Well you kinda do. You take the previous model and you double the processing power. When GPT3 announced it was kinda clunky but chatGPT fixed many of it's issues,so I'm guessing GPT4 will fix most of the issues of chatGPT,and then GPT5 will fix all the remaining issues of GPT4 but will make it even more useful,and by the time GPT6 will be announced it will already be 2 times better than GPT5,so it's safe to assume we will reach AGI by then. We are in a thread that everyone just speculates things based on when we will reach singularity,so why is there even a reply like that. So based on the processing power of the previous models and the rapid growth of technological innovation,I do feel AGI will be reached by GPT6.