NO = more time
In many professional and private life situations, it is very hard to say "no" to requests. Because we know how rejection feels, we try to avoid denying to our colleagues. So, why not let a machine come up with a reason to not accept their request? :)
Voila, here we use the GPT-2 language model to generate us a reason (not an excuse :) ) to reject and let it roll a completely biased dice with all sites pointing to the same answer. By doing so, we try to gain more own time using the art of saying this simple complex word "NO". Enjoy !
Should I ?
Friendly NO
(C) 2020. A stupid :) one-day project by Hendrik Strobelt using huggingface transformers.
Disclaimer: Be aware that the model used (GPT2-medium) was trained on internet text and can become rude or offensive. I did spent most of my time trying to filter as much bad/controversial words as possible - but it's not perfect. The model output does not reflect my opinion (or those of my affiliations) in any way. If you want to know more about bias in models like GPT2 start reading here. The webpage uses Google analytics, but no other cookies.