Chat GPT-4: multimodality and videos

Chat GPT-4: multimodality and videos

Berlin, March 10, 2023

According to a report from Heise.online (3/9/2023 3:53 p.m.=), OpenAI’s GPT-4 will be presented next week. Andreas Braun, the CTO of Microsoft Germany, says that the multimodality of the new model „will offer completely different possibilities such as videos,“ makes one sit up and take notice.

This note should make it clear that GPT-4 can analyze, react to and generate images and videos in addition to text.

The CEO of Microsoft Germany, Marianne Janik, who was also present, emphasized that AI is not used to replace employees, but to take on repetitive tasks in a new way.

In his speech, Braun described large language models as „game changers“ because they would teach machines to understand natural language. In a statistical way, the machines understand what is otherwise only understandable to humans. Obviously the Microsoft developers are so far that this works with „all languages“. Thus, the multimodality will contribute to Microsoft being able to make the models „overarching“.