This is what this post (or rather just the title, tbh) immediately reminded me of. I remember learning about it in the early 2000s. It was considered a very c

This is what this post (or rather just the title, tbh) immediately reminded me of. I remember learning about it in the early 2000s. It was considered a very challenging problem, with very important applications, most notably speech recognition in natural settings. I wonder what is the current status on this. Is this considered solved nowadays?