Something I'm wondering: is it relatively easy to just dive into these libraries and start doing useful things with facial recognition? Or is it more of a toy

Something I'm wondering: is it relatively easy to just dive into these libraries and start doing useful things with facial recognition? Or is it more of a toy at this point? I have a few ideas in that general direction, hence my asking.