[edit: lost context after parent edited]

What I can think of is a smartphone app that's always listening and transcribing, best-effort, all it listens. This can capture in person speech and make it searchable.

Also, are any of the current videoconferencing options offering machine transcribing out of the box? I know users could always hand-feed recorded video to a separate tool, but ease-of-use matters.

Emulating face-to-face-ness remotely is precisely the hard problem. As pg once said, the real world is incredibly high-bandwidth.