Answering the question!
This question arises as Gemini is a new and promising language model developed by Google. While ChatGPT has gained popularity and recognition, Gemini’s entry into the field may pose a potential challenge. It will be interesting to see how these two models compete and evolve in the future. Taking into account the data provided by Google on their blog, we know that Gemini Ultra beats GPT-4. Benchmark data is on Google’s side and indicates the performance of their product.
Benchmark data provided by Google below:
Gemini is the first model to outperform human experts on MMLU (Massive Multitask Language Understanding), one of the most popular methods to test the knowledge and problem solving abilities of AI models. ~ Google writes on its website.
Mind-Blowing Features!
Google released a video presenting another interesting feature of its product, which is beyond my understanding at the moment. Gemini will be able to play games with you when he sees your camera. Google showed us 2 examples. Gemini, seeing a map of the world on the camera, asked riddles with descriptions of different countries. The person had to point with his finger where it was. But don’t click on the screen! Fingers on the camera. It’s not everything. When Gemini was provided with a camera image showing 3 cups, the model decided to play the world-famous game with 3 cups and a ball. I still don’t believe that such things are possible.
Even though I have described many functions and statistics in which Gemini crushes its rival, this list is not the end. As it turned out, this model can do even more. On the website we can see how Gemini, seeing a photo of two yarns and asked what can be made with it, Gemini is able to generate photos of potential products to sew from the yarn. Another example is an animation of birds flying. Gemini provided the exact code to program such animations. For me, this model is an entry to a higher level. It cannot only transform text. According to Google, it will be able to transform text, sound, recordings and even photos, all in real time!!!!
To sum up, in my opinion, if Google is able to deliver the product as they currently describe it, we will see a change in the leadership in terms of the “best” AI model. I’m really positive and I hope that they will release Gemini without any unnecessary delays or errors.
Here is a short 90 seconds film delivered by Google about their new product in which they shortly describe what can we expect when the Gemini launches:
Sources:
https://deepmind.google/technologies/gemini/#introduction
https://www.youtube.com/watch?v=qu9-aPDoStc
https://www.youtube.com/watch?v=_TVnM9dmUSk
Very interesting article! It’s amazing how many different chats are already created and how helpful they are for us. Personally I really like the feature where Gemini can provide us with the ideas of what to create from what we have. Can’t wait to test it and hope they will release it soon.