TECH
In just a month, the ChatGPT neural network, which has become the most popular in the world, has not yet had time to reach its full potential. For example, she was recently taught how to create games in a computer language.
Although artificial intelligence is capable of improving, it is still an ordinary computer program that needs to be updated and refined.
OpenAI developers presented GPT-4, a new version of the language model, which is not only able to work with text, but also learned to process information from images. Experts call this update the ultimate in deep learning technology.
The main feature of GPT-4 is the ability to recognize information in photos. The neural network is now able to perceive data not only in text form, but also to process visual data. For example, if you ask the AI to describe what is happening in the image, the neural network will describe the image.
"GPT-4 is more robust, creative, and capable of handling much more complex tasks than the previous version of GPT," the developers note.
Even the new version of the GPT can produce errors - for example, the network calls Elvis Presley the actor's son. Therefore, the CEO of OpenAI admits that humanity is still a long way from developing a perfect AI.
by: mundophone
No comments:
Post a Comment