Positron and Negatron are Machine Learning natural language models, they generate original text phrases to motivate or aggravate you, whichever you choose.

These models are based on OpenAI GPT-2 774M, and fine-tuned by publicly available internet content. Because they are trained on things humans have written on the internet, incidental rudeness may abound. These thousands of generated texts were all run through a plagiarism check again the dataset, using the process I outline here to hopefully ensure they don't return anything from the dataset verbatim.

The views expressed in these generated texts are not the views of the site, or of its' creator. This site is an experiment to show the unintended consequences of training ML/AI on the vastness of humankind's internet text.