资讯

Wired3 年

It’s Like GPT-3 but for Code—Fun, Fast, and Full of Flaws

OpenAI’s new tool sped up my work and wrote entire pieces of software from a simple prompt. But it might riddle the internet with even more bugs.
The New York Times2 年

Let Us Show You How GPT Works — Using Jane Austen

An interactive explanation of how language models learn to mimic language, from Shakespeare to Star Trek.