On the topic of AI, I think there's 1 area you'd be hard pressed to find even the most staunch anti AI person arguing it's bad. Voice typing or voice coding for the purpose of accessibility, in the mess of new tech I think it's important to remember good use cases like this
Conversation
Notices
-
Embed this notice
Brodie Robertson (brodieonlinux@mstdn.social)'s status on Tuesday, 13-May-2025 09:26:13 JST Brodie Robertson
-
Embed this notice
Brodie Robertson (brodieonlinux@mstdn.social)'s status on Tuesday, 13-May-2025 09:37:56 JST Brodie Robertson
@frumble The way we discuss AI has also greatly changed over the years, fundamentally we're looking at statistical models and back in the 80's IBM's Tangora was making use of a Hidden Markov Model. This is the same sort of question as, at what point does the switch from calculator to computer happen.
-
Embed this notice
Maxi 11x 💉 (frumble@chaos.social)'s status on Tuesday, 13-May-2025 09:37:58 JST Maxi 11x 💉
@BrodieOnLinux Sure but voice recognition software was a thing even 20-30 years back. It has gotten better for sure with recent tech but it’s not like a whole new frontier.
-
Embed this notice
Brodie Robertson (brodieonlinux@mstdn.social)'s status on Wednesday, 28-May-2025 08:22:26 JST Brodie Robertson
@fireborn @J3317 At least from my usage, where whisper is wrong it tends to be so wrong that you can't even mistake what it is saying for something incorrect. So it's easy to go back and verify
-
Embed this notice
aaron (fireborn@dragonscave.space)'s status on Wednesday, 28-May-2025 08:22:27 JST aaron
@J3317 @BrodieOnLinux I totally agree. Speech-to-text with things like Open AI's whisper modal is really good now. Even as someone who doesn't need this tool I use it to transcribe audio messages I'd rather read than listen to.
-
Embed this notice