Fake Google Gemina video with GPT-4 model recreated for real + video

After Google released a stunning video demonstrating Gemini in action, it turned out to be a fake. Now a user has recreated this video with the help of Vision in GPT-4. In this video, we see that what Gemina claimed to do, is done by OpenAI artificial intelligence.

According to a recent report, a YouTube user posted a video showing off the power of Gemini's main competitor, the GPT-4. Greg Technology's YouTube channel has released a video in which the Gemini demo has been recreated in GPT-4. Unlike Google's fake demo, this video is recorded in real time with voice prompts.

In this video, the user asks this artificial intelligence to recognize his movements with the help of GPT-4 vision. After showing a heart with his hand, this user was able to get the correct answer from artificial intelligence with a little delay. Also, GPT-4 was able to recognize the duck painting shown to Gemina very quickly. OpenAI artificial intelligence was even able to detect the rock, paper, scissors game, but showed more delay in responding.

Performance of GPT-4 artificial intelligence model compared to Google Gemina

Google's Gemina Large Language Model (LLM) is the company's most powerful set of AI models to date. In an effort to demonstrate the multifaceted capabilities of its large language model, Google released a video of Gemina, which appears to respond to a user's voice requests quickly and in real time. At first, this demo surprised many, but then it became clear that Google has made manipulations in the output display to make the video look smoother and the performance of this artificial intelligence model more advanced.

In this regard, the video of Greg Technology's channel can be considered very important, because this user conducted this experiment with OpenAI artificial intelligence model, which is considered the main competitor of Gemina. But unlike Google's demo, this video looks very real.