[j_barbossa]

Crowdsource by Google aka "Help a billion dollar company to create datasets for free"

[mhh__]

Make that a trillion

[samatman]

Rare sighting of the long scale in the (English speaking) wild!

[visarga]

We are already providing search clicks and video likes as indicators of content quality. They train AIs on our preferences.

[krono]

It's not working so well though, in my experience anyway