A smarter way for large language models to think about hard problemsnews.mit.edu·2 pts·pedriquepacheco·0