It's weird zoom's automatic transcription can't consistently distinguish speakers at transitions - it should be able to run separately on each stream and get the attribution right basically 100% of the time... Though maybe there's a problem with the mics capturing output from speakers.

listening to the US Copyright Office "Technical Measures" opening plenary discussion

certainly some sparks. unfortunately, the plenary is a day long and I have to work :(

I'll try to post the recording when it's posted.