Take a look at GPT-3 doing a programming phone screen: https://twitter.com/lacker/status/1279136788326432771 https://twitter.com/lacker/status/12791367883264327

Take a look at GPT-3 doing a programming phone screen: https://twitter.com/lacker/status/1279136788326432771 https://twitter.com/lacker/status/1279136788326432771 This is an unedited output of a language model which was not trained to pass programming phone screens. The only thing it was trained to do is to predict the next word in a sentence. Just 10 years ago this would be pure sci-fi. Today it's "meh". More importantly, there's no sign of slowing down.