Mind blown, what a huge leap forward in language model performance! looking forward to seeing what this means for future applications. https://www.reddit.com/user/madkimchi