This week, Adam Engst joins us to talk about some work he did to try to measure the accuracy of several different AI models in creating audio transcriptions. He let me help him with some of the transcriptions so I was invested in his findings. We'll learn about ways to measure accuracy, including Word Error Rate aka WER. You can read Adam's article where he describes in detail the process he followed and the rather unsatisfying results at [tidbits.com/...](https://tidbits.com/2025/02/28/comparing-audio-transcription-in-notes-audio-hijack-and-macwhisper/).

Read an unedited, auto-generated transcript with chapter marks: CCATP_2025_03_23