A smarter way for large language models to think about hard problemsnews.mit.edu2 points·pedriquepacheco··0 commentsOpen articleSaveView on HN