Large Language Models (LLMs) face significant obstacles in effectively assisting with coding tasks due to issues with tokenization, context windows, and training methods.
