Jensen Huang, the CEO of Nvidia, recently offered insights at the GTC in Paris regarding the AI models he utilizes in his daily routine. Despite overseeing one of the largest tech companies globally, Huang depends on merely three AI models: ChatGPT, Gemini Pro, and Perplexity. This selection demonstrates a pragmatic approach to AI, indicating that most daily requirements can be effectively addressed with a limited selection of models.
**1. ChatGPT – The Classic**
ChatGPT, created by OpenAI, serves as a multifunctional resource for general activities such as composing emails, editing content, and brainstorming ideas. Renowned for its thorough and organized replies, ChatGPT often includes extra valuable details. It functions as a dependable starting point for many, including Huang, possibly owing to Nvidia’s involvement in its creation.
**2. Gemini Pro – The Alternative**
Gemini Pro provides succinct responses, often in bullet format, making it a favored option when accuracy is essential. It particularly shines in mathematical subjects, although users are encouraged to validate results with other models due to the inherent difficulties language models encounter with mathematical concepts.
**3. Perplexity – Google, Only Different**
Perplexity functions as an advanced search tool, offering exact answers along with follow-up information. It distinguishes itself by citing sources for its data, ensuring both accuracy and dependability. This model proves especially beneficial for confirming intricate or technical details.
Huang’s preference for these models emphasizes their ability to handle a broad array of tasks effectively, illustrating that even within the AI domain, quality frequently outweighs quantity.