[ad_1]
rewrite this content and keep HTML tags (remove this from content : rewrite this content and keep HTML tags)
You need a maximum of three AI models in everyday life. Jensen Huang knows that too.
Jensen Huang is the CEO of Nvidia, making him one of the most successful businesspeople in the world. After all, Nvidia is the second most valuable company in the tech sectorand ranked 47th among all companies worldwide
At the GTC in Paris yesterday (June 11, 2025), he mentioned that he uses three AI models in his private life:
ChatGPT
Gemini Pro
Perplexity
Why only three AI models?You might think that the CEO of one of the world’s largest companies, which also laid the foundation for modern AI companies, would use totally secret or special AI models. But that doesn’t seem to be the case.
In my opinion, the answer to the question of why only three AI models is simple: you don’t need any more. Most everyday needs can be easily covered with three models.
Why these models?That’s exactly what I’m going to tell you now. Or rather, I’ll tell you why I use these three models. I don’t know why Huang uses them, but I assume that our reasons at least overlap.
Let’s get started.
1. ChatGPT – The classic
You may have expected something more exciting, but ChatGPT from OpenAI laid the foundation for the ever-growing AI market and remains at the forefront when it comes to general tasks: writing emails, proofreading texts, brainstorming, and so much more.
ChatGPT often provides detailed and well-structured answers that are also well suited for creative tasks. ChatGPT is an all-rounder and will hardly disappoint you in everyday tasks. OpenAI’s AI often takes a little longer to do this – although in this case, “longer” means just a few seconds.
ChatGPT also likes to provide additional information that you haven’t asked for, but which is often useful.
ChatGPT is often the starting point for me. If I’m not satisfied with the results, I may switch to other models, such as Claude or Gemini Pro.
And what about Huang?I don’t know if Huang sees it the same way I do, but I suspect that one of the reasons he likes to use ChatGPT is because Nvidia’s hardware helped make ChatGPT big. At least that’s how I would feel.
2. Gemini Pro – The alternative
Gemini, like ChatGPT, has a lot to offer, but the results are often presented differently. Gemini is concise and likes to present answers in bullet points or sentences.
Unlike ChatGPT, Gemini often gives you exactly the answer you wanted. No more, no less. Whether you prefer this to the additional information provided by ChatGPT is a matter of taste.
Personally, I often only use Gemini when I am not satisfied with ChatGPT and don’t feel like spending time on prompt engineering.
Gemini has also helped me more with mathematical topics so far. However, when it comes to math, I can only recommend using several models and comparing results, because even though some models are better than others, math is still a difficult field for language models.
3. Perplexity – Google, only different (and better?)
Admittedly: It’s certainly not easy to break the habit of Googling, but over time it has more or less happened on its own. The reason for this is AI models, which not only give me a more precise answer to my question more quickly, but also answer follow-up questions.
Unfortunately, many AI models (including ChatGPT) are still quite unreliable, and complex topics or in-depth technical questions can sometimes be nonsense.
The solution is Perplexity This is a language model that searches the web for you in no time at all and presents the results neatly.
You don’t just get information that basically comes from humans and is only summarized by AI. You get at least one source for each statement, which you can check yourself afterwards.
This way, you can be sure that the information you receive is accurate – at least as accurate as you can be when it comes to information from the internet.
Maybe Jensen Huang will read this article somedayand tell us exactly why he uses these three models.
Until then, write your favorite models in the comments below!
[ad_2]
Source link
