> Lots of interesting ML output, even if they've somehow failed to capitalise on it. > The problem is what to do with the other 170,000 employees. This is act

> Lots of interesting ML output, even if they've somehow failed to capitalise on it. > The problem is what to do with the other 170,000 employees. This is actually the problem and solution in one go - they're all doing data annotation.