According to experts, GPT-4 will have a staggering 100 trillion parameters, compared to the 175 billion of GPT-3. This increase in power means that the new model will be able to produce even more precise answers at a much faster rate. Like its predecessor, GPT-4 will be able to generate, translate, and summarize text, answer questions, serve as a chatbot, and generate content in an automated way. But it will also have new capabilities such as customer service and education, and it should be more efficient in interpreting human intentions during interactions, leading to fewer instances of misinformation. ChatGPT, which is a refined version of GPT-3, will be “like a toy” compared to GPT-4, says developer Simon Høiberg. The new technology is expected to be deployed by OpenAI in 2023 and is already causing a lot of excitement in the AI sector. It’s likely that the price of a new iteration of ChatGPT based on GPT-4 will be greatly increased, but it represents “the opportunity of the century” for professionals looking to integrate AI into their services.