I Beta Tested OpenAI's Codex, and the Results Are Spooky Good

Last week, artificial intelligence company OpenAI launched Codex, a new deep-learning-driven platform which writes fully functioning software code automatically. The system -- which was trained on a vast corpus of publicly available code-- originally debuted as part of Github's Copilot, a feature which helps programmers improve or update their software automatically. Codex is based on OpenAI's wildly successful GPT-3. When I tested GPT-3 last year, I felt like I was witnessing a technological revolution. The system can generate everything from fully-formed blog posts to songs, recipes -- even sea shanties.