Wow, this is great, surely GPT-3 can provide more sophisticated results and can also help in improving the data set here. Would really appreciate if you would, could run the current data set through GPT-3 and share your results?

You can share it via opening an issue here (https://github.com/AkashRajpurohit/howtoprofessionallysay/is...), so it becomes a bit easier to track