namely a GPU with 8 GB VRAM, 16 GB RAM, Windows 10 or 11, and Nvidia GPU drivers version 535.11 or newer. This application also requires 35 GB of storage space and does not require an internet connection when in use.
This application is free to download, although it is still in demo version. However, users must be patient during the installation process, and this application can sometimes be power-hungry. In addition, this application does not remember the context of the conversation, so users cannot ask follow-up questions from questions that have already been asked.
A Safer and More Private AI Chatbot
Chat with RTX is similar to Google’s Gemini chatbot or OpenAI’s ChatGPT, which also have document or video summarizing features. However, unlike cloud applications, Chat with RTX is a personal digital assistant that works locally on the user’s PC.
This means that this application does not send user data to Nvidia servers or other third parties, making it safer and more private. Users do not need to worry about their data privacy, as this application only accesses data directly provided by the user.
Nvidia’s Step to Retain Clients
The launch of Chat with RTX is one of Nvidia’s steps to retain their clients who plan to develop AI chip projects while continuing to collaborate with Nvidia. Nvidia is developing a special business unit focused on designing and selling semi-custom AI chips, which are developed collaboratively between Nvidia and the client.
Using Nvidia’s IP blocks as a basis, clients also have the option to integrate third-party IP blocks according to their needs. Nvidia estimates a market potential of USD 30 billion for this special business unit.
In addition, Nvidia is also partnering with Cisco, a network infrastructure provider, to offer technology solutions that make it easier for companies to build their own internal AI computing infrastructure while expanding the reach of AI technology beyond large data center providers.
Cisco will provide M7 server computers that use Nvidia’s latest processors, software, AI models, and development tools from Nvidia. Cisco will also provide its own cloud-based service to monitor and manage the system.
This collaboration opens the door to the use of Nvidia technology in various sectors, while Cisco can capitalize on the growing use of AI that has made Nvidia one of the world’s most valuable chipmakers.