@Em0nM4stodon@infosec.exchange I prefer to let speech-to-text dictate, then I go through and summarize key points myself. The thing is, all current transcription thingies see “AI” now. They don’t need to be. Speech-to-text has been “solved” for decades. It doesn’t need GPU-intensive “AI.” My Pentium III in 1999 could do a competent enough job.