Here’s a demonstration of the new autocorrect and predictive text features coming to the Apple operating systems this fall… This is a post for MacSparky Labs Members only. Care to join? Or perhaps do you need to sign in?
MacWhisper, the app that uses an AI Large Language Model to create transcripts, just got its third major update. Improvements include:
- New full transcript mode
- Podcast mode, which can transcribe a podcast by adding the audio files for each host.
- System App transcription. Record and transcribe your zoom meetings or any other audio source on your Mac
- Faster speed (up to 40%).
I can’t help but feel that we are at an inflection point with voice to text and it is getting a lot easier to use and implement. MacWhisper is part of this wave.
I’ve been running tests on iOS 15 voice-to-text dictation. The below video demonstrates an iOS 15 vs. an iOS 14 comparison on two iPhones with the same processor. iOS 15 is slightly faster and more accurate. Because it is being done on-device, you are also freed from the whims of your Internet connection.
Where iOS 15 really pulls ahead is the removal of the blasted timer. No longer do you have to randomly restart dictation as you dictate. This will be a game changer when it releases. The below video shows demonstrates this difference.