I used to think voice typing was a poor technology invented for people who couldn’t type correctly. It’s full of errors after all, especially in Chinese, where many words share the same sound and differ only in tones.

I changed my view after I tried Speech Note, which integrates common voice recognition models. The evolution of voice recognition surprised me with its accuracy. Thanks to my graphics card’s CUDA, it transcrbes Chinese sentences in just a few seconds. Its performance is far better than Google’s voice recognition on my phone.

The minor errors from the model can be easily corrected with AI, making voice typing a much more efficient input method now.

Maybe I type too much and have forgotten how to speak. And voice typing is a great way to practice.