Dictate something into just about any recent software, even if it's for the first time, and it will almost certainly spit something out that’s at least understandable. Make it good enough, release it, and let the machines start learning.įor all its flaws, voice transcription already does rough approximation pretty well. But if you only need to translate a news article from Spanish, it’s typically good enough. Think about it this way: You wouldn't use Google Translate to convert Borges into English. It only needs to get just enough right to let two people speaking two languages grope out some meaning.
And the thing is, it doesn’t need to be perfect. Why It's Time for Devices to Understand What We Mean, Not Just What We Sayīut not impossible.
0 kommentar(er)
