OpenAI’s GPT-4o-powered ChatGPT can now sing, crack jokes, teach languages, solve math problems & more (WATCH)
OpenAI’s GPT-4o-powered ChatGPT is what Apple’s Siri, Samsung’s Bixby, Microsoft’s Cortana, and Google Assistant should have been right from the beginning. The abilities it has gained from GPT-4o make it as human as an AI can be.
![OpenAI GPT 4o powered ChatGPT can now sing crack jokes teach languages solve mathematics problems and more watch gcw](https://static-gi.asianetnews.com/images/01hq81eqdsn1hetkj7g4xfezs6/gettyimages-1246494152-594x594--1-.jpg?impolicy=All_policy&im=Resize=(1200) )
The new ChatGPT upgrade, known as GPT-4o, is causing a stir in the artificial intelligence community. It's not just a little improvement; it's one of the most significant improvements we've seen in an AI model thus far, and it's not even a full generational update.
Finally, GPT-4o-powered ChatGPT is what Apple's Siri, Samsung's Bixby, Microsoft's Cortana, and Google Assistant should have been since the beginning. This latest version of ChatGPT is extremely adaptable; it can crack jokes, perform songs, sense and interpret human emotions, and even add sarcasm to the mix.
OpenAI introduced GPT-4o on Monday, describing it as a significant step towards making human-computer interaction seem more natural. The interesting thing about this model is that it can accept text, audio, or photos as input and reply in any of these media. It is also intelligent enough to recognise emotions and keep up with a conversation nearly as rapidly as a person.
Mira Murati, CTO of OpenAI, highlighted during a live-streamed presentation that GPT-4o provides the intelligence level of GPT-4 to everyone, including those who utilise the free version. This makes utilising AI much more convenient for everyone.
During the lecture, OpenAI demonstrated some remarkable GPT-4o skills. It translated live between English and Italian, assisted a researcher in solving a maths issue on paper in real time, and even provided advice on deep breathing simply by listening to someone's breath.
Unlike its predecessors, which required users to voice questions and then transcribe the answers into text, GPT-4o handles all inputs and outputs using the same neural network. This implies that interactions are more smooth and retain tone and emotion, resulting in faster and more natural responses.