ChatGPT will soon be unbeatable on all subjects, that’s why

OpenAI is said to be designing the next generation of GPT, the language model used in ChatGPT and Dall.e, artificial intelligence technologies made available to the public since September 2022. GPT-4 promises to be even more powerful than his predecessor.

ChatGPT, the chatbot developed by OpenAI, is all the rage right now. It is a platform based on GPT-3, a machine learning model trained on large amounts of data to produce human-like text. If the motivations for using it are diverse and varied, everyone agrees that it is extremely advanced and impressive. Thanks to this AI, hackers can hack any website, and students can generate the text of their exams.

To read – ChatGPT AI writes so well it can fool scientists

Microsoft appreciates this technology so much that it plans not only to integrate it into its Microsoft 365 office suite, but also to become a majority shareholder of OpenAI. If the Redmond firm offers itself the OpenAI patents and implements ChatGPT in Bing, Google’s leading position in the field of search engines could clearly be threatened. Indeed, following a request, Google provides a list of links to check out, ChatGPT gives a real answer.

Thanks to GPT-4, the next generation of ChatGPT is likely to be even more powerful

The fourth iteration of the chatbot, ChatGPT-4, may be scary as it will be exponentially more powerful. Little information has yet filtered about this generation of AI, but it is rumored that it should be launched this year. The CEO of OpenAI claims that GPT-4 will offer no less than 100 trillion parameters, while the GPT-3 model only offers 175 billion. This makes not far from 571 times more distinct settings.

Read — ChatGPT will become paid, a premium “Professional” version is coming soon

Thanks to this increase in the precision of the answers, the conversational quality of ChatGPT should be even more human and indistinguishable. One of the criticisms of the current ChatGPT is that it can easily slip into misinformation. The next generation of the chatbot may also fix this