Transformer-based language models are revolutionizing the way neural networks process natural language. This deep dive looks at how organizations can put their data to work using transformer models. We consider the problems that business may face as these massive models mature, including training needs, managing parallel processing at scale, and countering offensive data.