The incredible news from the world of Artificial Intelligence: here are the chatbots that recreate user behavior. The news comes directly from OpenAI via a tweet announcing the launch of Gpt-4. The new Generative Pre-trained Transformer model takes advantage of the Deep Learning learning technique, in which artificial neural networks learn to perform tasks easily. Only recently, the artificial intelligence research organization OpenAI launched the ChatGpt “intelligence” chatbot, especially successful for the younger generation.
But let’s see specifically the new functions provided by this system.
Much more action
Accuracy will be one of the main skills provided by Gpt-4, especially in problem solving, thanks to the greater knowledge and functions compared to the previous model.
Gpt-4 will easily handle very large texts that can contain up to 25,000 words. This will allow the system to create much more extensive and complex content and conversations with the user than previous models.
Multimodality is another of the relevant characteristics of this system: from now on even images can be interpreted artificially. You can automatically generate captions or analyze images by reading them. It is expected that in the future the answer to every possible question posed to Gpt-4, can also be extended to video and audio. Furthermore, it will be possible to interpret and summarize pdf files that contain images, graphs and tables in a few simple steps using this intelligence.
Artificial intelligence goes further
Gpt-4 breaks records never seen before: it is positioned in the 10% of the best aspiring lawyers in the test to enter the profession in the USA. On the contrary, the model that precedes it was in the last places. This important result is given by the precise and codue to preparation and training on the system, which began in August 2022. But its ability to recreate human behavior lies in gathering user feedback and incorporating it into its own system.
Sam Altam, CEO of OpenAi, proud of the new technology, however, declares that the Gpt-4 will certainly have to be perfected and improved in some small functions such as, for example, suggestions to users, which are sometimes contradictory. The goal of Gpt-4: to spread transparency in content and educate users’ language.
How you will be able to use the Gpt-4
Microsoft has invested heavily in OpenAi artificial intelligence research, in fact Gpt-4 is backed by Microsoft Azure.
The Gpt-4 can be requested by ChatGpt Plus subscribers immediately, in the paid version at 20 dollars a month. Even if in reality to access the service, even subscribers will have to join the waiting list.
To use Gpt-4 it will be possible to use Bing in the chat version known as Bing Ai, but Microsoft has already implemented this system too.