Meta has developed a single AI model that can translate between 200 distinct languages. The project is being open-sourced in the hopes that others will expand on it. While most machine translation models manage only a few languages, Meta's model handles them all. Machine translation might serve as the foundation for a game-changing tool for future projects such as augmented reality spectacles. It outperforms prior state-of-the-art machine translation software by 44 percent. However, such a small sample size cannot convey the expressive potential of a complete language. To investigate some of the challenges they experienced, Meta's engineers conducted a series of interviews with 44 speakers of various languages. The team's plan to open-source as many aspects of the project as possible, from the model to the dataset and training code, should assist to balance the power imbalance inherent in such an endeavor.