Everything To Know About The 'Chat With RTX' AI Chatbot

While generative AI chatbots have been around for a couple of years, the fact that all rely on cloud processing poses challenges in handling sensitive data due to instances of hackers accessing such information from servers. In response to this issue, Nvidia has introduced its own AI chatbot called “Chat with RTX,” enabling users to host a personalized AI chatbot on their computers without an internet connection. Here’s everything you need to know about it.

Prerequisites

Before delving into the specifics, it is important to understand that Chat with RTX is a standalone application downloadable on your PC. As a result, it requires at least 40GB of free space and an RTX 30- or 40-series GPU with a minimum of 8GB of VRAM.

Capabilities of the Chat with RTX chatbot

Screenshot of the Chat with RTX AI model

Now that we’ve addressed the prerequisites, let’s discuss what you can do with the chatbot. Running on the Mistral or Llama 2 AI language model (depending on your preference), the potential applications are significant, especially in areas such as data research for employees or journalists, where analyzing multiple confidential documents is crucial.

In addition to document analysis, users can input video URLs and ask the chatbot to search transcripts for specific mentions or create comprehensive video summaries. Moreover, unlike Microsoft’s Copilot, which struggles with PDFs, the Chat with RTX chatbot can easily scan PDFs and fact-check data. Furthermore, the near-instantaneous response times could make it the new favorite tool for many professionals.

However, it is important to note that despite its features, the chatbot is still in beta and exhibits some bugs and issues. These include instances where the chatbot downloads transcripts of random videos, crashes, and displays inaccurate source information. Moreover, the chatbot currently lacks context retention, requiring users to start a new chat for follow-up questions.