←back to thread

470 points ph4evers | 3 comments | | HN request time: 0.964s | source

I've been working on a little side project that combines Duolingo-like listening comprehension exercises with real content .

Every video is transcribed to get much better transcripts than the closed captions. I filter on high quality transcripts, and afterwards a LLM selects only plausible segments for the exercises. This seems to work well for quality control and seems to be reliable enough for these short exercises.

Would love your thoughts!

1. owenpalmer ◴[] No.43548734[source]

I absolutely love the idea. I would honestly use this. However, when I tried the English learning, it incorrectly marked words wrong several times. Something to check out.

replies(1): >>43549900 #
2. ph4evers ◴[] No.43549900[source]

Thank you! It seems that the video is pretty well transcribed but it unluckily selected segments with a few words missing :/

replies(1): >>43553181 #
3. owenpalmer ◴[] No.43553181[source]

The words were correct in the YouTube subtitles, both in spelling and sequence. It must have been a problem with the transcription or some other bug (whitespace?)