Interesting thoughts about the limits encountered in the quest for better speech recognition, the implications for probabilistic approaches to AI, and “mispredictions of the future”.
As noted in the comments, artificial natural speech recognition is in the 85-95% range, and human natural speech recognition is also around 95%. I was skeptical of that article when I first read it because it did not even mention how good humans are at it to compare to the machines.
The unrecognized death of speech recognition
Interesting thoughts about the limits encountered in the quest for better speech recognition, the implications for probabilistic approaches to AI, and “mispredictions of the future”.
What do y’all think?
As noted in the comments, artificial natural speech recognition is in the 85-95% range, and human natural speech recognition is also around 95%. I was skeptical of that article when I first read it because it did not even mention how good humans are at it to compare to the machines.