Unlike the previous OpenAI AI, the GPT-4 can interpret up to 25,000 words, understand, and respond to pictures.
If you were one among those who were impressed with ChatGPT? “OpenAI,” the ChatGPT developers, have officially announced the introduction of their latest AI version, known as the GPT-4, so prepare to be blown away.
Until now, ChatGPT has been run on the GPT-3.5, whose performance is nothing near that of the GPT-4. According to OpenAI, the ChatGPT powered by GPT-4 will be able to grasp up to 25,000 words and even recognise and respond to images.
This means you may request a recipe from the GPT-4 by simply emailing it a snapshot of your ingredients. Following an examination of the image, the chatbot will provide a detailed description of your dish. Should we include “cookbook authors” to the list of vocations that AI is likely to replace?
The GPT-4 is trained on a far larger database than the previous edition, making it far more accurate and powerful. After using the GPT-4, Microsoft CEO Bill Gates declared that he is now a fan of ChatGPT.
OpenAI has spent more than six months establishing a variety of safety precautions, but it continues to warn users that the GPT-4 may provide erroneous information on occasion. According to the company, the AI can “hallucinate” on occasion, which means it can generate facts that are completely untrue and employ erroneous logic.
The GPT-4 is already available to ChatGPT Plus subscribers as an integrated function on the Microsoft Bing browser. A subscription that costs $20 to join.