OpenAI is having a discussion about what to publicly announce for the accuracy of GPT-4.

Apparently, there is a public debate going on right now among the directors of OpenAI about what to publicly declare about the accuracy of GPT-4. <br><br>It's known that accuracy is one of the most important factors for an AI to be classified as AGI. So OpenAI is having many internal debates on whether to declare GPT-4 as AGI now because of the accuracy of 92% which is very close to that of human beings.<br><br>However, many directors are commenting that even if GPT-4 achieves an accuracy of 100% or higher, it cannot be considered AGI without reasoning capabilities like human common sense. <br><br>So, there is a strong argument that GPT-4 is not a real AGI because it's just a large language model with no reasoning capabilities. <br><br>Some have even gone as far as to say that if they decide to declare GPT-4 as AGI, they would be lying to the world. <br><br>One thing I can tell you is that GPT-4 is almost indistinguishable from human beings. But we'll see in the next 24 hours whether it's officially classified as AGI or not. <br><br>Source: OpenAI internal meeting