Variants: * use OpenAI's text-davinci-003 model: https://pypi.org/project/openai-summarize/ * different approaches