After the automated captions on YouTube, you now get improved captions that can even recognize sounds to let you know what's happening when people aren't talking.
You've probably laughed a lot with YouTube's automated captions. But should you try it again, you'll notice that it has improved... a lot. Not only it has become a lot better at understanding (correctly) what people are saying, now it can even go beyond words.
Should you enable caption on YouTube, you'll find that it can now describe other types of sounds, like applause, music, and laughter..
Though speech recognition is a very complex affair, recognizing these sounds isn't that much easier; even more so, if you consider it can be happening simultaneous (like music and applause, or applause and laughter). But, Google's machine learning systems seem to be up to the task, and you can check it out yourself in videos like the one below.
Enable closed captions and enjoy...