Get instant answers using wikipedia and transformer models. Genie currently works best with research type questions (ie stuff you would look for on wikipedia).

Genie is somewhat accurate. Some examples include:

> genie-nlp "What is Agar.io?"
Answer: Agar.io is a massively multiplayer online action game created by Brazilian developer Matheus Valadares...

> genie-nlp What is the meaning of life?
Answer: The meaning of life, or the answer to the question: "What is the meaning of life?", pertains to the significance of living or existence in general... 

• Use rust bert and wikipedia to compile answers
• Make Genie give somewhat relevant answers
• Make Genie a command
• Use other methods to verify an answer's validity
• Make it robust and friendly
• Deploy it to a package manager (like Brew)

Written in rust. Libraries incoperated include

• rust-bert: transformer-based models implemented in pure rust
• wikipedia: a way to crawl and scrape wikipedia articles

The only transformer based model being used (for now) is RoBERTa, and it is used to determine if a given answer makes sense relative to the question (on a shallow level).