Once you’ve got a task to do, it’s better to do it than live with the fear of it. - Logan Ninefingers

We have developed an Artificial Intelligence algorithm called The Great Leveller (TGL) that massively improves over OpenAI’s GPT-2 in terms of practicality.

As an example we make use of Google Translate. We select the translation path English-Spanish-Latin-Greek-Russian-English. You can try this yourself.

Below is an example paragraph run through this model.

Having been a student for the majority of my life, I understand the pressures associated with grades. In middle school, the pressure came from your parents who expected high scores. In high school, the pressure shifts to something that seems more important for your future: getting into college. But when you finally get into college, regardless of which one, the pressure persists: get accepted into your major, an internship opportunity, graduate school, etc. Over time, the concept of a “grade” evolves into more than just a measure of academic progress, and students begin associating the number with their inherent skill or ability within certain subjects. It becomes part of their identity.

becomes

Even when I dedicated most of my life to these, I understand the urge of my compatriots to work with ratings. In high school parents are expected to score high. In high school, the change in some pressure is likely to be college. But when, finally, he entered the university, regardless of the impression he insisted, he took the opportunity of the film, graduate school, etc. Over time, the concept of “absolutely” is more than just an indicator of academic progress, and the number of students begins to the ability or ability inherent in certain subjects. This is part of their own identity.

One can then revise the result freely. Of course, the full algorithm is significantly more complex, involving a custom translator network, loss based on minimizing low-level similarity and maximizing high-level similarity, a learned language path, and selection of how much to translate at once.

We believe this new algorithm will help humanity work together to achieve free will for all.

Please email critique to albertwujj@gmail.com (Albert) or timerox888@gmail.com (Aaron). Feel free to experiment with our algo. After all, these are ideas, not original thoughts.