Y
Hacker News
new
|
ask
|
show
|
jobs
by
serf
1287 days ago
run voice through 3 models, when one model is different than the other two then throw it away.
problems abound with that approach, but I could envision it.