3 ms·
30,000 hours of video per hour. That's not that much. We can do a lot with speech to text as a starting point. And you can easily detect if speech to text isn'
by tertius 6y ago
30,000 hours of video per hour.
That's not that much. We can do a lot with speech to text as a starting point. And you can easily detect if speech to text isn't working very well.
And for those that are graphic you can do image detection. It's not out of reach.
- jedimastert 6y ago> We can do a lot with speech to text as a starting point. And you can easily detect if speech to text isn't working very well. > And for those that are graphic you can do image detection. It's not out of reach. YouTube does both of those things already. How well is it working out? Honestly, surprisingly well. But many, probably most, would argue not nearly enough, especially if you're legally liable for every frame of that video
- tertius 6y agoI'm just arguing that it's not as big of a problem as some make it seem.