Solving a machine-learning mystery | MIT News | Massachusetts Institute of Technology

MIT researchers have explained how large language models like GPT-3 are able to learn new tasks without updating their parameters, despite not being trained to perform those tasks. They found that these large language models write smaller linear models inside their hidden layers, which the large models can train to complete a new task using simple learning algorithms.
— Read on news.mit.edu/2023/large-language-models-in-context-learning-0207

Could ChatGPT do my job? | MIT Technology Review

This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here. In the last week there has been a lot of talk about whether journalists or copywriters could or should be replaced by AI. Personally, I’m not worried. Here’s why. So far, newsrooms have pursued…
— Read on www.technologyreview.com/2023/01/31/1067436/could-chatgpt-do-my-job/