There’s a tendency nowadays to throw automation at everything. Especially in problems involving machine learning and data science.
But sometimes, crowdsourcing human labor may be just as, or even more effective. For example, we human beings are very good at processing images and videos. Computers are not that good. Yet. (Good thing too, otherwise we’d probably be slaves to our computer overlords!)
As a result, many systems involving crowdsourcing have been developed. Perhaps the most famous of these is Amazon’s Mechanical Turk system. You’re probably familiar with it. And you might have even used it yourself!
However, to truly optimize your Mechanical Turk experience, you need to know exactly what kind of tasks such an approach is useful for. You also need to figure out what is a fair remuneration for your Turkers.
Edwin Chen has more to say about this subject. (Remember him? And his unique views on solving AI problems? I talked about this in an earlier article.) His insights on how to make this system work for you is well worth a read. Take a look.