From <https://freedom-to-tinker.com/2020/08/13/gpt-3-raises-complex-questions-for-philosophy-and-policy/>: GPT-3, a powerful, 175 billion parameter language model developed recently by OpenAI, has been galvanizing public debate and controversy ... Why the hype? GPT-3 is unlike other natural language processing (NLP) systems, the latter of which often struggle with what comes comparatively easily to humans: performing entirely new language tasks based on a few simple instructions and examples. Instead, NLP systems usually have to be pre-trained on a large corpus of text, and then fine-tuned in order to successfully perform a specific task. GPT-3, by contrast, does not require fine tuning of this kind: it seems to be able to perform a whole range of tasks reasonably well, from producing fiction, poetry, and press releases to functioning code, and from music, jokes, and technical manuals, to “news articles which human evaluators have difficulty distinguishing from articles written by humans”. Also note this: ... GPT-3 has been trained on us—on a lot of things that we have said and written—and ends up reproducing just that, racial and gender bias included. OpenAI acknowledges this in their own paper on GPT-3,where they contrast the biased words GPT-3 used most frequently to describe men and women, following prompts like “He was very…” and “She would be described as…”. The results aren’t great. For men? Lazy. Large. Fantastic. Eccentric. Stable. Protect. Survive. For women? Bubbly, naughty, easy-going, petite, pregnant, gorgeous.